The Homeownership Experience of Low-Income and Minority Families

U.S. Department of Housing and Urban Development

Ofﬁce of Policy Development and Research

The Homeownership Experience of

Low-Income and Minority Families

A Review and Synthesis of the Literature

The Homeownership Experience of

Low-Income and Minority Families

A Review and Synthesis of the Literature

Prepared for

U.S. Department of Housing and Urban Development

Office of Policy Development and Research

Prepared by

Christopher E. Herbert

Eric S. Belsky

Abt Associates Inc.

Cambridge, MA

February 2006

Acknowledgements

The authors would like to thank Bulbul Kaul for her assistance in analyzing the American Housing

Survey, Mark Duda for his assistance in reviewing the literature on the financial returns to

homeownership, and Missy Robinson for her efforts to produce the final report. In addition, we

would like to thank Harold L. Bunce, William J. Reeder, and Marina L. Myhre of the U.S.

Department of Housing and Urban Development for their careful and thoughtful review of the report.

The contents of this report are the views of the contractor and do not necessarily reflect the views or

policies of the U.S. Department of Housing and Urban Development or the U.S. Government.

Foreword

Throughout American history, homeownership and the independence that goes with a place

of one's own has been a primal aim of Americans. Every President in modem times has sought to

improve opportunities for homeownership and President Bush has placed special emphasis on

extending the

American Dream

to low-income and minority households with his goal and efforts to

bring about

5.5

million new minority homeowners by the year

2010.

The benefits of owning a home are well known. Perhaps best known as a principal source of

wealth accumulation for a majority of Americans, it also provides a greater degree of insulation from

rising housing costs allowing more income to be saved or put to other uses. Research has also

documented a range of non-financial, individual and social benefits of homeownership, including

improved housing quality, satisfaction. and conditions for

childhood development. However, the

extent to which low-income or minority homebuyers share in the traditional benefits of

homeownership noted above has received far less attention.

HUD's Office of Policy Development and Research initiated this study,

77ze Homeownership

Experience of Low-Income and Minority Families:

Review and Synthesis of the Literature.

provide a comprehensive examination of the homeownership experience of low-income and minority

families over time and determine the extent to which low-income and minority homeowners are

themselves reaping the traditional benefits of homeownership. Four shorter empirical papers listed

below were also completed in conjunction with the study:

I. The Impact of House Price Appreciation on Portfolio Composition and Savings

The Growth of Earnings of Low-income Households and the Sensitivity of Their Homeownership

Choices to Economic and Socio-Demographic Shocks

Wealth Accumulation and Homeownership: Evidence for Low-Income Horseholds

Is Manufactured Housing a Good Alternative for Low-Income Families? Evidence from the American

Housing Survey

The study finds that over the last ten to fifteen years, large numbers of new low-

income and minority homeowners have made

good

initial choices, buying good quality

homes in decent neighborhoods. More generally, the study finds that low-income and

minority owners are as likely as others to benefit from homeownership. These owners

are

just as likely to

see

their homes appreciate

others; in fact, for the vast majority, housing

wealth

their only source of wealth. There is also evidence that children of low-income

owners have greater success in education and the job market than children of low-income

renters.

The study also finds that low-income and minority families are more apt to encounter

difficulties sustaining homeownership and points to the importance of continuing and enhancing

policies that help owners weather and resolve circumstances that threaten their ability to sustain

homeownership. While the study identifies many areas where our understanding could be enhanced

with more research,

takes a giant step toward providing the basic understanding that will support

policies that ensure homeownership provides lasting

benefip to low-income and minority families.

barlene

Williams

Assistant Secretary for

Policy Development and Research

Table of Contents

Executive Summary ............................................................................................................................iii

: Introduction..................................................................................................................1 Chapter One

Individual Benefits of Homeownership.......................................................................................3

Process for Realizing the Benefits of Homeownership................................................................6

Outline of the Report ...................................................................................................................9

Related Studies...........................................................................................................................10

Chapter Two: Initial Housing Choices Made by Low-Income Homebuyers................................11

Trends in the Number and Characteristics of First-Time Homebuyers .....................................13

Housing Choices of Low-Income Buyers..................................................................................18

Neighborhood Characteristics....................................................................................................24

Housing Costs............................................................................................................................30

Mortgage Financing Choices.....................................................................................................33

Summary....................................................................................................................................39

Chapter Three: Key Experiences and Decisions of Low-Income and Minority Homeowners....44

Mobility, Defaults, and Length of Time as Homeowners..........................................................44

Mortgage Financing Choices After Purchase ............................................................................56

Investing in Home Maintenance and Improvements .................................................................63

Summary....................................................................................................................................67

Chapter Four: Financial Benefits of Homeownership....................................................................72

Housing Appreciation and the Financial Return to Homeownership ........................................73

Comparing the Costs of Owning and Renting ...........................................................................81

Homeownership and Long-Run Wealth Accumulation.............................................................91

Summary and Conclusions.........................................................................................................95

Chapter Five: Social Impacts of Homeownership.........................................................................100

Impacts on Children.................................................................................................................101

Impacts on Social Involvement................................................................................................107

Impacts on Psychological and Physical Health........................................................................111

Summary and Conclusions.......................................................................................................116

Chapter Six: Summary of Findings, Policy Implications, and Areas for Further Research.....118

Summary of Findings...............................................................................................................118

Policy Implications ..................................................................................................................124

Areas for Further Research ......................................................................................................129

Bibliography......................................................................................................................................134

Appendix A: Estimate of Tax Benefits of Homeownership by Income .......................................148

Table of Contents

Executive Summary

Aided by a favorable economic climate, concerted efforts by the public and private sectors alike have

succeeded in significantly increasing homeownership rates for low-income and minority households

across the country over the last decade. Despite these gains, efforts to increase homeownership

opportunities continue to receive important emphasis from policy makers as significant gaps still

remain in homeownership rates by income and race/ethnicity. But the success of efforts to increase

homeownership has highlighted the need for policy makers to evaluate the extent to which these new

low-income and minority homeowners are reaping the expected benefits of homeownership, and, if

not, what can be done to increase the chances that they will realize these benefits. In fact, in recent

years both housing advocates and the popular press have raised concerns that the emphasis on

promoting homeownership may be luring families and individuals into buying homes when they

would be better off renting. These critiques cite rising foreclosure rates, increases in the share of

buyers shouldering substantial financial burdens, and accounts of buyers being trapped in poor quality

homes as evidence that a move to homeownership is in many cases not beneficial for the low-income

and minority households who are the focus of these efforts.

In the interest of supporting the development of effective policies for promoting and supporting

homeownership, as well as to address the concerns raised about those who fear there is too great an

emphasis on promoting homeownership, the purpose of this report is to review and synthesize what is

known about the homeownership experience of low-income and minority households to assess the

extent to which homeownership is likely to benefit these groups. While there have been several

recent reviews of the literature that have assessed the empirical evidence on the benefits of

homeownership, this study is unique in an explicit focus on what is known about the homeownership

experience of low-income and minority households.

Efforts to promote homeownership are routinely justified by a range of financial and social benefits

that are thought to result from owning a home. One of the key rationales for encouraging

homeownership is that it is the principal source of wealth accumulation for a majority of Americans.

Other financial benefits include reduced income tax obligations, protection from inflation in housing

costs, and a resulting increased ability to amass other savings. But equally important are a range of

non-financial or social benefits, including improved housing quality and satisfaction, increased social

engagement (with positive impacts on the surrounding community), enhanced conditions for

childhood development, and improved psychological and physical health.

Summary of Findings

Overall, our own analysis and review of the literature documents the significant increase that occurred

during the 1990s in the number of low-income and minority first-time homebuyers. For the most

part, low-income and minority households were found to have made good initial choices in the homes

they buy, obtaining good quality housing in decent neighborhoods. The share of these homes that are

moderately or severely inadequate is only about 7 percent—no worse than the average for the country

as a whole. While it does not seem that the move to homeownership is associated with an

improvement in neighborhood conditions for many buyers, the new neighborhoods on average seem

Executive Summary

iii

to be decent areas with few signs of distress. There is also no evidence that these first-time buyers

faced higher interest rates than other purchasers—at least on their first-lien mortgages through 2003.

The most worrisome aspect of the situation of these new homeowners is that roughly one-in-five

buyers during 1995 to 2003 faced a severe payment burden, spending more than 50 percent of their

income on housing.

The study primarily relied on a review of the existing literature to assess whether over time low-

income and minority homeowners were as likely as other owners to realize the financial and social

benefits of owning a home. Our general conclusion is that for the most part these owners are as likely

as others to benefit from homeownership. With regard to homeownership’s financial benefits, these

owners are just as likely to see their homes appreciate in value as other owners. Since housing is a

leveraged investment, even modest appreciation in value combined with paying down mortgage debt

over time, results in fairly significant wealth accumulation. In fact, for the vast majority of low-

income households, housing wealth is their only source of wealth. In terms of social benefits, there is

modest evidence that owners do benefit from improved psychological and physical health, although

the research is not strong and there has been little attention to whether there are differences in these

outcomes for different income or racial/ethnic groups. Moreover, there is fairly convincing evidence

that the children of low-income owners have greater educational success, and more modest evidence

that they have greater success in labor markets, are less likely to have behavioral problems, and are

more likely to become homeowners themselves.

Nonetheless, there is also evidence that low-income and minority individuals and families face a

greater risk of being unable to sustain homeownership. Since the benefits of homeownership mostly

accrue slowly over time, a failure to maintain homeownership will greatly reduce the chance of

realizing these benefits. While it can be argued that the risk of foreclosure remains fairly low for

most owners, recent research on the rate at which households exit homeownership find that for every

household that faces foreclosure there are several more who voluntarily leave their homes. Several

recent studies have used longitudinal panel surveys to trace the tenure choices of households over

fairly lengthy periods of time and found that between 43 and 53 percent of low-income buyers will

not sustain homeownership for more than five years, compared to between 23 and 30 percent of high-

income buyers. These studies also find that minorities at all income levels are between 22 and 39

percent more likely to leave homeownership than whites. These statistics reveal that the notion that

“once an owner, always an owner” is not at all true—especially for low-income and minority

families. While there may be substantial benefits from sustained homeownership, there are also

significant costs of failed attempts at owning. Cases ending in foreclosure undoubtedly impose

significant financial and personal costs on these families. Much less is known about other early exits

from homeownership, but these situations may also impost non-trivial financial and personal costs to

the extent that owners are compelled to leave homeownership.

The research that has been done on exits from homeownership draw upon data that extends back

before the sharp rise in homeownership rates in the 1990s. Thus, it is not the case that these relatively

quick exits from homeownership are a new development. But there is reason to believe that the

homeownership gains of the 1990s may have increased the number of owners at risk of being unable

to sustain homeownership. Perhaps most importantly the development of more flexible mortgage

products has made it possible to buy a home with higher levels of debt, lower levels of savings, and

worse credit histories than was previously possible. Thus, it would be expected that these buyers

Executive Summary

would also face a greater risk of being unable to meet their mortgage obligations. The

homeownership boom of the 1990s also brought many more single adults into homeownership, who

may have less ability to carry their mortgage obligations in the wake of a financial crisis than

households headed by two adults.

Policy Implications

Nonetheless, given the benefits that result from sustained homeownership, there is no reason to retreat

from the goal of increasing homeownership opportunities for low-income and minority households.

There is, however, a clear need for policies to increase the likelihood that homeownership will be

sustained and its full benefits realized. A concerted policy effort to improve homeownership

experiences will have three broad thrusts: efforts to improve the initial homebuying choices made by

these families and individuals—including whether owning is the right choice; efforts to ensure that

homeowners optimize their mortgage choices after purchase and make appropriate investments in

maintenance and improvements to their homes; and efforts to help owners resolve crises that threaten

their ability to sustain homeownership. For the most part, there are a variety of existing efforts to

support homeowners in each of these areas. As a result, the recommendations may be thought of

more as an indication of where greater emphasis is needed rather than where there is currently a lack

of effort. Among the specific approaches that need to be emphasized are pre-purchase counseling to

ensure that prospective homebuyers make informed choices about buying a home, post-purchase

counseling to provide support for families once they are in their homes, affordable refinance

programs to help owners minimize the costs of homeownership, and loss mitigation programs to

provide options for owners in financial crises to help them keep their homes.

Areas for Further Study

This review of the existing literature has also revealed a number of areas where not enough work has

been done to fully understand the circumstances facing homeowners, the nature of their decisions, or

the outcomes realized. Further research is needed to provide a better understanding of the extent to

which low-income and minority families and individuals benefit from homeownership as well as the

challenges they face in sustaining homeownership over time. Perhaps one of the most important

issues identified in this review is that roughly half of first-time low-income homebuyers are not able

to sustain homeownership for at least five years, with minorities faring slightly worse still. Relatively

little is known about the experience of these households as homeowners—what challenges they face

and what resources they have to respond to these situations. Perhaps the most important area for

further research is to gather better information about the experience of low-income homeowners.

This information is needed for policy makers to be able to identify the type of support that is needed

to ensure that low-income and minority households are able to sustain homeownership over time to be

able to realize its many financial and social benefits.

Outline of the Report

This report consists of six chapters. Chapter 1 describes the motivation for the study, outlines the

benefits associated with homeownership, and describes the process by which these benefits are

realized. Chapter 2 uses the American Housing Survey to examine trends in the characteristics of

first-time homebuyers since the early 1990s and describes the housing choices made by these

homebuyers with an understanding that these choices have important implications for the likelihood

Executive Summary

that the long-run benefits of homeownership will be realized. Chapter 3 examines the choices and

experiences of low-income and minority homebuyers after they buy their first home, including how

often they move and leave homeownership, their experience with home maintenance and remodeling,

and their choices about when to refinance and the mortgage terms they obtain upon refinancing.

Chapter 4 reviews the literature evaluating the financial returns to homeownership and whether low-

income and minority homeowners are less likely to experience financial gains than other owners.

Chapter 5 focuses on what is known about the social benefits of homeownership for low-income and

minority owners, including impacts on owner’s psychological and physical health and the well-being

of their children. Each of these chapters concludes with a detailed summary of findings. The report

concludes in Chapter 6 with an overall summary of findings, a discussion of the policy implications

of these findings, and an identification of areas where further research is needed.

Executive Summary

Chapter One:

Introduction

There was a notable shift in federal housing policy in the early 1990s to place greater emphasis on

efforts to promote homeownership for low-income and minority families.

To be sure, federal

support for homeownership had a long history prior to this time—from the creation of the Federal

Housing Administration and Fannie Mae in the 1930s, through providing financial guarantees for

long-term, low downpayment mortgages that helped fuel the tremendous post-war homeownership

boom, to the interest rate subsidy programs introduced in the 1960s. But as Carliner (1998) notes, for

the most part the primary goals of these efforts were not to increase homeownership rates but rather

were designed either to spur economic activity, provide a benefit to returning servicemen, or remedy

urban blight.

The efforts that began in the 1990s were distinct from these earlier efforts in their explicit focus on

the importance of expanding homeownership opportunities for low-income and minority households

with the goal of helping these families realize the benefits associated with homeownership. In

announcing his administration’s commitment to increasing homeownership, President Clinton

justified this goal by describing the benefits of owner-occupied homes:

They make for a more secure environment for our children. They create pride and

self-esteem. They are the extension of our personality, our hopes, our dreams. For

most of us, they're the main harbor of all of our collected memories. They are the

most important investment in financial security that most Americans every make.

And most people who own homes care more about their own communities and have a

bigger stake in solving the kind of problems that we've been here talking about

today.

In June 2002, when President Bush announced his own commitment to increase minority

homeownership rates, he also described the benefits of homeownership as the motivation for this

effort:

It is a key to upward mobility for low- and middle-income Americans. It is an anchor

for families and a source of stability for communities. It serves as the foundation of

Throughout this report, “low-income” will be defined as households having income less than 80 percent of

the median household income for the geographic area where the household resides. The geographic area is

defined as either the metropolitan area or the non-metropolitan portion of the state. Moderate income is

defined as household income that is between 80 and 120 percent of the area median household income,

while high income is greater than 120 percent of the area median.

Speech by President William Clinton at the National Association of Realtors Conference, Anaheim,

California, November 5, 1994.

Chapter 1: Introduction

many people's financial security. And it is a source of pride for people who have

worked hard to provide for their families.

Notably, federal efforts to promote homeownership have not relied heavily on federal outlays, but

instead use the bully pulpit and regulatory powers to spur the public sector to heightened efforts to

reduce barriers to homeownership, most notably a lack of information about the process, an inability

to qualify for mortgage financing, and discriminatory treatment. One of the most important new tools

at the government’s disposal has been the housing goals for the government-sponsored enterprises

(GSEs), Fannie Mae and Freddie Mac, which were created as part of the Federal Housing Enterprise

Safety and Soundness Act of 1992. Through the housing goals, the government induces the GSEs to

lead the mortgage market in expanding access to mortgage capital for low-income and minority

individuals and communities. The government also fosters homeownership through mortgage

insurance programs administered by the Federal Housing Administration, the Veterans

Administration, and the Rural Housing Service of the Department of Agriculture, which are financed

in large part by premiums paid by borrowers. In terms of direct efforts by the federal government, the

HOME program, created by the Cranston-Gonzalez Act of 1990, is the largest source of federal funds

for homeownership, committing $3.9 billion for homebuyer efforts from 1992 through 2004. The

Bush Administration’s American Dream Downpayment Initiative of 2002 was designed to increase

use of the HOME program to fund downpayment assistance for first-time homebuyers. In addition to

the HOME program, another concrete effort to promote homeownership by the government was to

increase financial support for homeownership counseling from $3.5 million in 1991 to $12 million

during the mid-1990s, and ultimately to $45 million by 2004 (Hornburg, 2004).

With the assistance of sustained economic growth and historically low interest rates, these efforts

succeeded in fostering a dramatic increase in homeownership among all segments of society

beginning in the early 1990s and continuing through 2004. Between 1993 and 2004, homeownership

rates among very low-income households, blacks, and Hispanics increased by 6.4, 7.7, and 8.7

percentage points, respectively (Herbert et al., 2005). These sharp increases in homeownership were

all the more remarkable coming as they did on the heels of more than a decade of stagnant or

declining homeownership rates (Green, 1996).

Despite these gains, efforts to increase homeownership opportunities continue to receive important

emphasis from policy makers as significant gaps still remain in homeownership rates by income and

race/ethnicity. But the success of efforts to increase homeownership has highlighted the need for

policy makers to evaluate the extent to which these new low-income and minority homeowners are

reaping the expected benefits of homeownership, and, if not, what can be done to increase the chances

that they will realize these benefits. In fact, while the gains in homeownership have widely been

hailed as a significant accomplishment, in recent years there has been a growing chorus of concerns

that the emphasis on homeownership may have gone too far (Baker, 2005; Apgar, 2004; Coy, 2004;

Kosterlitz, 2004; Shlay, 2004; Pitcoff, 2003). A common theme in these articles is that the single-

minded pursuit of homeownership as a solution to the housing needs of low-income families has in

some cases made families worse off. The expansion of mortgage underwriting has made it possible

for homebuyers to become financially over extended and far too often to end up losing their homes, at

Speech by President George W. Bush, quoted in “A Home of Your Own: Expanding Opportunities for All

Americans,” June 2002.

Chapter 1: Introduction

significant financial and personal cost. Furthermore, even if buyers are able to maintain their housing

payments, they may be stuck in poor quality housing or devoting an excessive share of their income

for housing. In short, while the goal of expanding homeownership is meant to allow these households

to realize the many potential benefits of homeownership, including wealth accumulation, residential

stability, and better social outcomes for the owners and their children, critics have come to question

whether many low-income and minority buyers have actually been able to realize these benefits.

In the interest of supporting the development of effective policies for promoting and supporting

homeownership, as well as to address the concerns raised about those who fear there is too great an

emphasis on promoting homeownership, the purpose of this report is to review and synthesize what is

known about the homeownership experience of low-income and minority households to assess the

extent to which homeownership is likely to benefit these groups.

The primary methodology of this

study will be to review and synthesize the relevant literature from academic, public policy, and

housing industry sources. While there have been several fairly comprehensive literature reviews

assessing the benefits and costs of homeownership generally (McCarthy, Van Zandt, and Rohe, 2001;

Rohe, McCarthy, and Van Zandt, 2002; and Dietz and Haurin, 2003), this review differs in having an

explicit focus on low-income and minority homeowners. In addition, while it is not the primary

purpose of this study to conduct original research, unlike these other studies some amount of

supporting descriptive analysis is used to document and evaluate the experience of low-income

homeowners. This review is also intended to serve as a basis for identifying the types of policies and

programs that are needed to mitigate the risks, maximize the benefits, and minimize the negative

impacts of homeownership for low-income and minority households. Public policy must also

recognize that there may be circumstances where homeownership is not recommended for certain

households given the low likelihood of realizing the benefits of homeownership. Finally, there is also

a growing recognition that we know less about the homeownership experiences of low-income

families than we know about the causes of homeownership disparities by income and race-ethnicity.

Thus, a final goal of this review will be to highlight the areas where further research is needed to

enhance our understanding of this issue and to better inform the policy-making process.

In order to frame the discussion to be presented in this report, the remainder of this introduction

outlines the benefits that are believed to be associated with homeownership and describes the process

by which these benefits may or may not be realized. The chapter concludes by presenting the outline

for the remainder of the report.

Individual Benefits of Homeownership

Advocates of efforts to promote homeownership cite a wide variety of benefits that accrue to both

individual homeowners as well as to society more broadly. The focus of this report is on the benefits

that are realized by individual homeowners and so for the most part the report will not discuss societal

Herbert et al. (2005) provide a comprehensive review of the literature to synthesize what we know about

the causes of differences in homeownership rates by race/ethnicity and income as well as policies to

promote homeownership.

Chapter 1: Introduction

benefits.

Individual benefits of homeownership are generally divided into two classes: financial and

social.

Financial Benefits

One of the principal financial benefits of homeownership is as a vehicle for wealth accumulation,

both through appreciation in value and through the forced savings associated with paying down

outstanding mortgage principal. Wealth accumulation through homeownership is enhanced by tax

law provisions that shield most appreciation in home values from capital gains taxes.

One of the

unique aspects of homeownership as a vehicle for wealth accumulation is that it is one of the few

leveraged investments available to households with little wealth allowing homeowners with very little

equity in their homes to benefit from appreciation in the overall home value. For example, a buyer of

a $100,000 home with a $5,000 downpayment will experience a 100 percent return on their

investment if home prices rise by a mere 5 percent in the first year of ownership. Of course, financial

leverage is a two-edged sword and housing is not a risk-free investment. If home prices were to fall

by 5 percent, the buyer’s initial investment would be wiped out. As long as the owner can continue to

meet their monthly mortgage obligations they can recoup these losses over time assuming home

prices recover and assuming they are not forced by other circumstances to have to move. But should

owners experience a simultaneous loss in income and housing equity, which can happen during

economic recessions, there may not be a way to avoid the lose of the home through foreclosure.

It is also important to note that a key factor in the financial returns to homeownership is the high

transaction costs associated with buying and selling homes. Real estate agent fees alone are typically

five to six percent of the sales price. In addition, sellers can face transfer taxes, legal fees, or buyers’

closing costs paid by the seller. If buyers are forced to move either shortly after buying or during a

down market, these transaction costs can greatly erode or eliminate any financial returns to

homeownership.

Nonetheless, equity in homes is the single largest source of wealth for all households, and is

particularly important for low-income and minority households. In 2000, housing equity accounted

for 32.3 percent of aggregate household wealth, with stocks and mutual fund shares accounting for

the next largest share of wealth at 15.6 percent (Orzechowski and Sepielli, 2003). But among

households in the lowest income quintile, housing equity accounted for 56.2 percent of aggregate

wealth, while stocks and mutual funds only accounted for 7.7 percent. Home equity is also a very

important source of wealth among minorities, accounting for 61.8 percent of aggregate wealth among

blacks and 50.8 percent of aggregate Hispanic wealth. Recognition of the critically important role

One category of societal benefits relates to improved neighborhood conditions (such as higher quality

public services, better maintained properties, and higher levels of property appreciation) that are argued to

result from higher levels of homeownership. This report will touch upon this category of benefits to the

extent that owners themselves benefit from improved neighborhood conditions. Another class of societal

benefits relates to improved macroeconomic performance due to higher levels of investment in housing that

is associated with owner-occupants. This latter issue is beyond the scope of this paper.

As of 1998, capital gains of up to $250,000 for single filers and $500,000 for married couples filing joint

returns may be exempt from taxation.

Chapter 1: Introduction

that homeownership plays in wealth accumulation is one of the keystones supporting efforts to

promote homeownership among low-income and minority households.

There are two other ways in which homeownership is thought to contribute to an individual’s

financial well being. First, owner-occupants are insulated from rapidly rising housing costs,

particularly if they have fixed-rate financing. Because the real cost of housing declines over time,

homeowners can have greater capacity for accruing savings in other financial assets or can enjoy a

higher level of consumption. Second, the deductibility of mortgage interest and property tax

payments serves to lower the after-tax cost of homeownership, also contributing to owners’ ability to

increase savings or consumption, although many low-income owners may not benefit from these

provisions as the standard deduction often exceeds interest and property tax payments.

Social Benefits

There are also a wide variety of non-financial benefits attributed to homeownership, generally

referred to as social benefits. One of the principal social benefits is that owners are thought to have

higher satisfaction with their homes, both in terms of the housing unit itself as well as the

neighborhood where they live. A key factor in homeowners’ greater housing satisfaction is that as

owners they have greater ability and incentive to invest in their homes to suit their tastes. For this

reason, homeownership rates increase as households age and enter a more home-centered phases of

life, typically as they begin to raise children. Of course, the flip side of owners’ ability to invest in

the home as they see fit is the responsibility for maintaining the home. Individuals who do not have

the interest or ability for conducting routine housing maintenance may find this aspect of

homeownership to be more of a burden than a benefit.

In addition to the ability to investment in the home, homeowners may enjoy higher quality housing

because of the segmentation of housing units between owner and renter markets. Larger and higher

quality housing bundles are more likely to be available in the owner-occupied housing market. Rossi

and Weber (1996) note that the rental stock includes many fewer single-family detached housing

units, while the owner-occupied housing stock includes many fewer units in multifamily structures.

While to some extent differences in housing demand between owners and renters may account for

some of these observed differences in the type of units that are owned and rented, it is nonetheless

true that in many areas anyone seeking a single family detached home will have many more options

among units available for sale compared to units available for rent.

The argument that owner occupants are more likely to be satisfied with their neighborhoods is based

on the idea that owners are both more likely both to invest in their own homes and to be actively

engaged in efforts to improve their neighborhoods to protect their investment. To the extent that

homeowners tend to cluster together, the collective activities of owners to improve their communities

and their individual units would be expected to result in better neighborhood conditions.

Herbert (1997) estimated measures of the degree of segregation in 1990 between owners and renters as

measured by the dissimilarity index for 50 metropolitan areas. He found that the degree of segregation by

tenure was moderate, suggesting that homeowners do, in fact, tend to cluster in neighborhoods. While

segregation by tenure was much lower than the levels of segregation experienced by blacks, it was similar

to the levels of experienced by Hispanics and Asians and higher than segregation by income or education.

Chapter 1: Introduction

Another significant benefit thought to be associated with homeownership is higher life satisfaction

and better psychological health. Owners are thought to have higher self-esteem both due to the higher

social status associated with homeownership as well as the sense of accomplishment that results from

having achieved a significant life goal. Owners are also thought to benefit from a feeling of greater

control over their life, derived from the fact that owners do not have to worry about being forced out

of their home by landlords’ actions. The wealth created through homeownership may also contribute

to this greater sense of control by providing a financial cushion that can be tapped to meet emergency

needs. Owners are also thought to have better physical health, perhaps in part as a result of their

better psychological health and in part due to the better quality of their homes. Of course, to the

extent that owners are financially stretched to meet the costs of homeownership, they may feel less

control over their lives and more vulnerable to financial and personal shocks. In these situations

owners may fare worse than renters in terms of both psychological and physical health.

Finally, an important social benefit of homeownership is better life outcomes for children that grow

up in owner-occupied homes. Homeownership is thought to benefit children by several mechanisms.

Homeownership may enable greater residential stability, which benefits children by providing a stable

social and educational environment. The more home-centered lifestyle associated with

homeownership may provide children with a more nurturing home environment. Given owners’

incentive to invest in their homes and the fact that owner-occupied housing is much more likely to be

in single-family detached housing, the greater quality, size, and privacy of these homes may also help

support children’s development. Finally, to the extent that homeownership helps to foster wealth

creation, owners will have more financial resources available to invest in their children’s education

and health care and to generally provide a supportive environment for their development. A wide

range of better outcomes in children have been attributed to homeownership, including higher

educational attainment, less delinquency, lower rates of teenage pregnancy, and higher rates of

subsequent homeownership. On the flip side, there may also be reason to be concerned about efforts

that succeed in increasing low-income homeownership by having these households buy into

distressed neighborhoods. In these situations, the benefits of homeownership may be offset by having

children locked into these distressed communities.

Process for Realizing the Benefits of Homeownership

The potential benefits of homeownership outlined above are by no means guaranteed. Whether these

benefits are actually realized depends on a broad range of factors, including:

• When (age and timing) household heads first become homeowners;

• Where they chose to buy;

• How much the household spends on housing;

• The condition and age of the home they buy;

• How much they reinvest in maintaining and improving their homes;

• The mortgage products they can qualify for, have access to, and choose;

Chapter 1: Introduction

• If and when they refinance mortgages or tap into home equity;

• If income or budget shocks force them to default on their mortgage loans or house price

declines spur them to do so; and

• How often they move and their tenure choice and level of expenditures at each move.

Exhibit 1 presents a conceptualization of the determinants of homeownership outcomes, delineating

the key choices that affect outcomes as well as the types of events that affect these choices.

Importantly, many of the benefits of homeownership—such as the accumulation of wealth and

positive impacts on children or health—would only be expected to accrue over a long period of time.

One of the key insights from the process outlined in Exhibit 1 is that it is not the outcome of single

experiences with homeownership that matters but the timing of tenure and mortgage choices

throughout the life cycle. Thus, in evaluating whether an individual household benefits from

homeownership it is necessary to consider not just the outcome from the time spent in a single home,

but rather their cumulative experience in a sequence of homes. Few studies take this perspective,

probably because the number of paths that individuals can trace is so great and the sample sizes of

panel studies so small. As a result, most of the literature examines behavior and outcomes across

single episodes of homeownership, such as equity accumulation from purchase to sale of a home, or

examines cross-sectional behaviors and outcomes, such as who refinances during a refinance boom or

default and delinquency behavior in a single year. Nonetheless, the absence of a life cycle

perspective contributes to important gaps in the existing literature.

Of particular importance for this study, virtually all of the factors that contribute to the outcomes from

tenure choices are strongly influenced by a household’s income, race, and ethnicity. Lower average

incomes restrict the range of housing options available to homebuyers to only lower cost units, often

in lower quality neighborhoods. Segregation of residential space by income as well as race in turn

may influence the average house price appreciation experience of low-income and minority owners.

Research has consistently found significant geographic segmentation of mortgage markets by race

and income, suggesting that where an owner lives exerts an important influence on their access to

financial services and mortgage products. Low incomes also make it harder to save enough to be able

to buy a home, harder to save a cash cushion against budget and income shocks, and harder to cover

costs of maintenance and replacements. Lower income typically entails lower wage work and more

unstable employment, which tends to leave low-income households more prone to reductions in

income through job loss. Because Hispanics and blacks have lower levels of education on average

than whites and receive lower earnings on average for comparable levels of education, the problems

confronting low-income homebuyers and owners disproportionately affect minorities.

Chapter 1: Introduction

Exhibit 1

Conceptual Model of Lifetime Returns from Tenure Choices

Individual Housing Unit Spell

Sell/Move Default

Stay (inc. Remodel or Refinance)

Financial and Social Returns

from Individual Housing

Spell

Re-evaluate Housing Choice

Owner Choices

Property

•

Type, quality, value, location

Mortgage

•

Terms, rate, fees, underwriting

ratios

Renter Choices

Unit

•

Type, quality, price and location

Investments

•

Use of potential downpayment

and closing cost funds

Trigger Events for Renters

•

Income and wealth changes

•

Household composition changes

•

Neighborhood changes

•

Housing and mortgage market

changes

•

Housing related budget shocks

•

Non-housing budget shocks

Trigger Events for Owners

•

Income and wealth changes

•

Household composition changes

•

Neighborhood changes

•

Housing and mortgage market

changes

•

Housing related budget shocks

•

Non-housing budget sho cks

Re-evaluate Housing Cho ice

Financial and Social Returns

from Individual Housing

Spell

Supply Factors

Housing Options

•

Price, quality, and structure type by

location and tenure

•

Real estate broker practices

Mortgage Supply

•

Mortgage competition in area

•

Mortgage offeri ngs (price and

terms)

•

Mortgage underwriting (income

ratios, LTV, credit history)

Key Demand Factors

•

Race, Ethnicity & Income

Other Demand Factors

•

Wealth

•

Occupation/education

•

ge, family type

•

Past credit history

•

Family/peer housing histories

Move Stay

Tenure Choice

Number, Timing, Length, and Characteristics of All Housing Unit Spells

Financial Returns

Social Returns

•

Home equity accumulation

•

Housing and neighborhood

•

ccumulation of other financial

satisfaction

assets

•

Psychological and physical health

•

Cumulative housing costs

•

Impacts on children

•

Cumulative tax benefits

Chapter 1: Introduction

Finally, while the deductibility of mortgage interest on federal income taxes provides an incentive for

homeownership, lower income households derive fewer benefits from this provision both because

they have lower marginal tax rates and because their interest payments may be small relative to the

standard deduction reducing the chance that they will choose to itemize their deductions. Based on

estimates of the number of tax returns claiming the mortgage interest deduction, approximately 15

percent of homeowners with income under $30,000 claim this deduction, compared to 50 percent of

those with incomes between $30,000 and $50,000, and 64 percent of those with income over

$50,000.

Because Hispanics and blacks have sharply lower average wealth than whites of comparable incomes,

and because low-income households have sharply lower average wealth than higher income

households, the neighborhood and housing options of low-income individuals and minorities is

further restricted, their vulnerability to income and budget shocks greater, and the speed at which they

can achieve homeownership thereby slower.

Taken together, many of the systematic variations in income, wealth, location, and education related

to race, ethnicity, and income drive living arrangements and family choices, the number and timing of

moves, number and timing of tenure choices, mortgage choices, refinance behaviors, repair and

remodeling behaviors, and vulnerability to house price declines or housing payment increases,

income disruptions, and unforeseen but necessary non-housing expenditures. These variations give

rise to expected differences in the average experiences, risks, and returns to homeownership for low-

income and minority homeowners. Thus, the “odds” of different outcomes are expected to vary by

race, ethnicity, and income. The overarching goal of this study is to sort through available

information to evaluate how the different factors outlined in Exhibit 1 contribute to difference

homeownership experiences for low-income and minority homeowners.

Outline of the Report

As noted above, for a variety of reasons much of the literature examining the benefits of

homeownership does not take a life cycle view of housing choices, but rather focuses on a short-run

outcome—e.g., the appreciation in house values over the course a set period of time. In addition,

there is a variety of research that is not explicitly focused on examining the benefits of

homeownership but rather examines either specific housing choices, such as a decision to choose a

certain type of mortgage or undertake remodeling activities, or intermediate outcomes, such as the

choice of moving to a new home. But the process outlined in Exhibit 1 helps to place this research in

context in considering how specific housing choices and intermediate outcomes ultimately contribute

to the benefits realized by low-income homeownership.

With the process outlined in Exhibit 1 in mind, the literature reviewed by this report is organized as

follows. In Chapter 2, we begin by examining the housing choices made by first-time homebuyers.

These figures are derived from estimates of the number of tax returns by claiming the mortgage interest

deduction by filer income in 2004 as reported in “Estimates of Federal Tax Expenditures for Fiscal Years

2005-2009” prepared by the Joint Committee on Taxation, January 12, 2005, and the authors’ tabulations

of the number of homeowners in these income categories from the March 2004 Current Population Survey.

Chapter 1: Introduction

These choices have important implications for the likelihood that the long-run benefits of

homeownership will be realized. Chapter 3 then examines choices made after initial purchase,

including decisions about whether to move, remodel, refinance, or default. Chapters 4 and 5 then

examine literature that sheds light on whether low-income and minority homebuyers are likely to

realize the financial and social benefits of homeownership. The report concludes in Chapter 6 with a

discussion of policy implications of the reports findings and areas where further research is needed.

Related Studies

This report is part of a larger project to examine the experience of low-income and minority

homeowners. In addition to this report, four other studies were undertaken to analyze specific aspects

of the homeownership experience of these groups. The findings from these other studies are included

in the review presented here. They include:

• The Growth of Earnings of Low-income Households and the Sensitivity of Their

Homeownership Choices to Economic and Socio-Demographic Shocks, by Donald

Haurin and Stuart Rosenthal;

• Wealth Accumulation and Homeownership: Evidence for Low-Income Households, by

Thomas Boehm and Alan Schlottman;

• The Impact of House Price Appreciation on Portfolio Composition and Savings, by

Donald Haurin and Stuart Rosenthal; and

• Is Manufactured Housing a Good Alternative for Low-Income Families? Evidence from

the American Housing Survey, by Thomas Boehm and Alan Schlottman.

Full copies of these studies are available at http://www.huduser.org/publications/homeown.html.

Chapter 1: Introduction

Chapter Two:

Initial Housing Choices Made by Low-Income

Homebuyers

The chapter presents information on the initial housing choices of low-income and minority first-time

homebuyers. As described in the introduction, these characteristics are of interest because they

influence the extent to which the long-run financial and social benefits of homeownership are

realized. Of particular interest for this report are the millions of low-income and minority households

who bought their first home during the homeownership boom that began in the early 1990s. Because

there is not an extensive literature describing the housing choices of low-income first-time

homebuyers from this period, much of the information presented in this chapter is derived from

tabulations of the American Housing Surveys (AHS) from 1991 through 2003. The AHS, a national

survey conducted in every odd-numbered year, is a rich source of information on characteristics of

the U.S. housing stock and is one of the few sources of information on first-time homebuyers.

In order to place the housing choices of low-income and minority homebuyers in context, information

on the housing choices of several comparison groups is also presented. First, the housing choices of

white, moderate-, and high-income first-time homebuyers are used to examine the extent to which the

choices of low-income and minority buyers differ from these groups. Second, the housing choices of

recent-mover low-income renter households are also presented to see how the choices of homebuyers

differ from those of renters. Recent movers are used rather than all renters so that the choices reflect

the renters’ optimal housing choice subject to the constraints imposed by current market conditions.

As a final point of reference we also present information on all households.

Since the sample sizes for first-time homebuyers in specific income or racial-ethnic categories in any

one survey can be fairly small, survey results are generally combined for the seven survey years from

1991 through 2003 to provide more robust estimates of how the characteristics of first-time buyers

and their housing choices differ across the income and racial-ethnic groups of interest. But since

trends in first-time buyers over the course of the recent homeownership boom are of interest, we also

compare results for two time periods: those corresponding to the 1991 through 1995 survey years and

those from the 1997 through 2003 survey years.

Of course, there are important differences in the characteristics of the various comparison groups,

which will contribute to the differences in the housing choices made. The first section of this chapter

presents basic demographic information on these groups so that these differences can be borne in

mind when evaluating differences in housing choices. This section also presents information on

trends in the number and characteristics of first-time homebuyers since 1991.

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

In the remaining sections of the chapter, four main aspects of housing choices are discussed:

• Housing characteristics;

• Neighborhood characteristics;

• Housing costs; and

• Mortgage finance characteristics.

An assessment of housing characteristics is used to assess whether low-income homebuyers do, in

fact, benefit from larger and higher quality housing. In addition, housing characteristics will also

influence the cost and effort associated with maintaining the home. Finally, structural qualities may

influence the likelihood of future wealth accumulation. Manufactured housing, in particular, is of

special interest because of its important role in increasing low-income homeownership, especially in

the South, during the 1990s (Belsky and Duda, 2002a). But manufactured housing poses special

issues for two reasons. First, since about half of manufactured homes are placed on leased land,

owners of these units do not share in appreciation of land values and are subject to increased costs

passed on by owners of the land. Second, financing of these units is often more expensive than

conventional mortgage rates. Housing characteristics of interest include the housing type (e.g.,

single-family detached, manufactured home, or condominium in multifamily structure), age, size

relative to household size, and quality (e.g., number and type of housing problems).

A number of benefits associated with homeownership derive from neighborhood attributes, including

the quality of public services and surrounding properties. To provide some indication of whether

homeowners are more likely to live in higher quality neighborhoods, this chapter examines

information from the AHS on the location of the home within a metropolitan area, measures of

neighborhood quality, and the homeowner’s satisfaction with the neighborhood. There is also some

literature on the characteristics of neighborhoods where low-income buyers have located, which will

be reviewed.

Housing costs are of interest to determine whether the move to homeownership has placed an undue

financial burden on these new owners. The discussion of housing costs focuses on measures of

housing costs relative to household income. Finally, since mortgage finance choices have important

implications for housing costs (both initially and over time) and for buyers’ exposure to risks of

interest rate and house value fluctuations, we will also examine mortgage finance characteristics. An

important issue to consider in this context is subprime lending, which increases the costs of mortgage

finance and has been associated with predatory lending practices. The extensive literature that

examines this latter topic will also briefly be reviewed.

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

Trends in the Number and Characteristics of First-Time

Homebuyers

Exhibit 2 provides information on trends in the annual number of low-income and minority first-time

homebuyers by income as captured by the AHS surveys from 1991 to 2003.

The relatively small

sample sizes of some subgroups of first-time homebuyers results in fairly sizeable sampling variation

in the estimates, which may cloud information on trends in the number of buyers over time.

Nonetheless, the annual estimates provide some indication of trends over time. During the early

1990s the number of low-income first-time buyers rose from a little more than 500,000 per year to

more than 750,000 per year by 1995-1997, an increase of nearly 50 percent. These trends are

consistent with the sharp rise in low-income homeownership that occurred over this period. After

1997, the number of low-income homebuyers moderated somewhat, but remained above the levels

that prevailed during the first years of the 1990s.

Exhibit 2

Average Annual Number of Low-Income and Minority First Time Homebuyers

(Thousands of homebuyers)

Years Low-Income Black Hispanic

1989-1991 514 128 88

1991-1993 578 96 120

1993-1995 594 180 152

1995-1997 761 252 196

1997-1999 693 228 200

1999-2001 643 192 219

2001-2003 690 156 230

Note: The overlap in years reflects the fact that each AHS survey covers the two-year period prior to the survey.

Source: Tabulations of the 1991-2003 AHS.

The AHS is conducted every other year and provides information on current occupants of the surveyed

units, including whether they are first-time homebuyers and what year they obtained their home.

Responses to these questions make it possible to identify first-time homebuyers who purchased their homes

in the two-year period between surveys. The AHS does identify the year of purchase so annual estimates

are possible. But since the sample of first-time buyers is somewhat small for any single year, the number

of homebuyers captured by the survey is divided by two to yield an estimate of the annual average number

of first-time buyers to smooth out this sampling variation.

Of note, there was a change in the methodology used to assign the relevant area median income for each

household in the AHS. As a result, the trends in the number of low-income first-time buyers between

1999-2001 and 2001-2003 must be interpreted with caution. Trends between the last two survey years of

2001 and 2003 suggest there was a very sharp fall-off in high-income buyers, a more moderate decline in

moderate-income buyers, and a slight increase in the number of low-income buyers. These trends may be

related to the economic recession that occurred during the 2001-2003 period. But it seems likely that the

change in how the relevant area median incomes are assigned contributed to this trend.

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

The increase in minority first-time buyers was even more pronounced. Over the same period from

1989-1991 to 1995-1997, the number of black first-time buyers doubled, while the number of

Hispanic buyers rose by 123 percent. As with low-income buyers, the number of black first-time

buyers moderated after 1997 but still remained above levels from the start of the decade. In contrast,

the number of Hispanic homebuyers continued to grow through the 2001-2003 period.

Exhibit 3 presents summary information on the age, household type, and racial composition of first-

time buyers over the period from 1989 through 2003. In terms of age, in general, there is a fair

amount of similarity in the age profile of the three categories of buyers, with the single largest

category age 25 to 34 followed by age 35 to 44. But low-income buyers are more likely to be both

younger (under age 25) and older (age 45 or above) than either moderate- or high-income buyers.

These two age groups may represent two distinct categories of low-income buyers: the younger

buyers are more likely to only temporarily be categorized as low-income as their incomes will

increase with age, while the older buyers are more likely to be long-term low-income households that

have needed more time to accumulate the savings needed to purchase a home.

In general, the earlier

that a household becomes a homeowner, the greater chance they will have to reap the benefits of

homeownership. The fact that low-income first-time buyers are more likely to be older means they

will have less time to realize homeownership’s benefits. But the proportion of older households

among low-income buyers (16 percent) is not substantially greater than for moderate-income (9

percent) or high-income (8 percent) households.

Exhibit 3 also shows the age distribution of recent-mover low-income renters. In general, as with the

other demographic characteristics shown, low-income first-time buyers lie in between low-income

renters and higher income owners in terms of age. Low-income renters have higher shares of both

younger and older households than low-income owners, who in turn have higher shares of these

groups than higher income owners. The greater concentration of homebuyers in the 25-to-34-age

category is consistent with the view that households below age 25 have both greater expected

mobility and less demand for housing and therefore are less likely to pursue homeownership. But

low-income renters also have a higher share of households that are age 45 and older. These may be

households who simply prefer to rent or they may be households who cannot amass the savings

needed to purchase a suitable home.

There are more significant differences across the first-time buyer income categories by household

type than there are by age. Specifically, low-income first-time buyers include a much lower share of

married couple households and a much higher share of single-earner households than either moderate-

or high-income buyers. While married couples account for nearly two-thirds of moderate-income

homebuyers and three-quarters of high-income buyers, these households comprise only 42 percent of

Since the AHS surveys the same housing units each time, it can be used to give a sense of the degree to

which households move between income categories over time. Of the low-income first-time homebuyers

identified by the 1991 survey, 60 percent of those in the same housing unit at the time of the 1999 survey

were still categorized as low-income, while 18 percent were moderate-income and 22 percent were high-

income. So while a majority did not change their income category, there is nonetheless a fair amount of

upward mobility. At the same time, there is a similar amount of downward mobility. Of those we were

low-income in the 1999 survey, 66 percent were also low-income in 1991, while 20 percent started the

period as moderate-income, and 14 percent started as high-income.

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

low-income buyers. In contrast, single parents with children and single person households account

for 45 percent of low-income buyers, compared to only 11 percent of moderate income buyers and 9

percent of high-income buyers. The share of single person households among low-income buyers is

particularly large, at 29 percent, compared to only 4 percent of higher-income buyers.

Exhibit 3

Selected Demographic Characteristics of First Time Homebuyers, 1989-2003

Characteristic

First Time Homebuyers

Recent-Mover

Low-Income

Renters

All

Households

Low

Income

Moderate

Income

High

Income

Age of Head

25 or younger 18% 11% 6% 26% 5%

25 to 34 43% 56% 62% 35% 19%

35 to 44 23% 24% 24% 19% 23%

45 or older 16% 9% 8% 20% 53%

Household Type

Married, No Children 14% 26% 36% 9% 28%

Married with Children 28% 38% 39% 15% 24%

Single Parent with Children 16% 7% 4% 22% 9%

Single Person 29% 4% 4% 37% 25%

Other 12% 16% 8% 18% 13%

Race/Ethnicity

White 67% 75% 77% 59% 76%

Black 14% 10% 8% 20% 12%

Hispanic 14% 9% 8% 15% 8%

Other 5% 6% 6% 6% 4%

Note: Low-, moderate-, and high-income defined as income less than 80 percent of the area median income

(AMI), 80 to 119.9 percent of AMI, and 120 percent of AMI or higher, respectively.

Source: Tabulations of 1991-2003 American Housing Survey.

The high proportion of single earner households among low-income buyers is not surprising—it is to

be expected that households with single earners will have lower incomes than those with two. But it

also highlights an important challenge for this group—with only a single earner to rely on the

household will have less ability to respond to a crisis, such as the loss of a job or a health problem in

the family. These households also have fewer adults in the household to share the burden of

maintaining the home. In part for these reasons, these households are more likely to be found among

renter households. Among recent low-income renters, 59 percent were headed by a single adult,

while only 24 percent were headed by married couples.

In terms of race and ethnicity, low-income first-time homebuyers include a higher share of minorities

than the upper-income groups. Non-Hispanic whites account for about three quarters of both

moderate- and high-income buyers, compared to two thirds of low-income buyers. Blacks and

Hispanics each account for 14 percent of low-income buyers, compared to 10 percent or less of the

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

other two income groups. Minorities account for a greater share of low-income first-time buyers than

they do of all households, although they account for even higher shares of low-income recent-mover

renters.

At other points in this chapter we will make comparisons between the housing choices of low-income

first-time buyers and low-income recent renters. The demographic differences between these groups

evident in Exhibit 3—specifically, that renters are both younger and older, have fewer married couple

households, and have a higher share minority—will account for some of the differences in housing

choices made. While both groups have income levels below 80 percent of area median incomes,

renters also have lower incomes than owners. Across the period studied, the average income among

low-income recent mover renters is 38 percent of area median income, compared to an average of 49

percent among low-income first-time buyers. In short, low-income first-time buyers are not perfectly

comparable with low-income renters. Nonetheless, some of the differences in housing choices

between these groups does reflect differences in the housing choices available in rental and

homeowner markets.

Exhibit 4 presents further information on the characteristics of first-time homebuyers by race and

ethnicity. One notable difference between minorities and whites is that first-time minority buyers

tend to be older than whites. While only 30 percent of white buyers are age 35 or older, 52 percent of

blacks, 45 percent of Hispanics, and 48 percent of “Other” minorities are in these older age

categories. The fact that minorities enter homeownership at later ages than whites means that they

have less time to accumulate wealth and realize the other benefits of homeownership.

There are also notable differences in the distribution of household types by race-ethnicity. Compared

to whites, first-time black homebuyers are less likely to be married (45 versus 58 percent) and more

likely to be a single parent (41 versus 28 percent). Thus, black first-time buyers are less likely to

have two earners to support the household. In contrast, compared to whites, Hispanics and other

minorities are more likely to be married couples with children (52 and 46 percent, respectively, versus

31 percent) and less likely to be in single person households (9 and 10 percent, respectively, versus 21

percent). While these minorities groups are more likely to have two earners supporting the

household, they are also more likely to have children, which increases non-housing costs and may

make it more difficult to meet unexpected financial demands.

Finally, Exhibit 4 also presents information on the distribution of each racial-ethnic group by income.

Both blacks and Hispanics are more likely than whites to be low-income, with about half of first-time

buyers in this category compared to 37 percent of whites. Other minorities have a similar income

distribution to whites.

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

Exhibit 4

Selected Demographic Characteristics of First Time Homebuyers by Race-Ethnicity,

1989-2003

Characteristic

First Time Homebuyers

White Black Hispanic Other

Age of Head

25 or younger 13% 6% 11% 9%

25 to 34 56% 42% 44% 44%

35 to 44 20% 34% 30% 33%

45 or older 10% 18% 15% 15%

Household Type

Married, No Children 27% 14% 18% 23%

Married with Children 31% 31% 52% 46%

Single Parent with Children 8% 23% 11% 8%

Single Person 21% 18% 9% 10%

Other 13% 14% 9% 14%

Income Category

Low 37% 50% 52% 37%

Moderate 28% 25% 23% 27%

High 35% 25% 25% 36%

Note: Low-, moderate-, and high-income defined as income less than 80 percent of the area median income

(AMI), 80 to 119.9 percent of AMI, and 120 percent of AMI or higher, respectively.

Source: Tabulations of 1991-2003 American Housing Survey.

Exhibit 5 shows trends in the characteristics of low-income first-time buyers before and after 1995 to

see to what extent the increase in homeownership rates over this period was associated with changes

in the characteristics of first-time buyers.

There are two notable trends in the data shown in Exhibit

5. First, there has been a decrease in the share of married couple households and a concomitant

increase in the share of single adults, either with or without children. In the period 1989 to 1995, 50

percent of low-income homebuyers were married couples while 38 percent were headed by single

adults. By 1995 to 2003 these shares had essentially reversed, with 38 percent headed by married

couples and 49 percent headed by single adults. While moderate- and high-income buyers also

experienced an increase in the share of single adult households, the rise among these groups was only

3 to 4 percentage points. Thus, it is true that many more low-income first-time buyers consisted of

households headed by a single adult.

Grouping the AHS survey years together provides increases the sample of low-income first-time buyers to

provide a more accurate depiction of trends.

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

Exhibit 5

Trends in Selected Demographic Characteristics of Low-Income First Time

Homebuyers, 1989-2003

Characteristic 1989-1995 1995-2003

Age of Head

25 or younger 17% 18%

25 to 34 46% 42%

35 to 44 22% 23%

45 or older 15% 17%

Household Type

Married, No Children 16% 13%

Married with Children 34% 25%

Single Parent with Children 14% 17%

Single Person 25% 32%

Other 11% 13%

Race/Ethnicity

White 71% 64%

Black 13% 14%

Hispanic 11% 15%

Other 5% 6%

Note: Low income defined as income less than 80 percent of area median income.

Source: Tabulations of 1991-2003 American Housing Survey.

A second notable trend was for a higher share of minorities among low-income first-time buyers. In

1989 to 1995, non-Hispanic whites accounted for 71 percent of these buyers, but since 1995 this share

had declined to 64 percent. Much of this increase in the minority share was due to a higher share of

Hispanics among low-income first-time buyers, which increased from 11 percent in 1989 to 1995 to

15 percent by 1995 to 2003.

Housing Choices of Low-Income Buyers

Exhibit 6 presents summary information on the housing units purchased by first-time homebuyers by

income and racial-ethnic categories during the period from 1989 to 2003. There is relatively little

difference in the choice of structure type by race-ethnicity, although blacks are slightly more likely to

live in single-family attached units and Hispanics are slightly less likely to live in manufactured

housing. There are more significant differences evident by income. Compared to both moderate- and

high-income buyers, low-income households are less likely to purchase single-family detached homes

and more likely to purchase manufactured homes. This is in keeping with the findings of Belsky and

Duda (2002a), who found that manufactured housing played an important role in the boom in low-

and moderate-income homeownership during the 1990s. Among low-income buyers, manufactured

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers Abt Associates Inc.

Exhibit 6

Selected Housing Characteristics of First Time Buyers by Income Category 1989-2003

Characteristic

First Time Homebuyers

Recent-Mover

Low-Income

Renters

All

Households

Low

Income

Moderate

Income

High

Income White Black Hispanic

Structure Type

Single Family Detached 61.3% 73.6% 81.3% 71.8% 68.1% 73.9% 19.5% 62.7%

Single Family Attached 7.4% 8.0% 8.0% 7.4% 12.1% 7.0% 7.9% 6.1%

Multifamily 7.5% 7.4% 7.2% 7.8% 6.7% 8.5% 68.1% 24.8%

Manufactured Home 23.8% 11.0% 3.5% 13.0% 13.0% 10.6% 4.5% 6.4%

Median Square Feet per Occupant

545 570 652 642 527 389 434 662

Units Built in 1970 or Earlier

49.7% 47.4% 40.2% 46.4% 45.3% 49.9% 53.3% 53.4%

Housing Adequacy

Moderately Inadequate 4.8% 2.7% 2.3% 2.9% 4.5% 6.3% 7.9% 4.6%

Severely Inadequate 2.0% 1.2% 1.1% 1.4% 2.0% 1.5% 3.0% 2.1%

Housing Satisfaction*

Average 8.1 8.3 8.4 8.3 8.6 8.4 7.4 8.1

Share rated 5 or lower 8.7% 4.6% 3.1% 5.6% 5.0% 6.3% 17.9% 9.3%

Note: Housing satisfaction is rated on a 10-point scale with 10 being the best and 1 the worst.

Source: Tabulations of the 1991-2003 American Housing Surveys.

homes accounted for 23.8 percent of homes purchased, compared to 11.0 percent among moderate-

income buyers and 3.5 percent among high-income buyers. One recent study found that low-income

owner’s satisfaction with the quality of manufactured homes is only slightly lower than owners of

traditional homes. Since these homes have much lower costs, the authors conclude that manufactured

housing represents a good value for low-income buyers (Boehm and Schlottmann, 2004a). However,

the study also notes that the fact that a large share of these homes are on leased land greatly limits the

potential for wealth accumulation from these types of units—an issue that will be explored more in

Chapter 4.

As noted in the introduction, there is a substantial difference in the types of housing units occupied by

first-time homebuyers and renters. Low-income renters are nine times as likely to live in multifamily

structures and a third as likely to live in single-family detached housing compared to low-income

buyers. While some portion of these differences are undoubtedly related to differences in the desired

quantity of housing between these groups, the differences are great enough that some amount is likely

to reflect the different opportunities available in the rental and owner-occupied housing markets.

Low-income owners clearly are able to obtain a much greater amount of privacy than renters.

In terms of the amount of living space available per resident, low-income first-time buyers do have

less space than their higher-income counterparts. The median square feet per occupant for low-

income buyers is 549. While this is only slightly lower than the 560 square feet for moderate-income

buyers, it is substantially less than the 653 square feet available to high-income buyers. Nonetheless,

low-income buyers have 26 percent more living space than recent low-income renters, who have only

439 square feet per occupant.

There are larger differences in the amount of living space per resident by race-ethnicity. While on

average white buyers have 642 square feet per occupant, blacks have only 527 square feet and

Hispanics have only 389. While black homebuyers still have much more space on average than low-

income renters, Hispanics actually have less space per occupant than renters. However, the small

amount of space per occupant among Hispanics primarily reflects the larger household sizes among

Hispanic owners. The families of Hispanic buyers average 3.7 persons, while white owner families

average 2.5 and blacks average 3.1. The homes purchased by Hispanics are also about 10 percent

smaller on average than white homes, but it is the larger household sizes that lower the space per

occupant so much. Still, Hispanic renters only average 313 square feet per occupant, so

homeownership is associated with an increase in living space for Hispanics.

One of the concerns cited about the emphasis on low-income homeownership is that too many buyers

are purchasing inadequate housing, which increases housing costs, raises the risk of being subject to

financial shocks from unexpected housing problems, and reduces the quality of the living

environment enjoyed by residents. Exhibit 6 presents information on the share of buyers purchasing

older housing, which might be expected to need more maintenance and generally be of lower quality

due to the age of the house. In terms of housing age, low-income buyers are found to be more likely

to purchase homes that were built in 1970 or earlier, with 49.7 percent in homes of this age, compared

to 47.4 percent of moderate-income buyers and 40.2 percent of high-income buyers. There is less

variation in housing age by race-ethnicity. Hispanics have the highest share of older housing at 49.9

percent, compared with 46.4 percent of white first-time buyers in older housing and 45.3 percent of

blacks. However, the share of all households living in these older housing units is higher still at 53.4

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

percent, which is essentially the same as the share of recent-mover low-income renters in older units.

Thus, regardless of income or race-ethnicity homebuyers tend to occupy somewhat newer units than

either all households or renters.

A more direct measure of housing quality is provided by AHS variables indicating whether a unit is

moderately or severely structurally inadequate. It is true that low-income first-time buyers are more

likely to live in moderately or severely inadequate units, with an inadequacy rate that is 75 percent

higher than for moderate-income buyers and roughly twice the rate among high-income buyers.

Nonetheless, the share of low-income buyers in moderately or severely inadequate housing is fairly

low as 4.8 percent live in moderately inadequate housing and 2.0 percent live in severely inadequate

conditions. Minority homebuyers are more likely to live in inadequate housing than whites, with 4.5

percent of blacks and 6.3 percent of Hispanics in moderately inadequate housing, compared to 2.9

percent of whites. With the exception of Hispanics, these rates are either better than or about the

same as the share of all households living in inadequate housing, which suggests that low-income and

minority buyers are no worse off than other households. In addition, the level of structural

inadequacy is higher among recent-mover low-income renters, with 7.9 percent living in moderately

inadequate and 3.0 percent in severely inadequate housing.

A similar pattern is evident with regard to housing satisfaction. As a measure of satisfaction, the

AHS asks each respondent to rate their home as a place to live on a 10-point scale, with one being

best and 10 being worst. Exhibit 6 shows both the average satisfaction rating and the share of

households reporting a level of satisfaction of 5 or lower. Low-income buyers are found to have

slightly lower average satisfaction ratings than moderate- or high-income buyers, but have similar

levels of satisfaction compared to all households and higher levels of satisfaction compared to recent-

mover low-income renters. In terms of the share with low satisfaction ratings, compared to moderate-

and high-income buyers low-income buyers are two to three times as likely to rate their satisfaction

level as 5 or lower. But the overall share of low-income buyers with low satisfaction is fairly small

(8.7 percent) compared to the share of either all households (9.3 percent) or recent-mover low-income

renters (17.9 percent) rating their housing this low. There is less difference in housing satisfaction by

race-ethnicity, with blacks and Hispanics actually having higher average satisfaction levels than

whites, and similar shares of households rating their housing 5 or lower across these three groups.

There is little evidence of any worsening of the quality of housing purchased by low-income buyers

over the last decade. In terms of structural adequacy, among low-income buyers the share of units

that were either moderately or severely inadequate actually declined from 8.1 percent to 6.2 percent

between the 1989-1995 and 1995-2003 periods. Over the same time, the share of inadequate units

among recent-mover low-income renters increased from 10.1 to 11.6 percent. There was a slight

decline in low-income buyers satisfaction with their homes, but the changes were fairly small. The

average rating among low-income first-time buyers dropped from 8.3 to 8.1, while the share of low-

income buyers reporting a satisfaction rating of 5 or less rose from 8.4 to 8.9 percent. There were

similar changes in satisfaction among recent-mover low-income renters as well.

An obvious deficiency of these tabulations of the AHS data is that they do not account for all of the

differences in household characteristics between the groups being compared. Unfortunately, there is

a very limited literature that employs multivariate analysis to examine housing outcomes of low-

income or minority homebuyers. Of the studies that exist, several examine the issue of how

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

homeownership affects housing quality. The most recent of these studies is Friedman and

Rosenbaum (2004), which uses the 2001 AHS to evaluate whether immigrants and racial-ethnic

minorities who achieve homeownership are more likely to experience housing crowding or live in

inadequate housing than whites. The study includes household income as an independent variable,

and finds that increases in income reduce the probability of experiencing these problems, but the

study does not present any estimates of the magnitude of differences between low- and upper-income

households. With regard to race-ethnicity, they find that blacks and Hispanics are more likely to

experience both crowding and inadequate housing than whites regardless of tenure and so conclude

that a move to homeownership does not eliminate these problems for minorities. However, while

blacks and Hispanic owners are worse off in these dimensions compared to white owners, the study

does not examine the question of whether a move to homeownership reduces the likelihood of

minorities experiencing these problems. However, the descriptive statistics presented in the report

suggest that this is the case.

An earlier study that examines a similar set of questions is Rosenbaum (1996). Rosenbaum estimates

a statistical model to predict the likelihood that a housing unit is structurally inadequate or has

abandoned buildings nearby based on the race-ethnicity and socioeconomic status of the occupant,

including whether they own or rent the unit. The analysis relies on data for the New York area from

both the AHS and New York City’s Housing Vacancy Survey. The analysis finds that minorities and

lower-income households are more likely to experience both of these problems. However, one of the

model’s strongest results is that, all else equal, owners are less likely to experience these problems.

However, since the study does not interact either race-ethnicity or income with tenure it does not shed

light on whether an owner’s lower likelihood of experiencing these problems varies by either race-

ethnicity or income.

While recent-mover low-income renter households are shown in the exhibits presented in this section

to provide an indication of whether a move to homeownership improves housing conditions for low-

income homebuyers, since we do not control for the many differences between these groups it is not

clear if this is a fair comparison. There have been a few studies that have examined the factors

associated with housing satisfaction controlling for differences in housing and household

characteristics. These studies consistently find that homeownership increases housing satisfaction

even after controlling for these other factors (Kinsey and Lane, 1983; Lam, 1985; and Danes and

Morris, 1986). While these studies do include income as an explanatory variable, they do not attempt

to evaluate whether the impact of homeownership on housing satisfaction varies with income. One

study (Kinsey and Lane) does have an explicit focus on differences between whites and blacks in the

factors explaining housing satisfaction. This study finds that homeownership is associated with

greater increases in housing satisfaction for blacks.

Finally, one recent study provides some insight into the question of how housing consumption

changes when low-income households become homeowners. Cummings, DiPasquale, and Kahn

(2002) examine the pre- and post-move housing characteristics of participants in homeownership

programs run by the City of Philadelphia. The main focus of the paper is a program that was

designed to promote neighborhood revitalization by constructing deeply subsidized owner-occupied

housing units in severely distressed neighborhoods. Since the program provided homeowners with

per unit subsidies in the range of $50,000 to $100,000 it is not surprising that this group experienced

significant increases in housing quality after moving. But the study also found that participants in a

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

program that provided a small subsidy ($1,000) to low-income buyers in the City of Philadelphia also

experienced significant improvements in housing quality. The new units were larger, more likely to

have a garage and to be in single-family structures. Overall, 75 percent of survey respondents

reported that the new home was better than their previous one. Thus, this study provides limited

evidence that a move to homeownership is often associated with an improvement in housing quality.

Neighborhood Characteristics

Exhibit 7 summarizes the information available from the AHS on the neighborhood choices of first-

time homebuyers. The top portion of the exhibit provides information on the prevalence of

neighborhood conditions that are indicators of blight, a lack of public services, or property uses that

are less well-suited to residential areas.

In general, low-income and, to a greater extent, minority

first-time buyers do experience worse neighborhood conditions that higher-income buyers. However,

the incidence of most of these conditions is somewhat rare. Low-income buyers are more likely to

have abandoned or vandalized properties nearby and to have trash or junk on the street, but in both

cases less than three percent of buyers experience these conditions. Blacks are more likely to have

abandoned or vandalized properties nearby, with 5.7 percent exposed to this condition. Bars on

windows, an indicator of greater potential for theft, are evident in 6.4 percent of low-income buyers’

neighborhoods, compared to about 4 percent of moderate- and high-income buyers’ neighborhoods.

This condition is much more common among minorities, with 11.3 percent of blacks and 15.7 percent

of Hispanics exposed to these conditions, compared to only 2.4 percent of whites.

The most common issue in low-income buyers’ neighborhoods is the presence of commercial or

industrial properties. These non-residential property uses are evident in about one in five cases for

low-income and black buyers and nearly one in four of Hispanic buyers. But these mixed-use

neighborhoods are also fairly common in the neighborhoods where white (15.4 percent), moderate-

income (16.8 percent) and high-income (14.6 percent) buyers are located. Again, low-income buyers

fare better in all of the dimensions compared to low-income recent-mover renters and have shares that

are fairly similar to those experienced by all households.

Similar to the question on housing satisfaction, the AHS also asks respondents to rate their

neighborhood on a scale of 1 to 10, with 10 being best and 1 the worst. Exhibit 7 shows the average

neighborhood rating and the share of households reporting a neighborhood rating of 5 or lower. In

terms of average ratings, there is very little difference across the first-time buyer groups by either

income or race-ethnicity, ranging only from a low of 8.0 on a 10-point scale among low-income

buyers to a high of 8.2 among moderate- and high-income and black buyers. However, the average

neighborhood rating does mask some variation evident in the share of households rating their

neighborhood at 5 or lower. Among low-income buyers, 11.7 percent rated their neighborhood 5 or

lower, compared to 7.7 percent of moderate-income and 6.2 percent of high-income buyers.

Minorities also are more likely to give a low rating to their neighborhoods, with 9.6 percent of blacks

These neighborhood characteristics are recorded by the field staff implementing the AHS. The questions

ask whether the indicated characteristic is evident within 300 feet of the subject property.

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

Abt Associates Inc. Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

Exhibit 7

Selected Neighborhood Characteristics of First Time Buyers by Income Category 1989-2003

Characteristic

First Time Homebuyers

Recent-Mover

Low-Income

Renters

All

Households

Low

Income

Moderate

Income

High

Income White Black Hispanic

Neighborhood Blight Within 300 Feet

Abandoned/Vandalized Properties 2.6% 1.5% 1.0% 1.3% 5.7% 1.7% 3.9% 1.8%

Bars on Windows 6.4% 4.1% 4.2% 2.4% 11.3% 15.7% 9.7% 5.9%

Trash/Junk on Street 2.5% 1.2% 1.2% 1.4% 2.3% 2.9% 4.1% 2.1%

Commercial/Industrial Properties 20.2% 16.8% 14.6% 15.4% 21.9% 24.6% 37.7% 20.1%

Neighborhood Satisfaction*

Average 8.0 8.2 8.2 8.1 8.2 8.1 7.3 8.0

Share rated 5 or lower 11.7% 7.7% 6.2% 8.5% 9.6% 10.3% 21.6% 12.3%

Metropolitan Location

Central City 30.4% 27.1% 26.7% 24.5% 39.5% 40.6% 46.6% 30.6%

Suburb 45.9% 54.7% 55.6% 52.5% 43.8% 48.7% 36.6% 47.3%

Non-metropolitan 23.8% 18.2% 17.7% 23.1% 16.6% 10.7% 16.7% 22.1%

Note: Neighborhood satisfaction is rated on a 10-point scale with 10 being the best and 1 the worst.

Source: Tabulations of the 1991-2003 American Housing Surveys.

and 10.3 percent of Hispanics giving a rating of five or lower, compared to 8.5 percent of whites. But

once again, all buyer groups compare favorably with low-income recent-mover renters, who on

average only rate their neighborhoods at 7.3 and 21.6 percent rate the neighborhood at 5 or lower.

Even compared to all households, recent buyers fare well, as the average across all households is a

rating of 8.0 and the 12.3 percent rate their neighborhood at 5 or lower.

Finally, Exhibit 7 also compares the distribution of these households between central cities, suburbs,

and non-metropolitan areas. While there is great variation in neighborhood quality within each of

these geographic categories, in general, neighborhoods in central cities are thought to be more likely

to have lower quality public services and more land uses that are less well-suited for residential areas.

Central cities also tend to have lower homeownership rates than suburban areas, and so owners in

these areas may be less likely to realize benefits from higher concentrations of owner-occupants. As

shown, low-income buyers are less likely to live in suburban areas than either moderate- or high-

income buyers (46 percent versus 55 to 56 percent), but this difference is split between a greater

propensity to live in both central cities and non-metropolitan areas. There is little difference between

the geographic location of low-income buyers and all households. In contrast, low-income renters are

much more likely than low-income buyers to live in central cities, with 47 percent in cities, but only

30 percent of low-income buyers in these areas. Both blacks and Hispanics are much more likely to

buy in central cities than whites and less likely to buy in non-metropolitan areas. Nonetheless, the

suburbs are still the most common destination for first-time black and Hispanic homebuyers, with 44

percent of blacks and 49 percent of Hispanics choosing to buy in these areas.

There are a small number of studies that have used Home Mortgage Disclosure Act (HMDA) data to

identify the characteristics of neighborhoods where low-income and minority homebuyers are

purchasing homes. It is not possible to identify first-time homebuyers from the HMDA data, but

because these data identify the census tract where homes were purchased it provides more precise

information on the neighborhood choices of homebuyers than other data sources. These studies shed

light on the extent to which low-income and minority buyers are gaining access through

homeownership to higher income neighborhoods and whether the location choices of minorities are

helping to reduce racial segregation.

Stuart (2000) examined home purchases in the Boston metropolitan area from 1993 through 1998 and

observes that while a significant share of blacks and Hispanics did purchase homes outside of the city

of Boston, these minorities were still much more likely to purchase in the central city. While 91

percent of whites bought in suburban areas, only 41 percent of blacks and 61 percent of Hispanics did

so. Importantly, half of the blacks and Hispanics who moved to the suburbs were found in just seven

communities. While the reasons for such constrained choices are not clear—that is, whether it

reflects discriminatory treatment, limits due to housing affordability, or preferences for specific

communities—the result may be the recreation of racially segregated living patterns in suburban

areas. In considering the location choices of low-income buyers, Stuart found that while low-income

buyers were distributed across communities of all income levels, they were more likely to purchase in

low-income communities (60 percent) than middle- (47 percent) or upper-income (34 percent) buyers.

Furthermore, he found that in suburban areas low-income whites were as segregated from upper-

income whites as blacks were from whites.

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

Immergluck (1999) also uses HMDA data to examine home purchase patterns by blacks in the

Chicago area. He also finds that black homebuyers were concentrated in a relatively small number of

census tracts. In the 1995 to 96 period, 45 percent of black homebuyers located in areas that were 75

percent black or more and 50 percent of all black homebuyers were concentrated in 5 percent of all

census tracts. Thus, like Stuart, he finds that black homebuying choices seem to reinforce patterns of

racial segregation. Immergluck and Smith (2001) also use HMDA data to examine patterns of home

purchase by different income groups in Chicago. They find that there was significant growth in

homebuying activity by low-income households in suburban areas of Chicago between 1993-1994

and 1999-2000. While these suburban buyers were mostly concentrated in older suburbs near the

core and outlying suburbs, there was nonetheless a strong movement of low-income buyers to

suburban areas. At the same time, the number of upper-income homebuyers increased rapidly in the

City of Chicago, but, again, concentrated in a few specific neighborhoods. Nonetheless, Immergluck

and Smith do find that there was some evidence of greater income mixing by homebuyers in the

Chicago during the 1990s.

Finally, Belsky and Duda (2002a) also use HMDA data for the period 1993 to 1999 to examine home

purchase activity by low-income and minority households in nine metropolitan areas. They also find

that large shares of low-income and minority homebuyers are purchasing in the suburbs. Significant

shares of low-income buyers were found to have purchased homes in moderate-income areas, leading

the authors to conclude that homebuying activity was contributing to some income mixing, although

there was a tendency for these households to be concentrated closer to the urban core than upper-

income households. Black purchases were also more clustered near the urban core and tended to be

concentrated in predominantly minority areas, leading the authors to conclude that homebuying by

blacks was not contributing materially to lowering levels of racial segregation.

In short, studies examining home purchase activity using HMDA data come to mixed conclusions

regarding home purchases by low-income and minority households. While buyers are gaining access

to suburban areas, there is a tendency for these buyers to locate in areas with greater concentrations of

low-income and/or minority households. In short, as Belsky and Duda conclude, “whether the move

to low-income homeownership has been associated with a move to opportunity remains an open

question” (Belsky and Duda, 2002a, page 52).

Another study that shed some light on the types of neighborhoods where minorities are buying homes

is Herbert and Kaul (2005), who use decennial census data at the census tract level for 1990 and 2000

to examine the characteristics of neighborhoods where minority homeownership rates increased the

most during the 1990s. This study reaches similar conclusions as those using HMDA data. In

general they find that areas with the greatest gains in minority homeownership rates were more likely

to be in suburban areas and were marked by higher incomes and house values and lower

concentrations of minorities than areas where there was little change in minority homeownership

rates. These findings suggest that the movement to homeownership is associated with a move to

areas of higher socioeconomic status and is supportive of greater racial integration. Still, the findings

also indicate that minorities live in areas with lower incomes and house values and higher minority

concentrations than the areas where whites live.

However, while cross-sectional comparisons may show that on average low-income and minority

buyers reside in better neighborhoods than low-income renters, this does not mean that individual

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

buyers actually improved their neighborhood conditions as a result of their move to homeownership.

It may be that among low-income households, those who achieve homeownership already resided in

somewhat better neighborhoods than other low-income renters. A more informative way to evaluate

whether a move to homeownership is associated with an improvement in neighborhood conditions is

to compare the characteristics of neighborhoods where low-income buyers lived prior to buying their

home to the area where they purchased. Several recent studies provide results from this type of

analysis.

Reid (2004) analyzes data from the Panel Study of Income Dynamics (PSID) covering a period from

1976 to 1993, using a special version of these data that includes characteristics from the decennial

censuses for 1980 and 1990 for the census tracts where respondents reside. The panel nature of the

PSID allows her to identify when renters become homeowners and to then compare the characteristics

of the neighborhoods where they lived before and after purchasing a home. The characteristics

examined include those related to demographics, economic status, and housing market conditions.

Reid groups buyers into three income groups (low, moderate, and high)

and two racial groups (non-

Hispanic white and all minorities). Reid concludes that the move to homeownership results in

essentially no change in neighborhood conditions for low-income whites, but fairly sizeable

improvements for low-income minorities. There are also small positive changes for moderate- and

high-income whites and minorities. For all groups except low-income whites, the move to

homeownership does result in an increase in the neighborhood homeownership rate. Low-income

minorities also experience declines in the shares of female-headed households, people in poverty,

households with welfare income, and unemployed adults.

Tempering the positive finding that minorities of all income levels experience some improvement in

neighborhood conditions when buying a home is the fact that compared to whites of the same income

category, minorities live in areas with lower economic status, fewer homeowners, and lower property

values. Thus, while a move to homeownership improves neighborhood conditions for minorities, it

by no means results in the same level of economic status as whites of similar income levels.

Another recent study that examines the before and after-purchase neighborhoods of low-income

homebuyers is Turnham et al. (2004). This study gathered data on 788 low-income homebuyers

assisted through the HOME program in 33 jurisdictions around the country during the period from

1993 to 2003. All of the homebuyers assisted through the HOME program have low incomes, with

74 percent of participants having between 50 and 80 percent of area median income. With a 55-

percent share, minorities account for a higher share of program participants than they do of all low-

income buyers.

Reid’s income classification is somewhat unique. The low-income category includes renters whose income

is less than 80 percent of the area median income in every year they are observed up through the time they

purchase a home. Moderate-income renters are those whose income exceeds the 80 percent threshold in at

least one year through the time when they purchase the home, but whose income is not consistently above

the area median income. High-income renters have income that exceeds the area median income every

year they are observed through the time they purchase their home.

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

The study found that a large majority of buyers (70 percent) moved at least one mile from their

previous residence and so were likely to have changed neighborhoods. Of these, 47 percent moved

between one and five miles, and 24 percent moved more than five miles.

This study did find some indications of increases in the housing status of the post-move

neighborhoods. Homeownership rates were slightly higher (58 versus 54 percent), as were the share

of housing in single-family units (52 versus 48 percent). But by a variety of other measures of

housing conditions, including age, vacancy rates, and values, there was essentially no difference.

Similarly, the pre- and post-move neighborhoods were remarkably similar in a variety of economic

and demographic characteristics, including poverty rates, share of households receiving public

assistance, household incomes, and share of adults with some college.

Turnham et al. also compare the characteristics of the neighborhoods to the broader jurisdiction

(either city or county) where they are located. In general, neighborhoods where low-income buyers

purchased are somewhat below average on a number of socioeconomic indicators. For example, the

neighborhoods have lower household incomes, lower house values, and lower education levels than

the broader jurisdictions. However, the neighborhoods are by no means distressed. The authors also

point out the average incomes in the neighborhood are much higher than the average income of the

HOME-assisted buyers. While the average buyer’s income was about $29,000, the average

neighborhood income was $42,000. The study concludes that while the move to homeownership did

not result in improved neighborhood conditions, it is also the case that the neighborhoods were

generally decent places to live, with moderate-income levels, a high share of working families and

little welfare dependence, and racially diverse.

A similar type of analysis was conducted by Turnham et al. (2003) on a small sample (84) of

homebuyers using housing vouchers in 12 markets around the country and found very similar results.

The profile of families assisted through the housing voucher program is similar to those assisted by

HOME. The typical buyer had income of less than $35,000, half of the participants were minority,

and most were single-parent households. As with the study of the HOME program participants, most

buyers (61 percent) were found to have moved at least one mile from their previous residence, with

21 percent moving 5 miles or more. However, half of the buyers who did not move more than a mile

purchased the same unit they had rented—and so experienced no change in either housing or

neighborhood as a result of the purchase. For the most part, neighborhoods where they moved were

similar to where they started, with only slight improvement evident in various socioeconomic

indicators. There was a slight increase in neighborhood homeownership rates (60 versus 57 percent)

and in the share of homes in single-family structures (54 versus 51 percent). Poverty rates were also

slightly lower (16 versus 18 percent) as was the share of single female-headed households (10 versus

11 percent).

The study also conducted a windshield assessment of 32 of the properties and their surrounding

neighborhoods. For the most part, the houses purchased appeared to be in better shape than

surrounding properties, exhibiting better exterior condition of the structure and surrounding grounds.

However, the differences were not large. For example, all of the purchased units were deemed to

have good or excellent outside housekeeping evident, but 90 percent of surrounding properties were

similarly rated. Overall, a majority of the neighborhoods where buyers had purchased were rated as

excellent (38 percent) or good (47 percent). In short, as with the study of the HOME program,

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

participants in the voucher homeownership program were not found to have experienced a significant

improvement in neighborhood conditions, but the areas where they bought were generally stable,

good quality neighborhoods.

Finally, Cummings, DiPasquale, and Kahn (2002) examine the pre- and post-move neighborhood

characteristics of participants in homeownership programs run by the City of Philadelphia. The main

focus of the paper is a program that was designed to promote neighborhood revitalization by

constructing deeply subsidized owner-occupied housing units in severely distressed neighborhoods.

Not surprisingly, the study found that this group experienced significant declines in neighborhood

quality after moving. But the study also reports on the pre- and post-move neighborhoods of

participants in a program that provided a small subsidy ($1,000) to low-income buyers in the City of

Philadelphia. The authors find that participants in this program experienced significant improvements

in neighborhood characteristics in a number of dimensions, including household income, house

values, and homeownership rates.

Taken as a whole, the literature that has examined the neighborhood choices of low-income and

minority homebuyers paints a somewhat mixed picture. For the most part, there is little evidence that

a move to homeownership by low-income households is associated with significant improvements in

neighborhood conditions. Although nor is there evidence that low-income homebuyers are being

relegated to distressed neighborhoods. For the most part, the areas with higher concentrations of low-

income buyers are suburban areas with moderate incomes. On the other hand, there are some

indications that minority homebuyers may fare better, with the national analysis by Reid and the

study of a Philadelphia homeownership program by Cummings, DiPasquale, and Kahn both finding

that minorities realized much more substantial neighborhood improvements with a move to

homeownership. But the downside of this finding is that even with these improvements, the

socioeconomic status of neighborhoods where minority owners are locating is lower than for whites

of comparable incomes.

Perhaps the most important concerns about the neighborhood choices of low-income and minority

buyers is what implications these choices have for the likelihood of realizing the financial and social

benefits associated with homeownership. These issues will be explored in depth in Chapters 4 and 5.

Housing Costs

Exhibit 8 presents the distribution of housing cost burdens across first-time buyers and other

household types. Housing cost burdens measure the share of income devoted to housing, including

rent or mortgage payments, utilities, property insurance, and property taxes. Traditionally, housing is

considered affordable if it accounts for less than 30 percent of income. Housing cost burdens of

between 30 and 50 percent are considered moderate, while those of 50 percent or more are severe.

Exhibit 8 further breaks down those with moderate cost burdens further into those that pay between

30 and 39 percent of income for housing and those that pay between 40 and 49 percent. Housing cost

burdens are shown for the first half of the 1990s and for the 1997 to 2003 period to identify trends in

cost burdens between these periods.

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

Abt Associates Inc. Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

Exhibit 8

Trends in Housing Cost Burden for First Time Buyers and Other Households

(Share of Households Spending Given Percent of Income on Housing)

Time Period/Housing

Burden Category

First Time Homebuyers

Recent Mover

Low-Income

Renters

All

Households

Low

Income

Moderate

Income

High

Income

White Black Hispanic

1989-1995

Less than 30% 52.7% 82.9% 94.2% 76.7% 71.0% 63.5% 42.4% 72.5%

30-39.9% 20.9% 12.9% 5.4% 13.8% 9.3% 15.1% 19.7% 11.1%

40-49.9% 11.9% 3.4% 0.4% 5.1% 8.4% 9.4% 11.1% 5.3%

50% or higher 14.5% 0.7% 0.0% 4.4% 11.2% 11.9% 26.7% 11.1%

1995-2003

Less than 30% 46.3% 78.4% 93.1% 74.1% 68.4% 54.3% 42.0% 71.5%

30-39.9% 20.9% 15.0% 5.0% 12.9% 14.1% 19.5% 18.9% 10.7%

40-49.9% 12.6% 4.1% 1.2% 5.6% 7.5% 10.8% 11.0% 5.3%

50% or higher 20.1% 2.4% 0.6% 7.4% 9.9% 15.3% 28.0% 12.4%

Change

Less than 30% -6.3% -4.5% -1.0% -2.6% -2.6% -9.2% -0.4% -1.0%

30-39.9% 0.0% 2.1% -0.4% -0.9% 4.8% 4.4% -0.8% -0.4%

40-49.9% 0.7% 0.7% 0.8% 0.4% -0.9% 1.4% -0.1% 0.0%

50% or higher 5.6% 1.7% 0.6% 3.1% -1.3% 3.4% 1.3% 1.3%

Source: Tabulations of the 1991-2003 American Housing Surveys.

As shown, in the first half of the 1990s low-income buyers were much more likely to face both

moderate and severe housing cost burdens than either moderate- or high-income buyers. In the 1991

to 1995 period, 32.8 percent of low-income buyers experienced moderate payment burdens, compared

to 16.3 percent of moderate-income buyers and 5.8 percent of high-income buyers. The differences

in the shares of buyers with severe payment burdens were even starker. While 14.5 percent of low-

income buyers paid more than 50 percent of their income for housing, only 0.7 percent of moderate-

income and no high-income buyers faced this degree of burden. While not as extreme as the

differences by income, minorities, particularly Hispanics, were also more likely to face housing cost

burdens than whites. During this period, 24.5 percent of Hispanics first-time buyers had moderate

payment burdens compared to 17.7 percent of blacks and 18.9 percent of whites, while 11.9 percent

of Hispanics and 11.2 percent of blacks had severe payment burdens compared to 4.4 percent of

whites.

Importantly, the share of first-time buyers facing severe housing cost burdens increased considerably

after 1995, particularly among low-income buyers. In the period after 1995, 20.1 percent of low-

income buyers had a severe housing cost burden, a 5.6 percentage point increase from the first half of

the 1990s. While the share of households facing moderate and severe payment burdens increased for

both moderate- and high-income buyers over the period, the increases were much smaller. Among

minorities, Hispanics experienced the largest increases in the share of households with both moderate

(5.8 percent) and severe (3.4 percent) payment burdens. As a result, in the latter half of the 1990s,

Hispanics had payment burdens that were nearly as high as those among low-income buyers. Whites

also saw a jump in the share of households with severe payment burdens (3.1 percent), while blacks

had an increase in the share with moderate payment burdens (3.9 percent). Compared to whites,

blacks were somewhat more likely to face both moderate (21.6 versus 18.5 percent) and severe (9.9

versus 7.4 percent) payment burdens.

For the most part, low-income renters face higher payment burdens than owners. In the period prior

to 1995, low-income recent-mover renters were much more likely to face severe payment burdens,

with 26.7 percent renters in this category compared to only 14.5 percent of low-income buyers.

However, while the incidence of severe payment burdens was rising sharply for low-income buyers,

there was only a small rise for low-income renters. In addition, while the share of low-income buyers

with moderate payment burdens increased by 0.7 percentage points, the share of low-income renters

in this category declined by 0.9 percentage points. As a result, in the period after 1995 more low-

income buyers than renters faced a moderate payment burden (33.5 percent versus 29.9 percent),

while the difference in shares with severe payment burdens narrowed to just 7.9 percentage points

(20.1 percent versus 28.0 percent).

In short, the increase in low-income homeownership does appear to have been associated with fairly

sizeable increases in the incidence of severe payment burdens among first-time buyers. Among

minorities, the share of buyers with high payment burdens is most evident among Hispanics. The

relaxation of mortgage underwriting requirements, which has been credited with helping to fuel the

rise in homeownership rates, may also have contributed to these increases in severe payment burdens.

While most mortgage products in the past required that housing costs (including the mortgage

payment and property insurance and taxes) generally could not exceed about 30 percent of income,

new products designed for low-income borrowers now commonly allow ratios in the upper 30s, while

subprime products may allow even higher payment burdens. When the cost of utilities are added to

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

other housing costs, these more flexible guidelines can result in total payment burdens of 50 percent

or more. Whatever the cause, it is notable that one in five low-income first-time homebuyers were

paying more than 50 percent of their income for housing in the period since 1997, as were one in

seven Hispanic buyers.

Mortgage Financing Choices

The mortgage terms selected by homebuyers can have important implications for their experience as

owners both in terms of long-run mortgage costs and the degree of risk of being unable to meet future

mortgage obligations. One of the most important mortgage characteristics is the interest rate. Higher

interest rates raise the monthly costs of homeownership and also decrease the share of mortgage

payments that go toward principal in the early years of the mortgage, slowing equity accumulation. A

notable characteristic of the mortgage market during the 1990s was the development of the subprime

mortgage market, which gave borrowers who might otherwise not have qualified for a loan an

opportunity to obtain mortgage credit—but at the cost of higher interest rates. Subprime lending has

consistently been found to be disproportionately concentrated among minority and low-income

borrowers and neighborhoods (see Apgar and Herbert (2005) for a review of this literature).

As the market developed, most subprime loans were used to refinance existing mortgages. As a

result, most studies of subprime lending patterns have focused on this segment of the market.

However, the share of home purchase mortgages has been growing steadily. In 1993, subprime

lenders accounts for a little more than 1 percent of all home purchase loans (Joint Center for Housing

Studies, 2004). By 2001 this share had increased to 6.5 percent and by 2002 it was more than 9

percent.

As with refinance loans, subprime purchase loans are more common among minority

borrowers generally, and particularly common in low-income, minority neighborhoods. In 2001,

subprime lenders accounted for 5.1 percent of purchase mortgages for whites, compared to 9.6

percent for minorities. In low-income minority communities, 13.4 percent of all purchase mortgages

were made by subprime lenders, compared to 8.9 percent in high-income predominantly minority

areas and 7.5 percent in low-income predominantly white areas (Joint Center for Housing Studies,

2004). Given the sharp rise in overall purchase lending volumes by subprime lenders in 2002, these

shares are likely to be even higher now.

The increase in subprime purchase lending to minorities and, to a lesser extent, to low-income

borrowers would be expected to be evident in the share of buyers obtaining high interest rate loans.

The top portion of Exhibit 9 presents information on average interest rates for first-time buyers by

income and race-ethnicity for the periods before and after the 1995 survey.

In the period up through

the 1995 survey there was a clear tendency for lower-income buyers to face higher interest rates. The

average interest rate for low-income buyers was 8.81 percent, compared to 8.48 percent for moderate-

The share for 2001 is from the Joint Center for Housing Studies (2004) while the 2002 share is from

http://www.huduser.org/datasets/manu/subprime_2003_distributed.xls.

Recent first-time buyers in each survey are those who purchased their home since the previous AHS survey

two years earlier. As a result, the interest rates reported by buyers in any one survey reflect rates prevailing

during the previous two-year period. For example, interest rates obtained by recent buyers in the 1991

AHS reflect interest rates from the 1989 to 1991 period.

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

income buyers and 8.46 percent for high-income buyers. To put these differences in perspective,

assuming a $100,000 mortgage, the higher interest rates faced by low-income buyers is equivalent to

paying about $25 more per month than higher-income buyers. There were smaller differences in

average interest rates by race, with the average interest rate obtained by whites of 8.54 percent,

compared to 8.72 for blacks and 8.34 for Hispanics, who somewhat surprisingly had the lowest

average interest rate of the three groups.

Exhibit 9

Trends in Interest Rates by Income and Race-Ethnicity for First Time Homebuyers

1989-2003

Income or Race-Ethnicity 1989-2003 1989-1995 1995-2003 Change

Average Interest Rates

Low Income 8.02 8.81 7.43 -1.39

Moderate Income 7.86 8.48 7.39 -1.09

High Income 7.76 8.46 7.24 -1.22

White 7.86 8.54 7.38 -1.16

Black 7.90 8.72 7.51 -1.22

Hispanic 7.72 8.34 7.47 -0.87

Share of Buyers with High* Interest Rates

Low Income 11.4% 12.9% 10.5% -2.4%

Moderate Income 8.3% 6.8% 9.4% 2.6%

High Income 6.3% 4.6% 7.3% 2.7%

White 8.7% 8.7% 8.7% 0.1%

Black 9.2% 8.7% 9.4% 0.7%

Hispanic 10.9% 8.4% 11.9% 3.6%

Notes:

"High Interest Rate" defined as being more than one standard deviation above the mean for the AHS survey

period. A standard deviation ranges from 1.32 to 1.70 over the seven survey periods.

Source: Tabulations of 1991-2003 American Housing Surveys.

The most notable aspect of the trends in average interest rates is the general decline that occurred in

the second half of the decade. For all groups except Hispanics, average interest rates declined by

more than a full percentage point. Interestingly, average interest rates declined more among low-

income buyers, helping to substantially narrow the difference in average rates between low- and

upper-income buyers. This trend suggests that the expansion of affordable mortgage lending products

did contribute to a reduction in interest rates available to lower-income buyers. Blacks also

experienced a slightly larger decline in average interest rates than whites, narrowing the difference in

average interest rates by these groups to only 0.13 percentage points. Hispanics, however,

experienced much smaller declines in average interest rates. But since Hispanics had started the

period with lower average interest rates, in the second half of the decade there was little difference in

the average rates obtained by these two groups—7.47 versus 7.38.

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

Given that subprime lending expanded rapidly in the second half of the 1990s and that this lending

has been disproportionately concentrated among minority and low-income borrowers, it is somewhat

surprising that the trends in average interest rates did not indicate a widening of differences by

income or race-ethnicity.

In order to examine whether trends in average interest rates may mask the

extent to which the share of borrowers facing very high interest rates was rising, loans were identified

as having “high” interest rates if the rate was more than one standard deviation above the mean

interest rate for any survey period.

By this measure there was only a slight increase in the overall

share of home purchase mortgages that were high interest rate. During the 1989 to 1995 period, 8.6

percent of mortgages were high interest rate, compared to 9.1 percent from 1995 to 2003.

The bottom panel of Exhibit 9 presents the share of buyers with high interest rate loans. Interestingly,

while high cost loans are more common among low-income buyers, both moderate- and high-income

buyers experienced larger increases in the share of high cost loans in the late 1990s. While there was

a decline of 2.4 percentage points in the share of low-income buyers using high-cost loans, moderate-

and high-income buyers experienced an increase of 2.6 and 2.7 percentage points respectively. In the

1995-2003 period, roughly one in ten of both low- and moderate-income first-time buyers used high

cost loans, while about one in fourteen high-income buyers used these loans. One possible

explanation for this pattern is that the expansion of conventional lending to low-income buyers offset

the growth in subprime lending to lower the share of buyers obtaining high cost loans. Since

moderate- and high-income buyers would not have benefited as much from the expansion of

affordable lending products, the growth of subprime lending may be more evident among these

groups.

Among racial-ethnic groups, there was little difference in the early 1990s in the share of buyers

obtaining high cost loans. While the share of whites obtaining high cost loans was essentially

unchanged over the decade, the share of blacks with these loans increased by 0.7 percentage points

and the share of Hispanics increased by 3.6 percentage points. This result is in keeping with findings

from the literature on subprime loan usage that minorities are much more likely than whites to

borrower through subprime lenders, but the result is at odds with the literature in that subprime

lending is more common among blacks than Hispanics.

The general conclusion from this analysis of AHS data – that there was not a significant tendency for

low-income and minority homebuyers to face higher interest rates – seems at odds with the fact that

One reason why the increase in subprime lending may not be evident from these data is that higher

borrowing costs come both in higher origination costs as well as higher interest rates. Since the AHS does

not gather information on origination costs, we cannot assess whether there are differences among borrower

groups in these costs.

The variation in interest rates observed across borrowers in any survey period will reflect both variation in

interest rates over the two-year period covered by the survey as well as variation in rates across borrowers

at any particular point in time. Unfortunately, the AHS does not capture the month when mortgages are

originated and so it is not possible to standardize rates by comparing them to some prevailing benchmark

for the month of origination. Across the seven survey periods, the standard deviation of interest rates

ranges from 1.32 to 1.70, with greater variation in the 1991 and 1993 survey periods when interest rates

were falling more rapidly.

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

subprime lenders share of home purchase mortgages increased from 1.3 percent in 1993 to 6.5 percent

by 2001 (Joint Center for Housing Studies, 2004).

However, it is also consistent with two recent

studies that have examined differences by race and ethnicity in the interest rates obtained by

homeowners. Both Susin (2003) and Boehm, Thistle, and Schlottmann (2005) analyze data from the

AHS and find that there is no significant difference in interest rates on home purchase mortgages by

race and ethnicity once differences in other available risk factors are accounted for. These same

studies do, however, find that blacks pay significantly higher interest rates upon refinancing. These

results suggest that the simple tabulations of the AHS showing little different in home purchase

interest rates by race and ethnicity may be a fair depiction of market experience—at least through

2003. The fact that low-income and minority buyers have fared better in the purchase mortgage

market than the refinance market may also be a reflection of the fact that the emphasis of affordable

lending programs has been almost exclusively for home purchase. This may be an indication that

greater attention is needed on developing efforts aimed at assisting owners in the refinance market.

Another important characteristic of the initial mortgage terms is the loan-to-value ratio (LTV). While

higher LTVs reduce the amount of savings buyers need to qualify for a mortgage, making it easier for

low-income households to purchase a home, they also increase the risk that small fluctuations in

home prices will erase the buyers’ equity in the home. The greater prevalence of mortgage products

that allow buyers to put down less than five percent of the purchase price has been cited as one of the

factors contributing to the increase in low-income homeownership since the early 1990s. Exhibit 10

shows the distribution of LTVs among first-time homebuyers by income and racial-ethnic categories

both for the entire period from 1989 to 2003 and the change in the distribution between the period

before 1995 and the years since 1995. As would be expected, low-income buyers generally have

higher LTVs than higher-income buyers. Over the entire period, 24.2 percent of low-income buyers

had LTVs over 95 percent, compared to 21.0 percent of moderate-income and 14.2 percent of high-

income buyers. Nonetheless, a fairly high share of low-income buyers had LTVs of 80 percent or

less, with 45.1 percent of low-income buyers in this category, compared to 41.8 percent of moderate

income and 45.7 percent of high-income buyers.

As a result, there was little difference in the

average LTV across these income groups, ranging from 83.0 percent for low-income buyers, to 84.6

percent for moderate-income buyers and 82.3 percent for high-income buyers. When racial-ethnic

groups are considered, minorities are found to have a higher proportion of high LTV loans than

whites. Among black and Hispanic first-time buyers, 27.2 and 24.0 percent, respectively, had LTVs

above 95 percent compared to 18.3 percent of whites.

In terms of changes over time in the distribution of mortgages by LTV, there was an increase in the

share of higher LTV loans among all categories of first-time buyers, with the largest increases among

blacks (6.3 percentage points), moderate-income buyers (6.2 percentage points), and Hispanics (4.2

While this might be an indication that the AHS does not accurately capture interest rate information, a

recent study by Lam and Kaul (2003) concluded that data from the AHS on interest rates is consistent with

other data sources. In fact, a comparison of interest rates on non-governmental loans found the AHS

averages to be slightly higher, which the authors conclude may be due to the fact that the AHS includes

subprime loans while the comparison data set did not.

One explanation for the fairly high share of first-time buyers with low LTVs could be that they are more

likely to use second mortgages to supplement a smaller first mortgage. The LTV calculation was only

based on the primary mortgage.

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

percentage points). As a result of these increases, in recent years 29 percent of black first-time buyers

and about a quarter of Hispanics and low- and moderate-income buyers have purchased homes with

less than 5 percent equity in the homes at the time of purchase. Should home values drop in the near

future, these buyers would be most vulnerable to a loss of their equity.

Exhibit 10

Trends in Loan-to-Value Ratio by Income and Race-Ethnicity for First Time

Homebuyers, 1989-2003

Income or Race-Ethnicity/

LTV Category 1989-2003 1989-1995 1995-2003 Change

Low Income

80% or less 45.1% 47.0% 44.1% -3.0%

80.1-90% 18.8% 20.4% 18.1% -2.3%

90.1 to 95% 11.9% 10.0% 12.9% 2.9%

Above 95% 24.2% 22.6% 25.0% 2.4%

Moderate Income

80% or less 41.8% 46.6% 38.7% -7.9%

80.1-90% 22.6% 21.9% 23.1% 1.1%

90.1 to 95% 14.6% 14.3% 14.8% 0.6%

Above 95% 21.0% 17.2% 23.4% 6.2%

High Income

80% or less 45.7% 44.0% 46.6% 2.6%

80.1-90% 26.6% 30.6% 24.4% -6.2%

90.1 to 95% 13.5% 13.1% 13.7% 0.6%

Above 95% 14.2% 12.3% 15.3% 3.0%

White

80% or less 44.8% 46.0% 44.1% -1.9%

80.1-90% 23.9% 24.8% 23.3% -1.5%

90.1 to 95% 13.0% 12.4% 13.3% 0.8%

Above 95% 18.3% 16.8% 19.3% 2.5%

Black

80% or less 37.0% 36.6% 37.2% 0.6%

80.1-90% 19.9% 21.6% 19.1% -2.5%

90.1 to 95% 16.0% 19.0% 14.7% -4.4%

Above 95% 27.2% 22.7% 29.0% 6.3%

Hispanic

80% or less 40.5% 43.7% 39.3% -4.4%

80.1-90% 20.0% 26.8% 17.5% -9.3%

90.1 to 95% 15.4% 8.5% 18.0% 9.5%

Above 95% 24.0% 20.9% 25.2% 4.2%

Note: Low income defined as income less than 80 percent of the area median income.

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

Source: Tabulations of 1991-2003 American Housing Surveys.

Exhibit 11 presents information on other key mortgage characteristics. Since adjustable-rate

mortgages often provide initially lower interest rates, this option can be attractive to homebuyers who

are trying to stretch their initial buying power and expect their incomes to rise over the next few years

to meet any increase in interest rates. Fixed-rate mortgages, on the other hand, provide homeowners

with protection against future increases in housing costs due to rising interest rates. The data shown

in Exhibit 11 indicate that there is little variation across income or racial-ethnic groups in the

prevalence of fixed-rate financing. Over the entire period, 87.4 percent of low-income buyers used

fixed-rate financing, compared to 87.6 percent of moderate-income and 86.1 percent of high-income

buyers. Blacks and Hispanics were actually more likely to use fixed rate financing than whites. All

groups increased their use of fixed-rate mortgages after 1995, reflecting the fact that interest rates

were generally lower during this period so buyers were both more motivated to lock these lower rates

in for the long-term and had less need of an adjustable rate product to lower initial interest rates. In

the 1995 to 2003 period about 89 percent of most buyers used fixed-rate financing, with the shares

slightly higher among blacks and Hispanics.

Of note, in the past year there has been considerable attention in the popular press on the growing use

of adjustable rate loans, including a sizeable portion of these loans that are interest only loans for

some number of years. As shown in Exhibit 11, these trends were not yet evident in the AHS data

available as of 2003. It is not known to what extent the reported increased use of adjustable rate loans

was prevalent among first-time buyers. The 2005 AHS survey may shed light on the characteristics

of borrowers that were drawn to these adjustable rate loans since 2003.

Exhibit 11

Trends in Selected Mortgage Characteristics of Low-Income First Time Homebuyers,

1989-2003

Characteristic 1989-2003 1989-1995 1995-2003 Change

Share with Fixed Rate Mortgage

Low Income 87.4% 85.0% 88.9% 3.9%

Moderate Income 87.6% 85.3% 89.3% 4.1%

High Income 86.1% 81.4% 89.0% 7.7%

White 86.1% 83.1% 88.4% 5.3%

Black 91.4% 89.4% 92.5% 3.1%

Hispanic 89.0% 86.2% 90.2% 3.9%

Share with 30-Year Term or Longer

Low Income 62.0% 53.6% 68.4% 14.8%

Moderate Income 77.4% 72.4% 81.2% 8.8%

High Income 80.5% 77.0% 83.2% 6.2%

White 72.9% 66.7% 77.5% 10.8%

Black 75.0% 72.1% 76.4% 4.3%

Hispanic 78.5% 69.1% 82.5% 13.5%

Note: Low income defined as income less than 80 percent of the area median income.

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

Source: Tabulations of 1991-2003 American Housing Surveys.

Exhibit 11 also shows the share of mortgages with terms of 30 years or more. Longer-term

mortgages have the advantage of lowering the monthly payment, but also build up equity more

slowly. Interestingly, low-income buyers have had a tendency to use shorter term financing than

higher-income buyers. This likely reflects the fact that a relatively high share of low-income buyers

choose manufactured homes, which are commonly financed with shorter-term loans than site-built

housing. Over the entire period 62.0 percent of low-income buyers chose 30-year terms or longer,

compared to 77.4 percent of moderate-income buyers and 80.5 percent of high-income buyers. There

is less difference across racial-ethnic groups, although minorities tend to be more likely to use long

term financing than whites. While 72.9 percent of whites had 30-year terms or longer, 75.0 percent of

blacks and 78.5 percent of Hispanics opted for such long terms. All groups experienced an increase

in the share of mortgages with these longer terms after 1995, with larger increases among low-income

and Hispanic buyers. Still, low-income buyers are more likely than higher income buyers to use

shorter-term mortgages, and thus will tend to build up equity more quickly. There is relatively little

difference by race-ethnicity, although Hispanics are slightly more likely to use longer-term mortgages

and so will build up equity more slowly.

Summary

This chapter has made extensive use of information from the AHS from 1991 through 2003 to

identify the characteristics of first-time homebuyers and their housing choices and to examine

whether these characteristics have changed over time. In keeping with the well-documented rise in

homeownership rates, the number of low-income and minority homebuyers rose rapidly beginning in

the early 1990s. Between 1989-1991 and 1995-1997 the number of black and Hispanic first-time

buyers roughly doubled, while the number of low-income buyers rose by nearly 50 percent. After

1997 the number of low-income and black homebuyers remained high but increases moderated

somewhat, while the number of Hispanic buyers continued to increase. One notable change

associated with the increase in low-income and black homebuyers over the decade was the greater

proportion of single-parent households and single persons among first-time buyers. While this trend

is positive in that it indicates greater opportunities among these households who have historically had

lower homeownership rates, it is also true that they may be exposed to greater risks from unexpected

crisis since there is only one wage earner in the household.

While the size and quality of housing purchased by low-income and minority homebuyers tends to be

not quite as good as that enjoyed by moderate- and high-income households, conditions are better

than for low-income renters and at least as good as for the average U.S. household. While there have

been concerns that low-income homebuyers may be much more likely to purchase housing in poor

conditions, the share of homes that are moderately or severely inadequate is only about 7 percent—no

worse than the average for the U.S., although slightly worse than the average for all homeowners.

Overall, low-income homebuyers are satisfied with their homes, with only 8.7 percent of these buyers

rating their homes as 5 or lower on a 10-point scale. In comparison, 9.3 percent of all households and

17.9 percent of recent mover low-income renters rate their homes as 5 or lower.

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

One of the notable differences between renters and owners is the share occupying single-family

detached housing. Low-income and minority owners are much more likely to live in single-family

detached homes than renters, and so do gain access to more living space and greater privacy.

However, a fairly large share of low-income buyers (23.8 percent) purchased manufactured housing.

While there is evidence that these homes do provide good quality at an affordable price, there are

concerns that since a large share of these buyers do not own the land on which the unit sits, they may

not benefit from appreciation in land values. This issue will be explored more in Chapter 4.

Similar to the conclusions regarding housing quality, data from the AHS suggests that low-income

buyers experience better neighborhood conditions and have higher satisfaction with their

neighborhoods than low-income renters, and are similar to all U.S. households in both dimensions.

However, minority homebuyers are more likely than low-income buyers to buy in central cities,

which is reflected in a slightly higher propensity to live near commercial or industrial properties or to

have bars on the windows of nearby buildings. Nonetheless, minorities are slightly more satisfied

with their neighborhoods than low-income buyers.

One strand of existing research has used HMDA data to examine the location choices of low-income

and minority homebuyers in a small number of metropolitan areas. While these studies cannot

identify first-time buyers, the findings are consistent with those from the AHS. Low-income

households are found to be gaining access to suburban areas. While these buyers do tend to locate in

closer-in, lower-income areas, they are also fairly likely to locate in moderate-income areas, which

suggests that a move to homeownership does support some degree of income mixing. These studies

also find that while minorities are gaining access to the suburbs, these buyers, particularly blacks, are

often concentrated in a small number of areas with an above average share of minorities. As a result,

the move to homeownership does not seem to be fostering greater racial integration. But this does not

mean that these neighborhoods are not otherwise fine places to live.

A comparison of neighborhood characteristics of low-income buyers and renters is intended to shed

light on the extent to which a move to homeownership is associated with an improvement in

neighborhood conditions. Several studies provide more direct evidence on the change in

neighborhood conditions associated with a move to homeownership through data gathered on pre-

and post-move neighborhoods for samples of homebuyers participating in subsidized homeownership

programs. In general, these studies find that for the most part there is little change in neighborhood

conditions for these buyers, although there tends to be a small increase in homeownership rates and

the share of housing in single-family units. One study of this type used a national panel study to

examine pre- and post-move neighborhood conditions and so may have broader applicability than the

studies that examine participants in government programs. This study found that while low-income

whites did not experience any real change in neighborhood conditions by purchasing a home, low-

income minorities experienced fairly sizeable improvements, while moderate- and high-income

minorities experienced small positive changes. Nonetheless, the study also finds that the areas where

minorities purchased generally ranked lower on various socioeconomic dimensions than the areas

where whites purchased. In short, collectively these studies suggest that moves to homeownership

are generally not associated with substantial improvement in neighborhood conditions. But nor do

the studies find that low-income or minority homebuyers are systematically being shunted into poor

quality neighborhoods. Instead, they appear to be moving to low- or moderate-income areas with few

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

signs of distress. The implications of these neighborhood choices for the financial returns and social

benefits realized by these owners will be considered in Chapters 4 and 5.

This chapter has also presented information on the mortgage terms obtained by low-income and

minority homebuyers. It is generally believed that the sizeable increases in homeownership rates over

the last decade have been supported by expansion in the availability of mortgage credit through more

relaxed underwriting guidelines. This would suggest that borrowers may have had greater access to

affordable mortgage products over the decade. However, at the same time, there has also been

significant growth in subprime mortgage lending, which expands the supply of credit but at the cost

of higher interest rates and fees. Evidence from the AHS on differences in interest rates across first-

time buyers by income and race-ethnicity suggests that on average low-income and minority buyers

only pay slightly higher interest rates compared to upper-income and white buyers and these

differences tended to narrow over the course of the last decade. The growth of subprime lending was

not yet evident in higher interest rates on purchase mortgages, at least as of 2003. However, the same

cannot be said of refinance mortgages, which will discussed more in the next chapter.

Another important loan term is the ratio between the loan amount and the house value. While low

downpayment loans are important for addressing the lack of wealth that is the principal barrier to

homeownership for most low-income and minority households, it also exposes buyers to greater risk

of losing their investment due to fluctuations in home prices. Low-income and minority homebuyers

are more likely to buy homes with little money down. Since 1995, about a quarter of low-income and

Hispanic homebuyers and 29 percent of black homebuyers have purchased homes with less than 5

percent down, compared to 18 percent of all white buyers. The shares of buyers using such high LTV

loans has increased somewhat from the early 1990s, with increases of 2 to 6 percentage points across

these groups. While the availability of these loans has undoubtedly helped fuel the increases in

homebuying, there are a fairly large share of buyers with little equity in their homes.

In terms of other mortgage characteristics, there is no indication that low-income and minority first-

time buyers are more likely to choose adjustable rate mortgages and thus be exposed to interest rate

risk—at least as of 2003. There is also little difference in the length of the mortgage term by income

or race-ethnicity. However, over the last year there have been numerous news accounts documenting

the rapid growth in market share for various types of adjustable rate mortgages, including those with

interest-only payments. What is not evident from these reports, however, is what are the

characteristics of homebuyers using these loans, particularly to what extent the borrowers are low-

income and minority first-time homebuyers. Perhaps data from the 2005 AHS can shed light on

whether there have been substantial changes in mortgage choices of first-time buyers.

Perhaps the most troubling aspect of the housing choices made by low-income and first time buyers is

the fairly large share facing significant housing cost burdens. In the period since 1995, there has been

a fairly significant increase in the share of low-income buyers having a severe payment burden, where

they devote more than 50 percent of their income for housing costs. During the period 1995 to 2003,

20.1 percent of low-income buyers faced such severe payment burdens, an increase of nearly 40

percent from the 14.5 percent of buyers in this situation in the early 1990s. While black homebuyers

are only slightly more likely to face moderate or significant payment burdens compared to whites,

Hispanics are much more likely to have significant payment burdens, with 30.3 percent having

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

moderate payment burdens (that is, paying between 30 and 50 percent of income for housing) and

15.3 percent having severe payment burdens.

Overall, the evidence from the AHS and the literature paints a somewhat mixed picture of the initial

housing conditions of low-income and minority homebuyers. On the one hand, for the most part

these buyers have obtained decent housing in decent neighborhoods. The houses and neighborhoods

are of higher quality than those occupied by low-income renters and of similar quality to the average

U.S. household. On the other hand, there is not strong evidence that a move to homeownership has

resulted in large increases in neighborhood quality for these buyers. But there is no indication that a

significant share of buyers are ending up in distressed neighborhoods.

Perhaps not surprisingly, there are also indications that there has been an increase in the number of

buyers exposed to the risk of being unable to meet their mortgage obligations. One example of this

trend is the increased prevalence of high LTV loans, with a quarter or more of low-income and

minority buyers purchasing their first home with relatively little money down. While this has

undoubtedly helped to fuel the increase in homeownership, these buyers are also more vulnerable to

fluctuations in house prices. The significant increase in single person and single parent homebuyers

also raises concerns about the ability of these households to respond to crisis with only one earner to

support the mortgage. Most importantly, there has been a growing share of low-income first-time

buyers that are devoting more than half of their income to housing costs, with one-in-five buyers

facing such a severe burden in recent years. These households clearly have little ability to adapt to

any increases in expenses or decrease in income. On a positive note, low-income and minority buyers

do not appear to face significantly higher interest rates at the time of purchase compared to other

buyers.

This chapter has relied much less on a review of the existing literature and more on descriptive

analysis of available data than is true of other parts of this report. In part, this reflects a desire to

present a strong factual base about the recent low-income and minority homeownership boom to help

inform the interpretation of studies about the experience of low-income and minority households as

owners which is the subject of subsequent chapters. But it also reflects the fact that the literature

examining initial housing choices is fairly thin. Several areas for further research stand out in

particular. First, there is a need for multivariate analysis of the housing choices made by low-income

and minority homebuyers to examine whether, in fact, homeownership is associated with greater

housing quantity, quality, and satisfaction, taking into consideration important differences in the

characteristics of renters and owners. Second, it would be interesting to make use of panel surveys of

households to examine how a move to homeownership changes the quantity and quality of housing as

well as its cost. Finally, further analysis of the mortgage choices made by low-income and minority

homebuyers is needed given the importance of these choices in determining the financial benefits of

homeownership.

One of the key findings of this chapter is that the gains in homeownership of the 1990s have been

associated with an increase in the risk profile of first-time buyers, raising concerns about how well

these households will be able to sustain homeownership. This issue is the focus of the next chapter,

which examines the experiences and choices made by low-income and minority buyers after home

purchase.

Chapter 2: Initial Housing Choices Made by Low-Income Homebuyers

Chapter Three:

Key Experiences and Decisions of Low-Income and

Minority Homeowners

In this chapter we examine several critical aspects of the experiences and decisions of low-income

and minority homebuyers after they purchase their home. First and foremost is the question of how

long these owners maintain homeownership. Many of the financial and social benefits of

homeownership are derived from residential stability. The first part of this chapter examines research

related to the question of whether low-income and minority first-time homebuyers are able to sustain

homeownership and realize this stability. A second important issue is the experience of these owners

both in refinancing their primary mortgage and in using debt to tap their accumulated home equity.

These decisions have important implications for the ongoing costs of homeownership and whether

these owners are able to accumulate wealth over time. Finally, this chapter also examines differences

by income and race-ethnicity in homeowners’ tendency to invest in maintenance and improvement to

their homes. Investments in maintenance are important to ensure that the home continues to provide

adequate housing and to maintain its value. Investments in improvements can also contribute to the

owner’s enjoyment of the home and increases in value. Conversely, repairs and replacements can

make it difficult for borrowers to meet their mortgage obligations or force them to increase their

mortgage debt. Hence, these needs constitute cost items for which homeowners do not always have

the resources to cover. The chapter concludes with a summary of findings in each of these areas.

Mobility, Defaults, and Length of Time as Homeowners

Perhaps the most important question regarding the experience of low-income and minority

homeowners is whether they are able to sustain homeownership once achieved. The critiques of

efforts to promote homeownership discussed in the introduction to this report emphasize the high rate

of foreclosure among these buyers as evidence that we may have gone too far in helping households

into homeownership when they are not capable of sustaining it. Without a doubt, there are reasons

for concern about failed attempts at homeownership as the costs of being forced to move out of a

home can be quite high. But even if a move is voluntary, if owners have a short stay in their home

they will be less likely to realize many of the benefits of homeownership that are associated with

residential stability.

In terms of the financial benefits, given the high transaction costs of buying and selling a home,

homeownership becomes very expensive if the household moves frequently. Also, while in the

longer run nominal house prices are very likely to rise, short periods of falling nominal house prices

are not uncommon (Belsky and Duda, 2002b). If an owner is forced to sell their home into a down

market, they will incur these nominal declines in values. Longer tenure in the home will allow

owners to ride out short-term nominal declines and avoid these losses.

Of course, if a household suffers a foreclosure, the costs include not just a loss of their equity, but

also the psychic distress of having failed at homeownership and being forced out of their home and

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

the damage done to the owner’s credit history and ability to obtain credit in the future. Finally, as

will be discussed in detail in Chapter 5, a variety of the social benefits are associated with residential

stability. In short, if low-income and minority homebuyers are more likely to move or more likely to

be unable to sustain homeownership they will be less likely to realize both the financial and social

benefits of homeownership.

The literature touching upon these issues can be divided into three strands. The first strand deals with

residential mobility, which is the tendency for households to move out of their homes. This literature

is generally not concerned with movers’ subsequent housing choice and so does not address the

question of whether owners leave homeownership or are simply trading one owned residence for

another. But these studies are of interest in examining the question of whether low-income and

minority owners might be more likely to incur the high transaction costs associated with moving. The

second strand of the literature relates to the prevalence of foreclosure among homebuyers. Given the

high costs of foreclosure, the frequency of this outcome is of obvious importance. But while the

frequency of foreclosures is important, foreclosures are also relatively rare. Many owners who are

unable to sustain homeownership may be able to take steps to avoid a foreclosure. Still, these owners

may face significant costs from being forced to leave homeownership, including higher future

borrowing costs from having defaulted on a loan. The third strand of the literature, which is relatively

new, uses panel surveys of households to track their tenure choices over time. This literature

provides a more direct indication of the extent to which low-income and minority homebuyers are

able to sustain homeownership over time. The literature dealing with each of these three issues is

discussed in turn.

Residential Mobility

Several recent reviews of the literature have concluded that the evidence is convincing that owners

move less than renters (Dietz and Haurin, 2003; Rohe, McCarthy, and Van Zandt, 2001; Rohe and

Stewart, 1996).

However, the question of interest for this study is whether low-income or minority

homeowners are more or less likely to move than other homeowners, not whether they are more likely

to move than renters. There are actually few studies that compare the mobility choices of

homeowners of different incomes or race-ethnicities. Most studies pool owners and renters and

include income and race-ethnicity as independent variables, but do not interact a household’s tenure

with these variables to examine whether the impact of income or race-ethnicity on mobility differs

between owners and renters.

Most of the handful of studies that do address this issue suggest that low-income households are

somewhat less likely to move than higher income groups (Gronberg and Reed, 1992; Henderson and

Ioannides, 1989; Haurin and Lee, 1989). This result is attributable to the fact that higher income

households have more choices in the housing market and are less deterred by transaction costs and so

The fact that owners move less than renters does not mean that the evidence is clear that homeownership

causes greater residential stability. In fact, individuals are more likely to buy a home when they know they

are less likely to want to move in the near future. In this case, lower expected mobility leads to

homeownership, not the other way around. Still, homeownership would be expected to lower mobility in

several ways. First, higher transaction costs of moving will make owners less inclined to move as their

household circumstances change. Second, owners also have a greater ability to tailor homes to meet their

needs and tastes and so may have less need to move to adjust their housing consumption.

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

are more likely to move than low-income households. However, Varady (1986) found that high-

income households were less likely to move, although this result was not statistically significant. One

difference from the other studies cited is that Varady includes a much more exhaustive list of

explanatory variables, including measures of housing and neighborhood problems and duration of

residence in the home. One explanation for Varady’s finding is that high-income households have

greater ability to move away from housing or neighborhood problems, but in studies that do not

include these types of measures we only observe that high-income owners are generally more mobile.

In fact, they may only be more mobile when confronted by disagreeable situations.

These same studies also generally find that white owners have lower mobility than minorities

(Gronberg and Reed, 1992; Henderson and Ioannides, 1989). This finding is it odds with the

explanation advanced for the higher mobility of higher income households. If it is the degree of

choice in housing markets that drives mobility then minorities ought to have lower mobility than

whites. Instead, it is possible that the greater mobility among minority homeowners is more likely to

reflect their greater difficulty in sustaining homeownership. Again, Varady (1986) reaches the

opposite conclusion, finding that black owners are less likely to move. He interprets this as indicating

that blacks have fewer choices in the housing market due to racial segregation. The different

conclusion reached by Varady may be related to the longer list of explanatory variables used. It may

be that blacks are more likely to be exposed to housing and neighborhood problems and when not

controlled for these factors may contribute to higher mobility by black owners.

In short, the little evidence that exists on differences in mobility among homeowners by income and

race-ethnicity is mixed. This may simply reflect that fact that there are not substantial differences in

mobility rates and so different samples and different methodologies come to different conclusions.

Most of these studies also rely on data from several decades ago and may be of less relevance for

present market circumstances. More recent studies that focus not only on differences in the

propensity to move but also examine the subsequent tenure choice of households are more relevant

for the purpose of this study. This literature is reviewed below.

Mortgage Delinquency and Default

There is a rich literature on the determinants of mortgage delinquency and residential foreclosure,

which the literature generally refers to as default. Perhaps not surprisingly, research consistently

finds that households with lower incomes are more likely to miss payments and default on their

mortgages (see Quercia and Stegman, 1992 for a thorough review of the default literature through that

time). Two more recent studies with an explicit focus on the difference in default experience by

borrower income are Van Order and Zorn (2002) and Deng, Quigley and Van Order (1996). Van

Order and Zorn study the performance of mortgages purchased by Freddie Mac that were originated

between 1993 and 1995 and then tracked through 1999. They divide borrowers into four categories

based on their income relative to the median incomes where they live (80 percent of area median

income or less, 81 to 100 percent, 101 to 120 percent, and above 120 percent). Even after controlling

for a variety of loan characteristics, they find that lower income groups consistently have higher

default probabilities than higher income groups.

The unadjusted default rates (that is, differences in default by income category without taking into

account other differences between these borrower groups) reported by Van Order and Zorn are also

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

instructive as they indicate the extent to which low-income borrowers are likely to experience

foreclosure. The data presented by Van Order and Zorn indicate that even for low-income borrowers

foreclosure is a rare event. Among their cohort of loans from the first half of the 1990s, only 0.8

percent of buyers with income below 80 percent of area median income experienced a foreclosure in

the four to six years following origination. This was only slightly higher than the 0.6 percent of high-

income borrowers that experienced foreclosure over the same period of time.

Of course, since these were loans purchased by Freddie Mac, they represent prime conventional

mortgages, which are less risky than the FHA or subprime loans that may be more common among

low-income borrowers. Recent data from the Mortgage Bankers Association National Delinquency

Survey indicates that the share of FHA loans in the foreclosure process has been about five times

higher than prime loans while the share of subprime loans has been eight to ten times higher than

prime loans.

Still, even for subprime loans, the share of loans in the foreclosure process has only

been about 4 to 6 percent. In short, foreclosure is an uncommon event even for low-income

borrowers.

Deng, Quigley, and Van Order (1996) shed some light on how the expansion of mortgage credit for

low-income borrowers in the form of low downpayment loans may affect foreclosure rates. They

develop a model predicting mortgage default based on the performance of loans purchased by Freddie

Mac that were originated between 1976 and 1983 and then tracked through 1992. This model is then

used to simulate the performance of mortgages over a 15-year period with different assumptions

about borrower income relative to area median incomes, loan-to-value ratio, and fluctuations in house

price appreciation and unemployment rates. Under favorable economic circumstances (long-run

average unemployment of 4 percent and house price appreciation of 5 percent) and assuming a

downpayment of 10 percent, they find little difference in expected 15-year foreclosure rates by

income: 3.56 percent of borrowers with income between 60 and 100 percent of area median income

would default compared to 3.09 percent of borrowers with income between 100 and 150 percent.

the downpayment is reduced to 0 percent, the differences in default rates by income grow larger: 6.58

percent of low-income borrowers would default compared to 4.74 percent of borrowers with incomes

above the area median. If the macroeconomic conditions are also made much more challenging (8

percent unemployment and 0 percent housing appreciation on average), the differences in default

rates grow to nearly five percentage points: 12.88 versus 8.00 percent.

There are several conclusions that can be drawn from these simulations. To begin with, while the

likelihood of foreclosure among all income groups is sensitive to downpayment levels and

macroeconomic conditions, low-income borrowers are more sensitive to these factors than higher-

income borrowers. But it is also true that except under extremely poor macroeconomic conditions,

foreclosure is unlikely to exceed the low single digits. In the prime market it occurs in only about one

Mortgage Bankers Association press release, March 17, 2005, “Residential Foreclosures and Delinquencies

Down From last Year, According to MBA National Delinquency Survey,”

http://www.mortgagebankers.org/news/2005/pr0317.html

The paper also reports default rates for those with income below 60 percent of area median income, but the

sample size for this group is relatively small. They also report default rates for those with income above

150 percent of area median income, but these rates are higher than for those with income between 100 and

150 percent.

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

in twenty cases over a fifteen-year period even when borrowers start with no equity in their homes.

Also, the absolute differences in default rates by income are not large. With 0 percent down loans,

the probability of foreclosure among low-income borrowers is only 1.87 percentage points higher

than among higher-income borrowers. However, should there be a period of sustained poor economic

conditions, with nominal house price growth averaging 0 percent for 15 years, much higher

foreclosure rates would occur, with the foreclosure rate among low-income borrowers more than 50

percent higher than among higher-income borrowers. These results indicate the importance of

providing support mechanisms for low-income borrowers, particularly those with low downpayments

and particularly during challenging economic environments.

The above discussion has focused on differences in default probabilities by income. The issue of

differences in default rates by race-ethnicity has also received a fair amount of attention in the

literature and has been a contentious issue. Berkovec et al. (1994) analyzed the performance of a pool

of FHA loans and found that, all else equal, blacks had a higher default rate than whites. They argued

that this finding was counter to the argument that blacks were discriminated against in the mortgage

origination process as they hypothesized that the rejection of marginally-qualified blacks should

result in higher quality among black borrowers actually funded and so lower default rates. This work

was subjected to a series of criticisms regarding the adequacy of the controls employed for credit

quality and other borrower characteristics as well as the reasonableness of the hypothesis that

discrimination in the origination process should result in lower default rates. Cotterman (2002)

replicated the analysis of Berkovec et al. using data on FHA loans from the early 1990s that had

information on borrower credit history, which was not available to Berkovic et al. Cotterman found

that once credit quality was controlled for, there was no difference in default propensities for

Hispanics and Asians, and generally no difference for blacks as well.

The above studies attempt to control for differences in default propensities by race and ethnicity after

controlling for all observable risk factors. However, for our purposes, default rates by race-ethnicity

that do not control for other factors are of more interest to indicate whether minorities are more likely

to experience default given their characteristics, the types of loans they choose, and the characteristics

of their homes. When foreclosure rates are compared by race-ethnicity it is clear that minorities do,

in fact, face a greater likelihood of losing their homes to foreclosure. Van Order and Zorn (2002)

report statistics for a pool of Freddie Mac loans originated in 1993 to 1995 showing that while only

0.6 percent of white borrowers experienced a foreclosure by 1999, 1.9 percent of blacks and 2.2

percent of Hispanics lost their homes to foreclosure. Cotterman’s sample of FHA loans from 1992,

1994, and 1996, which were tracked through mid 2002, also show higher foreclosure rates for

minorities compared to whites. Across the three sample years of 1992, 1994, and 1996, white

foreclosure rates were 4.1, 4.0, and 2.9 percent. In comparison, black rates were roughly twice as

high at 8.1, 7.6, and 4.8 percent, while Hispanic rates were higher still at 11.0, 8.5, and 5.4 percent.

As the findings of these papers show, differences in foreclosure rates are much larger by race and

ethnicity than they are by income. In addition, the figures presented in Cotterman’s paper also

Cotterman tested the model on different origination samples and using different explanatory variables. In

most cases the black indicator variable was not significant. But in one case it was significant and positive,

indicating blacks had a higher default probability, and in another it was significant and negative, indicating

they had a lower default probability.

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

highlight that foreclosure rates among FHA borrowers are much higher than among prime borrowers.

While it is still the case that the vast majority of FHA minority homeowners do not experience

foreclosure, the 8.1 percent of blacks and 11.0 percent of Hispanics who lost their homes within eight

years of purchase are not insignificant. As with low-income homebuyers, these figures underscore

the need to provide support for thee borrowers to be able to sustain homeownership.

Aside from the question of whether low-income and minority homebuyers face greater risks of

foreclosure, there is also a question of whether the risk of foreclosure has risen in recent years. There

have been a wide variety of studies that have documented increases in foreclosure rates in specific

markets or neighborhoods since the mid 1990s, with these increases often linked to increases in

subprime lending (National Information and Training Center, 1999; Bunce et al., 2000; Immergluck

and Smith, 2004; The Reinvestment Fund, 2005). However, while these studies have documented

sharp increases in the number of foreclosures in specific areas, long-term national foreclosure

statistics compiled by the Mortgage Bankers Association (MBA) indicate that nationally the increase

in the foreclosure rate has not been as dramatic as these studies would suggest. For example, MBA

data indicates that among all mortgages, the share entering foreclosure rose from 0.26 percent in 1986

to around 0.33 percent in the early 1990s, before jumping again in the late 1990s to reach 0.46 percent

in 2002. Since then the rate of foreclosures started has moderated somewhat to 0.42 percent in

2004.

But while the incidence of foreclosure has increased by about 60 percent since 1986, the start

of foreclosure still remains a rare event. However, it may be that the MBA data simply do not

adequately capture the full range of subprime lending activity as their subprime foreclosure data is

based on a small sample of subprime lenders (Denis and Pennington-Cross, 2005).

The MBA data highlight two factors that have contributed to the rise in foreclosure rates. First, the

MBA began to track foreclosures among subprime loans in 1998, which revealed a much higher rate

of foreclosures among these loans. In 2001, as the rate of foreclosures started among all loans

reached their peak of 0.46 percent, among subprime loans this rate was 2.34 percent, or more than 10

times the rate among prime conventional loans. But since that time, the subprime foreclosure rate as

reported to the MBA has steadily declined to 1.50 percent in 2004, reflecting a general increase in

credit quality in the subprime sector (Danis and Pennington-Cross, 2005). Despite the decline,

subprime foreclosure rates remain nearly eight times as high as the rate among prime conventional

loans. Another factor that has contributed to rising overall foreclosure rates has been a large increase

among FHA loans. Foreclosure rates among FHA loans have more than tripled since 1986, rising

from 0.32 percent to about 0.60 percent in the late 1990s and then again to 0.98 percent in 2004.

MBA data for the share of loans entering foreclosure as reported in U.S. Housing Market Conditions, May

2005.

The MBA also tracks the share of loans that have missed any payments or have not made a payment in 90

days or more (seriously delinquent). A similar pattern of much higher rates of delinquency are evident

among subprime and FHA loans relative to prime loans, although the share of loans in these categories are

obviously higher than the share entering foreclosure. Still, serious delinquency is a rare event, with only

0.29 percent of prime, 2.72 percent of subprime, and 2.75 percent of FHA loans in this state in 2004. We

have focused on the share of loans entering foreclosure since this rate more accurately captures the share of

borrowers who are likely to lose their homes.

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

The sharply higher foreclosure rates among subprime and, to a lesser extent, FHA loans accounts for

the findings of studies that have documented the fact that in specific markets the number of

foreclosures has increased by several fold since the mid 1990s. While these highly concentrated

foreclosure patters are clearly a cause for concern, it is important to bear in mind that the overall

incidence of foreclosure remains a fairly uncommon event. Even among riskier subprime and FHA

loans, less than 1.5 percent of these loans entered foreclosure in 2004. Thus, the vast majority of

borrowers are unlikely to face foreclosure.

Length of Time as Homeowners

One of the concerns with drawing conclusions about whether low-income and minority buyers share

in the benefits of homeownership from the literature analyzing residential mobility is that some share

of moves represent positive outcomes—owners trading up to better quality homes. On the other

hand, estimates of the share of borrowers losing homes to foreclosure may also underestimate the

failure of buyers to succeed as homeowners by ignoring cases where buyers are forced by

circumstances to move out of their homes, but do not experience a foreclosure. Cases where owners

reluctantly put their homes up for sale, possibly at a financial loss, are not captured in foreclosure

statistics. Recently, several studies have made use of panel surveys – surveys that track the same

households over time – to examine the question of how long low-income and minority first-time

buyers maintain homeownership (Reid, 2004; Haurin and Rosenthal, 2005a; Haurin and Rosenthal,

2005b; and Boehm and Schlottmann, 2004b). By capturing all cases where owners leave their home

and do not purchase another one, these studies provide a much better indication than either the

residential mobility or default literature of the degree to which these buyers are able to remain owners

over time and so reap the benefits of homeownership.

One of the surprising conclusions from these studies is that a fairly sizeable share of all first-time

owners – regardless of income or race-ethnicity – return to renting or living with others after first

achieving homeownership. Both Reid (2004) and Haurin and Rosenthal (2005a) find that about 40

percent of first-time homebuyers leave homeownership at some point after buying. These studies also

find that low-income owners face a higher risk of being unable to sustain homeownership over time.

Reid’s analysis of data from the Panel Study of Income Dynamics from 1976 through 1993 found that

53 percent of low-income buyers left homeownership within five years of buying their first home,

compared to 23 percent of high-income buyers.

Employing a less restrictive definition of low-

income, Haurin and Rosenthal’s analysis of data from the National Longitudinal Study of Youth from

1979 through 2000 found that about 43 percent of low-income buyers did not sustain homeownership

for more than five years, compared to 30 percent of high-income buyers.

While Reid cites specific survivorship rates for some subgroups in the text of her study, in some cases

specific rates had to be estimated based on survivorship graphs shown in the report.

Reid’s definition of low-income required that the household have income less than 80 percent of area

median income in every year through the year in which they bought their first home. High-income buyers

were those whose income was above the area median income every year through the year they purchased

their home. All other households were considered moderate income. Haurin and Rosenthal, in contrast,

defined households based on their income at age 25 relative to all other 25-year olds. Those in the bottom

quartile were considered low-income, while those in the top quartile were considered high income. The

survivor rates for Haurin and Rosenthal are unpublished figures obtained from the authors.

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

In some respects, given differences in the samples used and time periods studied, the results of these

two analyses are somewhat similar: as an approximation, roughly half of low-income buyers exit

homeownership within five years of purchase compared to a quarter to a third of high-income

buyers.

However, the difference in survival rates between low-income and high-income buyers is

much larger in Reid’s study than Haurin and Rosenthal’s. While Reid finds a 30 percentage point

lower survival rate for low-income buyers, Haurin and Rosenthal find a difference of only 13

percentage points. This may well be due to Reid’s more restrictive definition of “low-income,” which

requires that households have income less than 80 percent of the area median income in every year

they are observed through the year in which they buy. Haurin and Rosenthal, on the other hand,

define low-income as those in the bottom quartile of the income distribution in their sample at age 25.

Reid’s sample is also somewhat older, consisting of renters between the ages of 18 and 45, while

Haurin and Rosenthal begin with a sample of those age 14 to 22. Given the differences in the age

groups and definition of low-income, Reid’s results may well represent the experience of what might

be thought of as more permanently low-income households. Further work is needed to understand the

difference in findings between these studies for low-income owners.

Both of these studies also find that minorities are much more likely to return to renting or living with

others than whites. Reid’s five-year rates of exits from homeownership for minorities are between 22

and 38 percent more likely than for whites in the same income categories.

Haurin and Rosenthal

(2005a) find that blacks are 46 percent more likely than whites to be unable to sustain

homeownership, while Hispanic’s were 38 percent and Asians 39 percent more likely to leave.

Reid

reports that after five years about 29 percent of high-income minorities did not sustain

homeownership compared to 21 percent of high income whites, while 58 percent of low-income

minorities were no longer owners compared to 46 percent of low-income whites.

The high rates of exit for low-income and minority first-time buyers are a cause for concern as the

benefits of homeownership will generally be much greater for those who continue as owners for

longer periods. This is both because the odds of benefiting from appreciation increase with time and

the benefits of amortizing loans increase exponentially with the aging of the loan. In addition, social

benefits of homeownership are strongly linked with residential stability. One point worth noting is

that Reid’s study pertains to the period from 1976 through 1993, which is prior to the increase in

homeownership experienced over the last decade. Haurin and Rosenthal examine the period from

1979 through 2000, which does capture some of the period of rising homeownship rates, although

given the age of the persons in their sample many made their first foray into homeownership during

The analysis by Boehm and Schlottmann (2004b) of the PSID from 1984 through 1992 produces a higher

rate of success in maintaining homeownership for at least five years both for all households and for white

low-income owners. For example, among high income whites, they find that 95 percent survive 5 years,

while among low-income whites 82 percent survive five years. A key difference from the other studies is

that Boehm and Schlottmann do not limit their sample to only first-time buyers. The difference in results

may reflect the fact that repeat homebuyers are more likely to maintain homeownership over time.

Reid’s analysis does not distinguish between different types of racial-ethnic minorities.

Haurin and Rosenthal find that the differences between whites and both Hispanics and Asians are

completely accounted for by other household characteristics, while large, statistically significant

differences remain for blacks even after controlling for other factors.

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

the 1980s. Thus, these high rates of exit from homeownership identified in these studies are not a

new phenomenon, but have been evident for several decades. These studies do not shed light directly

on whether the rate of exit from homeownership has increased in recent years. However, the

continued gains in homeownership through 2004 suggest that homeownership exits have probably not

risen as an increased homeownership failure rate would have a dampening effect on overall

homeownership.

In fact, Haurin and Rosenthal (2005b) discuss the implications of more frequent exits from

homeownership and longer intervening spells as renters on overall homeownership rates. They find

that the difference in the shares of whites and blacks in their sample who ever attain homeownership

is much smaller than the difference in homeownership rates observed at the end of the sample period.

They also find that about a quarter of this difference in ending homeownership rates is accounted for

by the greater tendency of blacks to leave homeownership and to take longer to regain

homeownership.

If more low-income and minority families were exiting from homeownership in

recent years, this would serve to depress overall homeownership rates for these groups. Instead, they

have continued to rise.

Perhaps a more important caveat regarding these findings is that most exits from homeownership are

not permanent. Returning to renting or living with others can be an effective way to adjust to changes

in circumstances. A key question is what share of these households later return to homeownership.

While Reid’s analysis only extends up to the point when households leave homeownership, Haurin

and Rosenthal (2005b) continue to track households to identify how frequently households regain

homeownership. While their analysis does not examine differences across income groups, they do

report differences by race-ethnicity. Of white first-time buyers, 69 percent of those who moved back

to renting or living with others for a period are ultimately observed to return to owning. Thus, the

majority of these exits from homeownership are temporary. The rates of returns to homeownership

are lower for minorities, but a majority also succeeds in returning to ownership status, including 59

percent of blacks and 64 percent of Hispanics. Overall, of those who become first-time buyers, the

share who are observed to either never leave homeownership or to return to owning after a spell of

renting or living with others is fairly high for all racial-ethnic groups. Among whites, 86 percent who

become owners either continually sustain ownership or return to owning, compared to 81 percent of

blacks and 84 percent of Hispanics. However, Haurin and Rosenthal do find that compared to whites,

the average length of stay as owners is shorter for minorities and the period of renting or living with

others is longer. These differences contribute to overall differences in ownership rates.

Factors Contributing to Leaving Homeownership

The studies by Reid (2004) and Haurin and Rosenthal (2005a, 2005b) also estimate models to identify

the factors associated with a household leaving homeownership. The explanatory variables included

in these models include both the characteristics of the household when they first purchased as well as

changes in their personal circumstances and the macroeconomic environment after purchase. Aside

from income and race-ethnicity, one of the most important household characteristics is whether the

Specifically, while the white-black gap in homeownership rates in their sample in 2000 is 34 percentage

points, the difference in the shares of whites and blacks that become first-time buyers over the period is

only 26 percentage points.

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

owner is a married couple. Haurin and Rosenthal (2005a) find that this is the single most important

factor predicting the length of time that homeownership is maintained. Married couples are found to

have half the risk of exiting homeownership as households headed by single persons. Reid’s results

find a similar magnitude, although she finds that the importance of being married is somewhat less for

low-income households, who face only a 30 percent increase in risk of exiting homeownership if they

are not married. This finding supports the concerns raised in Chapter 2 that the growth in

homeownership among single persons and single-parent households may raise the number of owners

who are vulnerable to economic shocks.

Other household characteristics associated with the risk of leaving homeownership are age and

education. Younger households are found to be at greater risk of returning to renting or living with

others. As shown in Chapter 2, low-income first-time buyers include a relatively large share of both

younger and older households. While the younger households may be at greater risk of leaving

homeownership, those who come to homeownership later may be more likely to sustain it. Those

with more education are also more likely to sustain homeownership. Both Reid and Haurin and

Rosenthal speculate that education likely captures the long-run earnings potential of the owner, with

higher educated owners more likely to experience rising earnings. To the extent that greater

education is associated with greater financial literacy, this result would also be consistent with the

importance of financial knowledge to maintaining homeownership.

The studies also examine the importance of changes in household circumstances for precipitating an

exit from homeownership. It is generally believed that “trigger events,” which are unexpected

changes in a household’s circumstances, are important factors in producing defaults or otherwise

ending homeownership spells (Vandell, 1995; Elmer and Selig, 1999). The most commonly cited

trigger events are a reduction in earnings as a result of job loss, the splintering of the household due to

divorce or separation, or an increase in expenses or reduction in earnings due to a health crisis. Cutts

(2003) reports that among delinquent Freddie Mac borrowers during the period from 1999 to 2003, 40

percent reported unemployment or curtailment of income as the reason for their delinquency. The

next most common issue was illness or death of the borrower or someone in the family, which was

reported for 24 percent of delinquent borrowers. Marital difficulties and excessive financial

obligations each were cited in about 10 percent of cases.

To capture job loss or income curtailment, Reid includes an indicator of whether the borrower

experienced an unemployment episode, while Haurin and Rosenthal include a change in household

earnings compared to earnings in the year when they purchased the home. Both find these measures

to be significant predictors of whether a household will cease to own. Reid found an unemployment

spell more than doubled the probability of ending ownership, with larger impacts on higher-income

households. Haurin and Rosenthal found a somewhat smaller effect, with declines of $10,000 in

earnings raising the risk of leaving homeownership by 11 percent. Reid reports that job loss was

more common among low-income households, with about 9 percent of low-income households and

15 percent of low-income minorities having a spell of unemployment compared to 6 percent of high-

income households. Haurin and Rosenthal also find that among the low-income households that

switch back to renting or living with others, earnings declined on average by $13,629, or 37 percent,

in the year when this transition occurred, which suggests that declines in earnings are common among

those leaving ownership. Further supporting the importance of employment loss, Haurin and

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

Rosenthal find that increases in the state unemployment rate are significantly associated with moves

out of homeownership.

Both Reid and Haurin and Rosenthal find that divorce is the single event that is most strongly

associated with termination of an ownership spell. Reid’s findings indicate that a divorce raises the

probability of leaving homeownership roughly by a factor of 10, while Haurin and Rosenthal’s

estimate is a more modest, but still significant, 40 percent.

Haurin and Rosenthal include an indicator of whether there was a change in the buyer’s health that

limited the amount or type of work they could do. This situation was rare in the sample (occurring in

only 1 percent of cases) and was not statistically significant. Some of the effect of illness may have

been captured by their measure of a change in earnings. The other impact of a health problem is on

the costs incurred by the household, particularly if they do not have health insurance. The change in

health measure might have been expected to capture the impact of uninsured health care costs, but

this was not found to be an important factor—at least in this sample.

Haurin and Rosenthal also evaluate the impact of changes in both non-housing wealth and housing

values on the likelihood of leaving homeownership and find that neither is particularly important.

Non-housing wealth might be expected to act as a buffer in the event of income loss or an increase in

expenses, but the results do not support this role for wealth. Increases in house values might be

expected to support sustained homeownership by increasing overall household wealth, but the

coefficients on these variables are not statistically significant. One reason for this result might be that

they measure house values in real terms, while nominal changes in value may be more relevant for

housing choices.

One of the key findings in Chapter 2 is that a significant share of low-income homeowners during the

1990s started out with severe housing cost burdens, devoting 50 percent or more of their income for

housing. Neither Reid nor Haurin and Rosenthal include a measure of housing costs (either at the

time the home is purchased or during the time after purchase) in their models, so there is no direct

measure of the extent to which high housing costs precipitate a movement out of homeownership.

But Haurin and Rosenthal do include measures of prevailing mortgage rates, both at the time of

purchase and changes in these rates over time, all of which are found to be highly significant. A one-

percentage point increase in initial mortgage interest rates is found to increase the risk of leaving

homeownership by 16 percent. The authors note that this finding provides an indication of the

increased risks faced by low-income buyers using higher-cost subprime financing. They also find

that a one percentage point increase in rates over time increases the risk of leaving homeownership by

30 percent, while a one percentage point decline reduces the risk of leaving by 15 percent. They

interpret these latter results as indicating the risks faced by those using adjustable rate financing.

However, their data do not indicate whether borrowers actually have an adjustable rate product.

Another interpretation of this result could be that owners who are forced by circumstances to change

residence have a harder time maintaining ownership during periods when interest rates are high. In

sum, while Haurin and Rosenthal do not incorporate any direct measure of housing costs, their

findings regarding the importance of interest rate levels and changes do suggest that housing costs are

an important factor in sustaining homeownership.

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

Related to the issue of changes in housing affordability, Haurin and Rosenthal also investigate

changes in low-income first-time homebuyers’ earnings over time. They find that on average the

earnings of low-income owners grew at 13 percent, which is at least twice the growth rate of low-

income renters and moderate- or high-income owners. Thus, they conclude that for most low-income

owners strong income growth will help to lower housing cost burdens over time and help make

homeownership more sustainable.

The panel nature of the AHS makes it possible to track first-time homebuyers over time to see how

housing cost burdens change in the years after home purchase as well as whether those with higher

initial cost burdens are less likely to continue owning the same home. In terms of changes in housing

costs burdens, analysis of AHS surveys from the 1991 through 2003 suggests that housing costs as a

percent of income do drop over time, but only modestly. Pooling data from all AHS surveys from

1991 through 2003, 8 percent of first-time buyers were found to have a severe payment burden at the

time of purchase, where housing costs accounting for more than 50 percent of their income. Among

first-time buyers still in their home after four to six years, this share was unchanged at 8 percent.

However, there was a decline in the share of first-time buyers with moderate payment burdens (where

housing costs account for between 30 and 50 percent of income) from 20 percent to 15 percent of

owners. The median burden among first-time buyers declined over the first four to six years from

22.5 percent of income to 19.5 percent.

The AHS also provides some insights into differences in the propensity of first-time buyers to stay in

their first home given differences in the initial payment burden. It is important to note that since the

AHS tracks housing units over time, not households, we cannot tell if buyers who moved out of their

homes continued to own or not. But the rate at which first-time buyers leave their first home provides

some indication of the degree to which these households are able to maintain homeownership in the

face of high initial payment burdens. Across all the AHS surveys from 1991 through 2001, we find

that buyers who start out with initial payment burdens over 50 percent are less likely than other

buyers to still be in their home by the time of the next AHS survey two years later: 61 percent of

buyers with severe payment burdens at purchase are still in their homes, compared to 70 percent of

those with moderate burdens and 71 percent of those with burdens less than 30 percent of income.

However, the rate at which these buyers leave their homes stabilizes after the first period, so that by

the time that six to eight years since purchase have passed, there is little difference in the share of

buyers still in their homes by initial payment burden: 40 percent of those with severe initial burdens,

compared to 42 percent of those with moderate burdens and 40 percent of those with affordable costs.

One reason why the initial burden may not be such an important determinant of which households

continue as owners is that there appears to be fairly significant changes between survey periods in

whether a household has a severe payment burden. For example, of the households that have severe

payment burdens at purchase, only 19 percent are found to still have such an extreme burden in the

next survey period. Meanwhile, of those with payment burdens of less than 30 percent at purchase, 5

percent have severe payment burdens two years later. This suggests there is a fair amount of

volatility in the income of low-income owners.

Since the AHS is conducted every two years, new first-time buyers are those who moved into a home in the

two years since the last survey and so have been owners for between zero and two years. Those buyers still

in the home two survey periods later would thus have been owners for between four and six years.

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

While these studies do shed some light on the factors associated with exits from homeownership, they

also leave many questions unanswered. While it appears clear that a loss of income is an important

factor in precipitating an exit from homeownership, the specific factors that cause these declines in

income and what support is needed to help owners cope with these changes are not well understood.

Another puzzle is that while data on the causes of delinquency and studies of long-run trends in

foreclosure rates (Elmer and Selig, 1999) suggest that health problems and a lack of health insurance

may contribute to owners’ difficulties in maintaining homeownership, the analysis by Haurin and

Rosenthal does not find evidence to support this view.

It is also likely that many exits from homeownership are simply rational decisions in response to

changes in circumstances that do not impose significant costs on the owners. The fact that a fairly

high proportion of high-income households leave homeownership within five years suggests that a

failure to be able to afford homeownership is not the only reason for these departures. In short, there

is a clear need for more information about the dynamics of homeownership over time, including the

changes that occur in household circumstances, how different households respond to these changes

(including whether they can draw upon savings, debt, insurance, or resources provided by family and

friends), and how these responses are associated with different outcomes. A better understanding of

these dynamics would allow policy makers to identify which households are at greater risk of being

unable to sustain homeownership and what types of support is needed to increase their chances of

success.

Mortgage Financing Choices After Purchase

Mortgage financing choices made by homeowners after home purchase can have important

repercussions for the financial benefits realized from ownership. One important decision is to

refinance into lower interest rate loans when market conditions provide the opportunity to do so. The

failure to take advantage of such opportunities can result in much higher interest rate costs over the

life of the mortgage. Owners also can use loans to tap accumulated home equity. While the

availability of this wealth is one of the benefits of homeownership, changes in mortgage markets over

the last decade have made it much easier to tap home equity both through refinancing of existing

mortgages and through home equity loans or lines of credit. The ease with which owners can tap

their home equity may make it easier to use their wealth to support current consumption which both

increases housing costs and erodes the development of a nest egg to help weather financial crises,

fund investment in homes, business or education, or support the owner in retirement. This section

explores what is known about differences among homeowners by income and race in their propensity

to take advantage of refinance opportunities to lower interest rates or to cash out accumulated equity

for other uses. Each of these issues is considered in turn.

Refinance Activity

In general, analysis of refinancing activity has found that low-income and minority homebuyers are

less likely to refinance their primary mortgage than upper-income households or whites. Exhibit 12

presents data from the 2003 AHS as an indication of the difference in magnitude of the likelihood of

refinancing. As of 2003, 12 percent of low-income owners had primary mortgages that were

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

refinanced, which is about half the share of moderate-income owners and a third of the share of high-

income owners. Comparing different racial-ethnic groups, whites (24 percent) are more likely to

refinance than either blacks (14 percent) or Hispanics (21 percent). Exhibit 12 also shows that first-

time buyers who are still in their first home are least likely to have refinanced (5 percent), although

this likely reflects the fact that on average they have the shortest tenure in their homes and so have

had less opportunity to refinance.

Exhibit 12

Refinance Activity and Reasons for Refinancing by Income and Race-Ethnicity of

Owners

Low

Income

Moderate

Income

High

Income

White Black Hispanic

First-time

Buyers

Share of All Owners with

Primary Mortgage Refinanced

12% 26% 35% 24% 14% 21% 5%

Reason for Refinancing

(Share of those with refinanced mortgage)

To get lower interest rate 83% 90% 93% 90% 84% 87% 97%

To increase payment period 2% 3% 2% 2% 2% 2% 0%

To reduce payment period 9% 11% 13% 12% 9% 12% 5%

To renew or extend a loan 1% 1% 1% 1% 3% 1% 0%

To receive cash 14% 12% 11% 11% 12% 15% 5%

Other reason 12% 10% 6% 8% 11% 12% 4%

Note: Reasons for refinancing sum to more than 100 percent as more than one reason can be reported.

Source: 2003 American Housing Survey.

Exhibit 12 also shows the reported reasons for refinancing among owners who have refinanced. By

far the most common reason given among all groups of owners is to obtain a lower interest rate, with

at least 83 percent of all owners reporting this reason. However, low-income and minority borrowers

are more likely to report a desire to take cash out or “other reasons” for pursuing a refinance. Since

one of the common reasons for refinancing is to consolidate non-housing debt into lower cost and

longer-term mortgage debt, it is likely that this motivation is captured in the “other reasons” category

(Canner et al., 2002). Considering both the shares motivated to refinance to take cash out and for

other reasons, 26 percent of low-income owners report these reasons, compared to 22 percent of

moderate-income owners and 17 percent of high-income owners. Blacks (23 percent) and Hispanics

(27 percent) are also more likely to report these motivations than whites (19 percent).

While these overall refinance propensities from the AHS provide some indication of the prevalence of

this activity by income and race-ethnicity, these simple tabulations do not take into account other

differences in borrower circumstances that affect the likelihood of pursuing a refinance. For example,

borrowers with mortgages that are largely paid off would be expected to be less likely to refinance

because the small loan size reduces the benefits while financing costs are larger as a percent of the

outstanding loan balance. Two recent studies have examined differences in the propensity of

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

homeowners by income and race-ethnicity to refinance using multivariate techniques to control for

other differences in loan and borrower characteristics. In general, these studies find that low-income

and minority homeowners are, indeed, less likely to refinance when interest rates fall and so may not

be realizing the benefits of interest rate reductions to the same degree as other owners. However,

these studies also suggest that low-income and minority generally seem to be responsive to market

conditions, but they are more likely to be impeded from taking advantage of these opportunities by

other financial constraints.

Van Order and Zorn (2002) analyze a sample of loans originated in 1993 through 1997 and purchased

by Freddie Mac. These loans were tracked through 1999 to examine their refinance activity. In order

to evaluate whether there are differences in the probability of refinance by homeowner income or

race-ethnicity, they divide their observations into groups based on a comparison of the borrowers’

current interest rate and prevailing market interest rates at the time. Specifically, the samples they

report include cases where market rates are above the loan rate, near to the loan rate, and below the

loan rate. They then model the probability of a refinance occurring as a function of the borrowers’

income, race-ethnicity, and the age of the loan within each of these sub-samples. A comparison of

the coefficients for the income and racial-ethnic groups across the samples is used to indicate whether

low-income or minority homeowners behave differently in different interest rate environments. The

results indicate that there is no difference in the propensity to refinance by either income or race-

ethnicity when current interest rates are either higher or close to the borrowers’ existing mortgage

rate. This finding indicates low-income and minority borrowers are not more likely to refinance

when doing so would result in a higher interest rate. However, these owners are found to have a

lower probability of refinancing when current interest rates are below the owners’ existing mortgage

rate. Thus, low-income and minority owners are missing opportunities to lower their borrowing rates

and so facing higher costs of homeownership than is necessary.

Van Order and Zorn also estimate models including other controls for borrower risk, including credit

score, loan-to-value ratio, and their loan amount. Once these additional factors are included they find

that there are no longer any differences in the probability of refinance by borrower income. Thus, the

tendency of lower-income borrowers to miss refinance opportunities is related to the fact that they are

more likely to be credit constrained. Among minorities, the inclusion of these other variables

diminishes the likelihood of missing a beneficial refinance opportunity, but some differences do

persist. In fact, in these specifications both blacks and Hispanics are found to have a similarly lower

probability of refinancing in all interest rate environments. Since refinance activity during periods of

relatively high interest rates is more likely to reflect owners’ need to tap home equity for other

purposes, the lower rate of minority refinance activity during these periods suggests they are also less

likely to pursue cash-out refinances to meet other needs. This issue will be discussed more below.

The findings of Van Order and Zorn are consistent with those of Archer et al. (2002) based on an

analysis of data from the AHS. Archer et al. examine refinance activity in five different periods from

the mid 1980s through the mid 1990s. The focus of their analysis is on identifying any differences

between low-income and other owners in the importance of various factors in predicting the

likelihood of a refinance. They do this by interacting all of the explanatory variables with an

indicator variable for low-income owners. In general, they find little difference in the importance of

explanatory variables in predicting the refinance behavior of low-income versus other owners. Most

notably, they found no difference in their sensitivity to opportunities to refinance to a lower interest

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

rate. Among the few statistically significantly differences they found was that low-income owners

were more sensitive to restrictions imposed by high existing loan-to-value ratios. The authors

conclude that the lower propensity of low-income owners to refinance is due to the fact that they are

more likely to face credit constraints than higher income owners. While Archer et al. do include

controls for borrower race, they do not discuss these results. In general, the racial-ethnic coefficients

were not statistically significant, but this result may simply reflect the fact that the AHS may not have

had large enough samples of minority owners to detect an effect.

Canner et al. (2002) provide further support for the conclusion that low-income and minority owners

are less likely to refinance. This study analyzes a sample of 2,240 homeowners interviewed by

telephone in 2002. The survey gathered information on the refinance activity of these owners in 2001

and the first half of 2002. Using the results, they estimate a model predicting the probability that an

owner obtained a refinance loan during this period as a function of a variety of borrower

characteristics and loan terms prior to the refinance. Their results indicate that borrowers with

income under $40,000 were 1.4 percent less likely to refinance. Minorities were also 4.0 percent less

likely to refinance, although this result was not statistically significant.

Finally, like Archer et al., Nothaft and Chang (2004) use the AHS to estimate models predicting the

likelihood of refinancing. They pool the AHS from four different periods between the late 1980s and

2001 when interest rates declined and provided good opportunities for refinancing. Their results are

similar to those of Archer et al., although by pooling the AHS from several years they also obtain

statistically significant results for minorities, finding that blacks and “other” minorities are less likely

to refinance than whites.

One of the interesting contributions of Nothaft and Chang is their attempt to estimate the value of the

missed refinance opportunities by blacks and low-income owners. Given a prototypical upper-

income, married couple homeowner, they find that blacks are 16.5 percent less likely to refinance

than whites. Assuming a mortgage of $100,000, average interest rates of about 8 percent and an

average decline in the interest rate of 1.33 percentage points from refinancing, they estimate the

present value of lower interest payments of the mortgage of $12,394 per borrower. If 16.5 percent of

blacks miss this opportunity, this works out to an average lost benefit of $2,040 per black

homeowner, or $22.0 billion in lost benefits across all black homeowners. Employing the same

methodology, they find that 6.9 percent of low-income homeowners miss out on refinance

opportunities, with a total lost benefit of $21.9 billion. In short, the benefits lost by missed refinance

opportunities of these owners are substantial.

While not the subject of a detailed analysis, Nothaft and Chang also indicate that there is a substantial

difference between whites and blacks in the decline in interest rates obtained through refinance.

While whites average a decline of 1.33 percentage points, blacks average a decline of only 0.39

percentage points. Several other recent studies that have examined racial disparities in mortgage

interest rates have also observed that blacks obtain much less financial benefit from refinancing.

Susin (2003) and Boehm, Thistle, and Schlottmann (2005) both analyze data from the AHS and find

that blacks pay higher interest rates than whites. Susin’s analysis of all outstanding mortgages as of

2001concludes that blacks pay about 0.44 percentage points higher rates on average, with much of the

difference associated with differences in rates obtained through refinancing. Boehm, Thistle, and

Schlottman’s analysis of primary mortgages originated from 1990 through 2001 finds that the interest

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

rates obtained by blacks who refinance are on average 0.75 percentage points higher than the

refinance rates obtained by whites. When they estimate statistical models that take into account a

variety of borrower and loan characteristics, they find that the unexplained difference in refinance

rates increases to 1.01 percentage points.

The significant differences in mortgage rates obtained by blacks who refinanced is consistent with the

findings from a large number of studies that have found that blacks are much more likely than other

racial-ethnic groups to use subprime lenders (Pennington-Cross, Yezer, and Nichols, 2000;

Scheessele, 2002; National Community Reinvestment Coalition, 2003; Calem, Gillen, and Wachter,

2004; and Calem, Hershaff, and Wachter, 2004). It is telling that the Boehm, Thistle, and

Schlottmann and Susin find that the disparities in mortgage interest rates between whites and blacks

are not evident among purchase mortgages. This result is consistent with the fact the growth of

subprime lending during the 1990s was most evident among refinance loans and much less evident

among purchase loans—at least until very recently.

Boehm, Thistle, and Schlottmann also use their estimated model to disaggregate the reasons for

blacks’ higher interest rates into portions due to differences in borrower characteristics versus

differences in treatment in the market associated with race that cannot be attributed to other borrower

characteristics. They find that 87 percent of the difference in black-white refinance interest rates is

due to different treatment in the market and only 13 percent is due to differences in borrower or loan

characteristics. They note that some of the unexplained racial difference may be due to differences in

credit history, a factor that is not captured by the AHS.

Nothaft and Chang’s estimates of the value of missed refinancing opportunities do not take into

account differences in interest rates obtained by income or race-ethnicity, but the disparities found by

the above studies suggest they could be substantial. Carr and Schuetz (2001) present calculations

showing that every additional percentage point added to a 30-year mortgage increases the total

interest paid over the life of the mortgage by at least $20,000. If on average all black owners who

refinance pay about one percentage point higher rates than whites, the total aggregate costs of these

higher rates would be several multiples of Nothaft and Chang’s estimated cost of $22 billion in lost

benefits from the 16.5 percent of black owners who did not refinance.

Tapping Home Equity Through Cash-out Refinance or Home Equity Loans

As noted above, another issue to consider regarding mortgage finance decisions is the extent to which

owners reduce their equity in their homes either through cash-out refinancing or home equity loans.

Exhibit 13 illustrates the prevalence of these activities by homeowner income, race-ethnicity, and

first-time owner status. Both cash-out refinancings and home equity loans are more than twice as

common among moderate- and high-income owners compared to low-income owners. Home equity

loans are much more common among all groups. While 1.6 percent of low-income owners took cash

out of their homes through a refinanced primary mortgage, 8.1 percent had a home equity loan in

place in 2003. In comparison, among moderate- and high-income owners, 3.1 and 3.7 percent,

respectively, had taken cash out through a refinancing while 15.3 and 21.2 percent, respectively, had

home equity loans outstanding.

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

As Exhibit 13 also shows, blacks were less likely than whites to have either taken cash out through a

refinancing (1.7 versus 2.7 percent) or have a home equity loan (8.4 versus 15.5 percent). Hispanics

were also less likely to have a home equity loan than whites (8.7 versus 15.5 percent), but they were

slightly more likely to have taken cash out through refinancing (3.2 versus 2.7 percent). Once again,

first-time buyers are much less likely to tap home equity than any other group of owners, no doubt

reflecting the fact that they have less equity in their homes.

Exhibit 13

Refinance Activity and Reasons for Refinancing by Income and Race-Ethnicity of

Owners

Low

Income

Moderate

Income

High

Income White Black Hispanic

First-time

Buyers

Share of All Owners with:

Cash Out Refinance 1.6% 3.1% 3.7% 2.7% 1.7% 3.2% 0.3%

Home Equity Loan 8.1% 15.3% 21.2% 15.5% 8.4% 8.7% 7.8%

Note: Reasons for refinancing sum to more than 100 percent as more than one reason can be reported.

Source: 2003 American Housing Survey.

The differences in the propensity to tap home equity through borrowing evident in these simple cross

tabulations are supported by several recent studies that have used multivariate techniques. Using the

AHS, Nothaft and Chang (2004) estimate models predicting the incidence of both cash-out

refinancings and the use of second mortgages to draw down equity. In addition to income and race-

ethnicity, these models also control for the loan-to-value ratio, the size of the primary mortgage, and

the borrowers’ payment-to-income ratio. Higher income households are generally more likely to tap

home equity both through cash-out refinancing and second mortgages. First-time homebuyers are

less likely to use either type of financing to draw down their equity. With regard to minorities,

Nothaft and Chang find that “other” minority households are less likely to use a refinance to take cash

out, while blacks are no different than whites in this regard. Blacks are, however, much more likely

than whites to take out a second mortgage, while “other” minorities are no different than whites.

Finally, Canner et al. (2002) also examine the tendency for owners to take cash out through a

refinance. They do not show results for the probability of taking cash out, but instead indicate the

association between borrower and loan characteristics and the amount of cash taken out. The single

most important factor is found to be the race-ethnicity of the borrower, with minorities taking out

much less cash out than whites—$5,537 less on average. Homeowner income is not found to be as

important as race, with those whose income is below $40,000 found to take out $1,847 less on

average.

In short, existing studies suggest that low-income and minority owners are less likely to tap home

equity through cash-out refinancing. Thus, it does not appear to be the case that these owners are

drawing down their home equity more rapidly than other owners. Of course, tapping home equity is

not necessarily a bad outcome. One of the benefits of homeownership is that the accrued equity in the

home can be used to finance investments in businesses or education or be drawn upon to meet

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

emergency needs. The concern is that low-income and minority households will have a greater need

to tap equity to meet more routine consumption needs, which will reduce the availability of wealth for

these other purposes.

To the extent that some owners are tapping their equity, it would be interesting to know how the use

of proceeds from cash-out refinancing or home equity mortgages differs by income and race-

ethnicity. But, unfortunately, there is little information available on this topic. Canner et al. (2002)

report on the differences in the use of cash taken out among all owners and find that the most

common use is the repayment of other debts (51 percent), followed by home improvement (43

percent), and consumer expenditures (25 percent). More rarely, owners are found to use proceeds to

make investments in stocks or other financial instruments (13 percent) or real estate and businesses (7

percent). The AHS only reports on the proportion of cash used for home improvements. Low-

income households are slightly less likely to use cash out proceeds for home improvement expenses

compared to moderate- and high-income owners (57 percent versus 62 percent, respectively). With

regard to race-ethnicity, Hispanics (67 percent) are more likely to use cashed out funds for home

improvement than either whites (60 percent) or blacks (56 percent).

One study that provides some indication of the extent to which home equity is tapped for other

purposes is Haurin and Rosenthal (2005c). This study employs a different approach than the studies

discussed above to estimate the extent to which homeowners tap capital gains for other purposes.

Their basic approach is to predict levels of total household debt and non-housing assets as a function

of a variety of household characteristics, including changes in the value of their home. The

coefficient on the variable measuring the change in housing value is interpreted as the share of the

gain that is evident in increased household debt or non-housing assets. An important difference from

the studies focusing on mortgage debt is that this approach allows for the fact that increased housing

wealth may encourage owners to take on non-mortgage debt to finance purchases. The difference

between the change in debt levels and the change in non-housing assets is interpreted as the share of

debt spent on non-durable goods that do not add to household assets. The focus of the study is on

evaluating differences by homeowner income level in the propensity both to tap gains in house values

and to use these proceeds either to invest in other assets or to support consumption of non-durable

goods.

The study uses two different data sets, including pooled observations from the Survey of Consumer

Finance (SCF) from 1983 through 2001 and the National Longitudinal Study of Youth (NLSY) from

1980 through 2000. The authors conclude that low-income and minority households have a

somewhat higher propensity to tap capital gains, as for each dollar in gain the amount of household

debt among these owners increases by between 12 and 18 cents, compared to an increase of 8 to 17

cents for high income households. This conclusion differs from the studies focusing solely on

mortgage debt, which found that low-income households and minorities were less likely to borrow

against their homes. The results suggest that these owners may be more likely to use non-mortgage

debt as a means of tapping home equity.

One recent study (Hurst and Stafford, 2004) has found that owners are more likely to have taken equity out

of their homes through refinancing if they experienced a spell of unemployment and also had low levels of

liquid financial assets. However, the study did not explore whether there were any differences in these

tendencies by income or race-ethnicity of the borrower.

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

One of the key findings of this study is that there are important differences by income in how owners

use their housing wealth. High-income households appear to spend most of their gains on non-

durable goods as they experience little increase in the value of non-housing assets. For low- and

moderate-income households, on the other hand, much of the increased debt is associated with an

increase in the value of non-housing assets, with their estimates ranging from 11 to 15 cents. In short,

while high-income households are more likely to tap housing equity for consumer expenditures, low-

and moderate-income households are more likely to use their gains to finance the purchase of other

durable goods.

While the results from the SCF and NLSY were fairly consistent for both moderate- and high-income

households, there were some important variations in Haurin and Rosenthal’s results for low-income

owners. The SCF results found no evidence that low-income owners increased either their debt levels

or their non-housing assets in response to gains in house values. One explanation for this result is

that the low-income owners in the SCF sample were much older than the low-income owners in the

NLSY, with an average age of 60 compared to 35 in the NLSY. This suggests that older low-income

households are less likely to tap housing equity than younger ones, a result that is consistent with

Skinner (1993).

One important caveat regarding the findings of most of these studies is that there has been a

significant rise in recent years in the amount of home equity being extracted through refinances. Data

from Freddie Mac for conventional conforming loans shows that the amount of home equity

withdrawn through refinances more than doubled between the early and late 1990s, increasing from

less than $20 billion annually before 1997 to about $40 billion in 1999. After a decline in 2000, the

amount of cash being taken out then rose to $83 billion in 2001, $111 billion in 2002 and $147 billion

and $140 billion in 2003 and 2004, respectively.

The rise in cash out refinancing seems to be driven

by a combination of rapid growth in housing values, continued low interest rates, and increases in

consumer debt generally (Canner et al., 2002). Most of the studies cited above rely on data from

2001 or earlier, which is before the peak of this cash-out boom. Canner et al. come closest to

capturing this cash-out boom, but their study period of 2001 and the first half of 2002 is also before

the peak. Thus, these studies do not shed light on the extent to which this sharp rise in cash out

refinances has been distributed across different types of homeowners. But the figures cited from the

2003 AHS above indicate that higher income households were much more likely to have taken cash

out.

Investing in Home Maintenance and Improvements

Another important choice that borrowers face after purchasing a home is whether and how much to

invest in maintenance and improvements. These investments are important for several reasons. A

certain level of investment in the house is needed to counter the effects of depreciation to protect the

owner’s investment. Deferred maintenance can lead to larger problems and have significant impacts

See http://www.freddiemac.com/news/finance/docs/cashout_volume.xls. Since these figures exclude

government insured, jumbo, and subprime mortgages, they undoubtedly underestimate the total volume of

cash taken out, but the trends are nonetheless instructive.

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

on the home’s value. For example, ignoring a leaking roof can lead to substantial damage to both the

structure and the interior of the home. Aside from being necessary to maintain the value of the home,

another reason why maintenance expenditures are of interest is that they can also add to the financial

burdens of homeownership. If low-income and minority households purchase older homes that

require greater levels of ongoing maintenance, this will increase their total housing costs.

Investments in home improvements beyond routine maintenance also help to support increases in

home values over time. Analysis of the AHS by the Joint Center for Housing Studies (2005) found

that greater investments in home improvement over the period from 1995 to 2001 were strongly

associated with larger increases in home values over the same time.

However, while owners do

realize some return on their investments in home improvements, on average the increase in the value

of the house is less than the cost of the improvement. Industry estimates of the return on remodeling

investments indicate that 80 percent of the costs of typical improvement projects are recouped in

increases in home values.

The disparity between the cost of home improvements and the increase in home values that result

from these efforts highlights that, while investments in the home can have important financial

implications, for the most part owners choose to make these investments out of a desire to tailor their

homes to suit their tastes. In this regard, investments in home improvements may also be important

as an indication that low-income and minority owners are able to maximize the benefits they receive

from their homes. On the other hand, as long as the home is in sound condition and adequately

provides for the needs of the household, improvements may best be categorized as a luxury rather

than a necessity. That is, a failure of low-income and minority owners to invest in their homes to the

same degree as upper income or white owners may not be an indication that homeownership fails to

provide them with sufficient benefits.

In short, decisions by low-income and minority owners about investments in property maintenance

are probably most important to both protect their investment in the house and to ensure that the home

provides an adequate living environment. The extent to which these households choose to invest in

improvements to the property may provide some indication of their ability to tailor their homes to

meet their needs, but in many cases these investments may not be necessary to meet the fundamental

housing needs of the owners.

There is very little research that has examined the experience of low-income owners in having to

make investments in maintenance and repair to their homes. One study that provides some

investigation of this issue is Rohe et al. (2003). This study surveyed low-income participants in

homeownership counseling programs offered through affiliates of the Neighborhood Reinvestment

Corporation in eight locations around the country. The survey, conducted about 18 months after

participants received counseling, asked those who had purchased homes about their experience with

unexpected major costs associated with the house and whether there were any needed repairs that the

owner had been unable to afford. The responses from 343 homebuyers suggest that both of these

issues are fairly common among the low-income buyers participating in these programs. Almost half

As the Joint Center notes, the causality in this association goes both ways. That is, it is also true that rising

housing prices provide owners with equity that can be tapped to fund these home improvements.

“2004 Cost Vs. Value Report,” REMODELING Magazine, November 1, 2004.

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

(48 percent) of respondents indicated that they had experienced a major unexpected cost, with the

most common problem being a repair needed to one of the home’s major systems. A little more than

a quarter of buyers (28 percent) also reported having a needed repair that they were unable to afford,

most commonly including problems related to the roof, foundation, or major systems. The survey

results suggest that home maintenance issues may be fairly common among the low-income buyers

assisted by homeownership counseling programs.

While there is little work with an explicit focus on low-income homeowners, there is a fair amount of

literature that has evaluated the factors that are associated with an owner’s decisions about whether to

invest in home improvements and, if so, how much to invest. In general, this research has found that

low-income and minority households are less likely to make improvements and, when they do, their

investments are smaller (Mendelsohn, 1977; Boehm and Ilhandfeldt, 1986; Montgomery, 1992;

Harding, Miceli, and Sirmans, 2000; and Baker and Kaul, 2002). Tabulations of the 2003 AHS by

the Joint Center for Housing Studies (2005) provide some indication of the magnitude of these

differences. Among homeowners with income under $40,000, 23 percent are found to undertake a

home improvement project, compared to 30 percent of owners with incomes over $120,000.

However, since elderly homeowners account for a disproportionate share of these lower-income

owners, in large part these differences reflect the tendency of older owners to make more limited

investments in their homes. A comparison of the remaining income categories shows little difference

in the proportion that invest in their homes. Among those with incomes between $40,000 and

$80,000, 28 percent invested in home improvements, which is identical to the share of owners with

income between $80,000 and $120,000 and only slightly lower than the 30 percent share among those

with income above $120,000. A similar pattern is evident among owners by race-ethnicity; while

blacks are less likely to invest in their homes, the difference from whites is small. In 2003, 27 percent

of white owners undertook a home improvement project, compared to 25 percent of blacks.

Hispanics were actually slightly more likely than whites to invest in their homes, with 28 percent

undertaking a home improvement project.

While the share of owners investing in their homes is similar, there are larger differences in the

average amounts spent. Among owners with incomes between $40,000 and $80,000, the average

home improvement project was $5,300, compared to $8,900 for those with incomes between $80,000

and $120,000, and $14,200 for those with incomes above $120,000. Whites also spent more on

average than either blacks or Hispanics. The average white homeowner spent $7,300, compared to

$4,400 among blacks and $5,500 among Hispanics.

It is important to note that, for the most part, the Joint Center tabulations and the literature cited above

do not include expenditures for routine home maintenance. These types of activities include repairs

and maintenance to the home’s existing systems as well as cosmetic improvements, such as painting.

Investments in maintenance and repairs can be substantial. In 2001, homeowners spent a total of

$131.5 billion on home improvement and $34.3 billion on home repairs (Joint Center for Housing

Studies, 2004). As noted above, expenditures on home maintenance may be a more important factor

in evaluating whether low-income and minority homeowners are sustaining their investment,

continuing to occupy adequate housing, and are not facing undue cost burdens from high maintenance

expenditures.

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

Of those cited above, Montgomery (1992) is the only one that addresses whether there may be

differences in the factors associated with investments in home improvements versus maintenance. In

an analysis of data from the 1985 AHS, she estimates models predicting both the prevalence of home

improvement activities and the amount spent where home improvement is alternatively defined to

include or exclude home maintenance activities. Her results indicate that there is little difference in

the factors predicting if an investment is made whether maintenance is included or excluded. In both

cases, higher-income and whites households are more likely to undertake these activities. However,

there are some significant differences in her results regarding the amount spent. While household

income is still positively and significantly associated with expenditure levels, the size of the income

coefficient is smaller when maintenance activities are included. More importantly, the association

between expenditure levels and race-ethnicity switches when maintenance expenditures are included.

Without maintenance expenditures, whites are found to spend more; with maintenance expenditures

included, minorities are found to spend more (although the coefficient is only significant at the 10

percent level).

There are several possible explanations for this result. One reason might be that when minorities do

have a need to spend on home maintenance, the problems are more severe and so require greater

expenditures. However, another reason might be related to the fact that minorities are more likely to

hire a professional contractor for home improvement projects and less likely to undertake these

projects themselves—so called “do-it-yourself” or DIY activities. Two studies have examined the

factors associated with the decision of a homeowner to undertake a home improvement project

themselves or to hire someone to do the work (Mendelsohn, 1977; and Bogdon, 1996). Both of these

studies find that while lower-income households are more likely to engage in DIY, minorities are less

likely to do so.

The finding that low-income households are more likely to engage in DIY is consistent with the

notion that these households both have lower costs of forgoing paying work to undertake these

projects and also have less income to be able to hire someone. Since minorities tend to have lower

incomes than whites, it might have been expected that they would also be more likely to engage in

DIY activities. Yet, controlling for income and other housing and household characteristics, this does

not appear to be the case. Bogdon speculates that one reason why minorities might not be as likely to

engage in DIY is that they are less likely to have grown up in an owner-occupied home and so have

had less opportunity to develop the skills needed for these efforts. She also posits that low-income

households may have greater difficulty in undertaking DIY efforts if they have to work long hours to

compensate for lower hourly wages.

Again, tabulations of the 2003 AHS by the Joint Center for Housing Studies (2005) shed some light

on the variation by income and race-ethnicity in the prevalence of DIY versus hiring professional

contractors for home improvement projects. Across the four income categories defined by the Joint

Center, the lowest share of DIY projects for those undertaking some home improvement was among

those with incomes under $40,000 (48 percent) and those with income over $120,000 (47 percent). In

contrast, 59 percent of those with income between $40,000 and $80,000 and 57 percent of those with

income between $80,000 and $120,000 engaged in DIY activities. As before, the low shares of DIY

among the lowest income category may well reflect a higher share of elderly homeowners in this

category, who are much less likely to engage in DIY. Focusing on the next two lowest income

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

groups, the data do support the idea that lower income households are less likely to hire contractors

than the highest income group.

The Joint Center figures also show that while 53 percent of whites undertake DIY projects, only 42

percent of blacks do. Hispanics are actually 8 percentage points more than whites to undertake home

improvement efforts. One explanation for the differences between Hispanics and blacks may be that

Hispanic immigrants come from countries where it is very common for households to construct their

own housing (so called self-help housing). Another factor might be that Hispanic owners are even

more likely to be in married couples households than whites, and so have more adults in the

household to support DIY activities.

Bogdon suggests that to the extent that lower DIY activity is the result of less knowledge of how to

undertake these projects, the obvious policy response would be to make training available for new

homeowners to develop these skills. If in fact the lower level of DIY activity among blacks is

indicative of less ability to undertake repairs and improvements to the home, it also raises concerns

that these owners are deferring routine maintenance, potentially increasing the magnitude of these

problems. This issue further highlights the lack of research that provides a good understanding of the

maintenance needs of low-income homeowners, particularly blacks, and how they respond to these

needs.

A final issue regarding maintenance and improvement expenditures that has not been the subject of

much study is the relationship between neighborhood conditions and an owner’s decision to invest in

their home. In theory, poor neighborhood conditions would deter investment in the home since the

owner would be less likely to recoup their investment in a declining area. As shown in Chapter 2, for

the most part low-income and minority first-time homebuyers do not locate in distressed areas, but,

nonetheless, conditions are somewhat worse than they are for whites and higher-income buyers.

While most studies of improvement activities include controls for region and whether the home is in a

central city or suburban area, few have attempted to capture variations in neighborhood conditions.

Of those that have, Boehm and Ihlanfeldt (1986) find that owners in areas with high crime rates and

higher shares of surrounding buildings with structural defects are less likely to invest in their homes.

On the other hand, Montgomery (1992) includes the AHS measure of neighborhood quality but finds

that it is only weakly correlated with home improvement activities. When there is a significant

association, she also finds that higher quality neighborhoods are associated with less investment in the

home, a finding that is contrary to expectations. One study that directly assesses this issue is

Ioannides (2002), which analyzes the association between spending on home improvement by other

homeowners in the immediate vicinity on a homeowner’s own improvement spending. He finds a

strong positive impact of neighbors’ investments. But the study does not identify whether this effect

varies by the characteristics of the neighborhood. The question of whether investment in home

upkeep and improvement varies with neighborhood condition is not well studied.

Summary

Many of the financial and social benefits of homeownership are associated with residential stability.

Longer-term stays in the same house avoid the high transaction associated with selling a home, allow

owners to ride out short-term market downturns, and give families an opportunity to develop and

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

benefit from social networks in the community. Given the importance of residential stability, a

critical issue in evaluating whether low-income and minority owners realize the benefits from

homeownership is whether they are able to sustain homeownership over time.

One measure of stability is simply the rate at which owners move out of their homes, which includes

moves to other owner-occupied situations as well returns to renting or living with others. The main

implications of greater mobility relate to the expected financial returns to homeownership, since more

frequent moves mean greater costs of buying and selling homes and more chance to sell into down

markets. There is a rich literature on residential mobility, which indicates quite clearly that owners

move much less frequently than renters. However, there are surprisingly few studies that address the

question of whether there are differences in mobility rates among owners by income or race-ethnicity.

Of the few studies that have been done, most suggest that higher income households are more likely

to move, which is in keeping with the view that these households have greater choices in the housing

market. These same studies also find that minorities are generally more likely to move. Since

minorities are unlikely to have more housing choices than whites, the higher mobility likely reflects

greater difficulty in maintaining homeownership. Of note, one study comes to opposite conclusions,

finding that low-income and white households are both more likely to move. In short, the mobility

literature does not provide a clear picture of the extent of differences in the mobility rates of owners

by income and race-ethnicity.

The most obvious and costly failures to sustain homeownership are cases that end in foreclosure.

There is no question that low-income and minority owners are more likely to experience a foreclosure

than other owners. But that still leaves the question of at what point is the rate of foreclosure too

high. Among prime borrowers with at least 10 percent equity in their homes, foreclosure rates among

low-income and minority borrowers are fairly low (less than one in twenty owners) and the difference

from other owners are small (a few percentage points at most). However, simulations indicate that

foreclosure rates rise by several percentage points when owners have no initial equity in their homes.

If no-downpayment loans are combined with sustained periods without any house price growth,

foreclosure rates are estimated to rise to about one in ten borrowers, with low-income borrowers

about 50 percent more likely to face foreclosure than other owners. But while prime borrowers are

only projected to experience rates this high in relatively extreme circumstances, foreclosure rates of

one-in-ten were evident during the 1990s among minority borrowers using FHA insurance. Even

higher rates are likely to occur among subprime borrowers.

But even in these high-risk lending programs, the vast majority – on the order of 90 percent – of low-

income and minority owners do not experience a foreclosure. Thus, while foreclosure rates of one-in-

ten are a cause for concern, it does not mean that lending to these groups in general has gone too far.

However, it does highlight the need to identify the borrowers who are at greatest risk of foreclosure

both to see whether financing ought to be limited to these groups and to develop approaches for

providing more support for these owners over time to avoid foreclosure. In fact, during the 1990s

there was a great deal of innovation in loan servicing approaches both to improve lenders’ rates of

contacting delinquent borrowers and to offer these borrowers a range of options for curing their

problems short of foreclosure. These innovations in loss mitigation efforts were widespread in both

the conventional prime market and at FHA (Cutts and Green, 2004; Herbert et al., 2000). Given the

continued growth of subprime lending, particularly in the home purchase market, the implementation

of concerted loss mitigation efforts in this market segment is clearly needed. Recently, there have

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

been examples of innovative efforts by subprime lenders in the Chicago area to help reduce

foreclosures (Neighborhood Housing Services of Chicago, 2004).

But foreclosures are only the most visible way in which owners fail to maintain homeownership.

Several recent studies have examined panel studies tracking households over time to examine how

long low-income and minority first-time owners are able to stay in their first homes before returning

to renting or living with others. One unexpected finding from these studies is that exits from

homeownership are actually fairly common among all first-time buyers, with 40 percent of buyers

found to no longer own their residence five years after they first purchased. Importantly, these

studies also find that low-income and minority first-time buyers are more likely to leave

homeownership than other buyers. Estimates from two studies focusing explicitly on first-time

buyers differ somewhat, but they find that between 43 and 53 percent of low-income buyers will not

sustain homeownership for more than five years, compared to between 23 and 30 percent of high-

income buyers. They also find that minorities at all income levels are between 22 and 39 percent

more likely to leave homeownership than whites.

These studies do not identify why these households leave homeownership. The exits will include

cases where a change in circumstances make it impossible for owners to sustain homeownership—

including both instances where owners are forced out of their homes by foreclosure as well as

instances where owners leave to avoid foreclosure. But some share of these exits are also voluntary

and likely represent sound decisions given changes in household circumstances that impose few costs

on the owners. Assuming that high-income owners are only rarely forced out of homeownership by

an inability to meet their financial obligations, then the exit rate for these households of between 23

and 30 percent may be taken as an indication of the share of cases where households voluntary leave

homeownership within five years. This assumption would suggest that roughly a fifth of low-income

owners generally and about a quarter of low-income minority owners are forced out of

homeownership within five years of buying. Clearly, it is not at all a given that low-income first-time

buyers will be able to sustain homeownership long enough to realize its benefits.

These studies find that income loss, unemployment, and divorce are all strong predictors of who is

likely to leave homeownership. One study that included measures of mortgage interest rates both at

purchase and over time found a strong association between higher interest rates and shorter spells as

an owner, suggesting that higher housing costs are also an important factor. Aside from household

income and race-ethnicity, another important household characteristic is whether the owners are a

married couple. Households headed by a single adult are found to have a much greater chance of

leaving homeownership. This may well reflect the fact that with only one adult in these households

they have less ability to accommodate a loss of income, an illness, or other unexpected crisis.

While the share of low-income and minority owners leaving homeownership seems high, it is also

important to consider that not all moves out of homeownership are involuntary or associated with

financial losses. As noted above, the fact that a fairly high share of high-income first-time buyers

leaves homeownership within five years suggests that a lack of affordability is not the only factor that

precipitates a return to renting or living with others. One study tracked households after they left

homeownership to see if they subsequently returned to owning. While the study did not report results

by household income, they did find that a majority of first-time buyers in all racial-ethnic groups that

left homeownership ultimately succeeded in purchasing a second home. Among whites, 86 percent of

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

those who become owners either continually sustain ownership or return to owning, compared to 81

percent of blacks and 84 percent of Hispanics. These figures suggest that, despite fairly high shares

of households leaving homeownership at some point after they buy their first home, ultimately a large

majority of first-time buyers of all racial-ethnic groups do succeed as owners.

While these papers examining the length of time households sustain homeownership provide valuable

insights into the homeownership experience of low-income and minority owners, they also highlight

how little we know about the dynamics of an individual buyer’s homeownership experience over

time. Specifically, it would be very helpful to know more about the changes that occur in household

circumstances after home purchase, how different households respond to these changes (including

whether they can draw upon savings, debt, insurance, or resources provided by family and friends),

and how these responses are associated with different outcomes. A better understanding of these

dynamics would allow policy makers to identify which households are at greatest risk of being unable

to sustain homeownership and what types of support are needed to increase their chances of success.

Another important set of choices made by owners after purchasing their homes relates to the use of

mortgage finance. One important benefit available to owners is to take advantage of declines in

mortgage interest rates to refinance their mortgage and lower their housing costs. Studies that have

examined the propensity of lower-income and minority homeowners to refinance have found that

both groups are less likely to refinance during periods when interest rates decline. For low-income

owners these differences appear to be completely explained by the fact that these borrowers are

constrained from refinancing by their higher risk profile in terms of loan-to-value ratios and credit

scores. For minorities, some of the difference from whites is explained by minorities having higher

risk factors, but even after controlling for these factors they are less likely to refinance than whites.

One study estimated that in present value terms low-income and minority homeowners each lost in

aggregate about $21 to $22 billion in reduced interest costs over the life of their mortgages from

missed refinancing opportunities.

For blacks, another troubling issue is that when they do refinance, on average they obtain interest

rates that are about one percentage point higher than whites. Interestingly, these studies of racial

disparities in interest rates also find no significant difference in rates paid by blacks at the time of

purchase. This result is consistent with the findings of a range of studies that have found blacks to be

much more likely than whites or other minorities to use subprime lenders to refinance their primary

mortgages. While differences in credit risk may explain some of this difference, a variety of studies

that have included measures of risk still find that blacks are more likely to use subprime lenders, all

else equal. Given that each additional percentage point on a mortgage of about $80,000 raises total

interest paid over the life of the loan by at least $20,000, the additional costs imposed on black

owners by unnecessary subprime refinances could be substantial. The tendency for blacks to fare

worse in mortgage refinance markets than in purchase mortgage markets may be an indication that

affordable lending programs need pay greater attention to the needs of existing owners rather than

focusing most of their attention on efforts to support initial home purchases.

Another important aspect of mortgage finance choices relates to owners’ choices to tap accumulated

home equity for other purposes. Since one of the benefits of homeownership is that owners can use

accumulated wealth to support investments in education and businesses or to weather financial crises,

the use of home equity is not necessarily a bad outcome. However, one concern is that low-income

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

and minority homeowners might be more likely to use home equity to support current consumption

and so reduce the availability of wealth for these other purposes. Accumulated home equity also

helps to lower long-run housing costs, so if owners continually ratchet up their debt, they will not

realize this long-run benefit. Innovations in mortgage markets during the 1990s have made it cheaper

and easier to tap home equity through either cash-out refinances or home equity loans, raising the

chances that owners will draw down their home equity. Studies examining the factors associated with

the use of mortgage debt consistently find that low-income and minority owners are less likely to tap

their home equity either through cash-out refinancing or home equity loans. One exception to this

pattern is that Hispanics are slightly more likely than whites to take cash out through refinance. Thus,

in general, there does not appear to be concerns that these borrowers are more likely to draw down

their home equity. One study that examined changes in total household debt did conclude that low-

and moderate-income households were more likely to increase their debt in response to gains in house

values, but it also found that these owners were highly likely to invest in durable goods that increased

the value of household assets. In contrast, high-income households mostly tapped housing gains to

support consumption of non-durable goods that did not add to the household balance sheet.

A final aspect of the homeownership experience examined in this chapter dealt with investments in

home maintenance and improvements. These investments are important in two respects. First,

spending on home maintenance is important to ensure that the home continues to provide adequate

shelter and to protect the value of the home. Required home maintenance spending can also add to

housing costs and may increase the financial burden on low-income owners. Second, investments in

home improvement increase the benefits received by owners from their homes and help to support

increases in value. In general, studies have found that low-income and minority households are less

likely to invest in home improvements, although the differences are not large. However, existing

studies mostly focus on home improvements and shed little light on the incidence of maintenance

expenditures. Since these activities are arguably more important for preserving the home, more

research on differences in maintenance activities by income and race-ethnicity would be interesting.

One interesting finding from studies of home improvement activities concerns the tendency of owners

to either do the work themselves versus hiring professional contractors. While low-income

households are found to be more likely to do it themselves, minority owners, particularly blacks, are

found to be more likely to hire contractors. There are several possible explanations for this pattern.

One is that since blacks are less likely to have grown up in owner-occupied households, they may

have less experience with these types of tasks. To the extent that this is the case, one implication is

that blacks will face higher costs for home maintenance than other owners, which was found by the

one study that examined patterns of home improvement spending including maintenance

expenditures. Another implication is black owners may also be less likely to handle routine

maintenance tasks, which will increase the chances that more serious problems will result. An

obvious response to this situation would be make training available for new owners to develop the

skills needed to handle basic home maintenance and improvement tasks.

Many of the issues touched upon in this chapter have important repercussions for the financial return

realized by low-income and minority homeowners. The next chapter explores this issue in depth.

Chapter 3: Key Experiences and Decisions of

Low-Income and Minority Homeowners

Chapter Four:

Financial Benefits of Homeownership

One of the fundamental reasons that policy makers have made it an important goal to increase

homeownership opportunities for low-income and minority families is because it is believed to be the

primary means of wealth creation for many U.S. households. The most important ways in which

homeownership generates wealth is through appreciation in the value of the home and amortization of

mortgage debt over time. Secondary financial benefits, which can also contribute to wealth creation,

include lower annual housing costs both due to the income tax benefits associated with the

deductibility of mortgage interest and property taxes and to the protection that homeownership

provides against rising housing costs.

Arguments about the importance of homeownership as a means of wealth creation generally point to

the large share of household wealth that is held in housing, particularly among low-income and

minority households. In 2000, equity in homes accounted for 56 percent of aggregate wealth among

households with income in the lowest quintile, compared to 32 percent of aggregate wealth among all

households. Housing wealth was even more important among blacks, accounting for 62 percent of

their aggregate wealth (Orzechowski and Sepielli, 2003).

But while faith in the financial benefits of homeownership is strong, the literature examining owner-

occupied housing from a financial investment perspective raises significant questions about whether

housing is, in fact, the best investment a family can make. Chief among the concerns about housing

as an investment is the risk of experiencing weak or even negative appreciation in house values. Over

the past 30 years there have been a number of markets that have experienced significant booms and

busts in housing prices, exposing owners who are forced to move at the wrong time to potentially

significant financial losses. In other markets, the long-run appreciation in house prices has not kept

pace with inflation. In these areas, owners may well have been better off financially if they had

invested in a balanced portfolio of stocks and bonds. Further adding to the risk of homeownership are

the significant transaction costs associated with buying and selling homes.

To the extent that homeownership is, in fact, a good investment for the average household, there is

still the question of whether it is a good investment for low-income and minority households. Since

these households are buying lower-valued homes in different neighborhoods than higher-income and

white households they may not have the same likelihood of experiencing appreciation in home values.

Another important difference is that low-income households are less likely to benefit from favorable

income tax provisions since the standard deduction is more likely to exceed their deductible housing

costs. Finally, as the literature reviewed in Chapter 3 showed, low-income and minority first-time

buyers have a greater likelihood of leaving homeownership after a few years and so are less likely to

be able to ride out the ups and downs in housing prices.

In this chapter we review the literature that has examined both the general question of whether

homeownership is a good financial investment as well as what is known about differences in the

likelihood of realizing financial benefits for low-income and minority owners. The chapter begins by

reviewing the literature that has examined the role of housing appreciation in the financial return to

Chapter 4: Financial Benefits of Homeownership

homeownership and whether expectations regarding appreciation differ for homes purchased by low-

income and minority homebuyers. The next section then examines the importance of income tax

benefits in making homeownership financially attractive and how limitations on the ability of low-

income owners to take advantage of these benefits affect their financial return. The last section takes

a more comprehensive look at the role of homeownership in wealth creation. There is a complex web

of factors that contribute to the financial returns from homeownership, including appreciation in

values and tax benefits as well as the issues raised in Chapter 3 about sustaining homeownership over

time and decisions about financing, maintenance, and improvements. The literature reviewed in the

last section attempts to account for all of these factors by using longitudinal surveys of households

over time to track trends in the tenure choices of households and how these cumulative tenure choices

are related to wealth accumulation. The chapter closes with a summary and conclusions from the

literature reviewed.

Housing Appreciation and the Financial Return to Homeownership

Appreciation in house values is critical to the realization of financial returns to homeownership.

Given the focus of this study, we are primarily concerned with the question of whether low-income

and minority owners are as likely as other owners to benefit from housing appreciation. But before

turning to the question of what is known about how appreciation rates may vary by income or by the

race-ethnicity of the owner, it is important to consider the general question of how financial returns to

housing compare to other investments. In assessing the financial returns to housing, Goetzmann and

Spiegel (2002) take the straightforward approach of estimating the average nominal change in

housing values in 12 markets based on a repeat sales index covering the period from 1980 to 1999.

They then compare these rates of return to after-tax mortgage costs over the same period as proxied

by the annual return on mortgage-backed securities. They find that in 9 of 12 markets examined

house price growth was lower than the after-tax mortgage costs so the returns to housing over this 20-

year period were negative. From this result they conclude that housing offers lower returns than

Treasury bills and note that given this poor performance it is “surprising that housing continues to

represent a significant portion of American household portfolios.”

However, Goetzmann and Spiegel’s analysis does not take into a number of important considerations,

including the value of the housing services provided by the home (also referred to as imputed rent),

the full tax advantages of home ownership (including the ability to deduct property tax payments), or

the fact that investments in homes are generally highly leveraged through mortgage debt. The use of

leverage means that while the rate of appreciation in the overall house value may be fairly low, since

many owners have a fairly small amount of equity in the home but realize appreciation on the full

value of the property, even small increases in value can represent a high rate of return on the invested

equity.

Studies that do take a more comprehensive view of the financial returns to homeownership come to a

much more positive conclusion about the investment appeal of owner-occupied housing. Flavin and

Yamashita (2002) incorporate the value of housing services, a more complete estimate of the income

tax benefits, and estimates of other costs of homeownership (specifically, property taxes and

depreciation). They do not, however, incorporate the effect of financial leverage. They estimate

returns to housing using two methods—one using owners self valuations of their homes from the

Chapter 4: Financial Benefits of Homeownership

Panel Survey of Income Dynamics for the period 1968 to 1992 and another using a repeat sales

housing price index for four markets during the period from 1970 to 1986. By either method, they

find that the financial returns to homeownership are lower than the returns to stocks, but much higher

than the returns from either risk free investments or investments in mortgages. Flavin and Yamashita

also estimate the variance of the returns for individual owners and find that the variance of returns to

housing are higher than risk-free investments but lower than for stocks, which is in keeping with the

relative rates of return of these different investments. Finally, they also find that the correlation of

housing returns with stocks, bonds, and treasury bills is essentially zero—so homeownership helps to

diversify the risks of other investments.

Case and Shiller (1990) examine the same repeat sales index as Flavin and Yamashita covering the

Atlanta, Chicago, Dallas and San Francisco markets over the period from 1970 to 1986. Like Flavin

and Yamashita, their analysis of the financial returns to owner-occupied housing includes income

taxes, maintenance, and the value of housing services. But in addition they also estimate returns

alternatively assuming the home is owned without debt and with a mortgage equal to 80 percent of its

value. Case and Shiller then estimate the difference between the financial returns to housing and the

returns earned by 90-day Treasury bills, defining this difference as the excess return to housing above

a riskless investment. They conclude that in these four markets over the period studied the excess

returns on owner occupied housing were enormous in all four markets, particularly when the effect of

leverage is included.

These results are confirmed by Goetzmann (1993) who presents a similar analysis for the same four

markets. Over the period studied, house price appreciation rates were high in all four markets, so that

even when the estimated returns ignore the combined contribution of tax benefits, the value of

housing services, and financial leverage, the returns are found to compare favorably with other

investments taking into account differences in risk. Importantly, however, when Goetzmann does

include estimates of the tax benefits, the value of housing services, and financial leverage the returns

to housing are increased by a multiple of between 1.6 and 3.7 and greatly exceed the returns from

alternative investments.

In short, when tax considerations, imputed rent, and financial leverage are included in the financial

returns to owner-occupied housing, on average and over fairly lengthy holding periods this

investment is found to generate significant financial returns. When financial leverage is not taken in

account, estimates suggest that the financial returns are a little lower than for stocks, but with

appropriately lower risk. When financial leverage is included, the financial returns to homeownership

were found to be even greater than for stocks—at least in the four markets studied over the period

from 1980 to 1996.

Several of the studies cited above also examine the issue of how much housing should be owned to

optimize a household’s investment portfolio. In general, the household’s consumption demand for

housing is found to force lower-wealth households to hold a larger share of their wealth than is

optimal in housing. But these studies still find that given the constraint that a household needs to

occupy a home and given the risks and returns offered by owner-occupied housing, home ownership

does represent a constrained optimum investment strategy (Flavin and Yamashita, 2002; Brueckner,

1997; Goetzmann, 1993).

Chapter 4: Financial Benefits of Homeownership

Variations in Housing Appreciation by Income and Race-Ethnicity

While in general it may be true that homeownership offers the potential for significant financial

returns, the key question for this study is whether low-income and minority owners are as likely as all

homeowners to experience positive returns. Given both the importance of house price appreciation to

the overall returns to homeownership and the fact that low-income and minority buyers are likely to

buy homes in different submarkets than higher-income, white households, it is important to address

the question of whether these buyers are as likely to realize positive house price appreciation.

There is a fairly extensive literature evaluating differences in housing appreciation rates across

different value submarkets within metropolitan areas. Dietz and Haurin (2002) and McCarthy, Van

Zandt, and Rohe (2001) provide fairly thorough reviews of this literature. There are a variety of

approaches used to define submarkets by value in these studies. One approach divides individual

homes into different value categories (e.g., lower quartile or upper quartile) without regard to the

specific neighborhood where the home is located. Another segment examines appreciation rates for

individual neighborhoods as a function of neighborhood characteristics, including the median home

value. Finally, a few studies have used household micro data that provides information on the

characteristics of the occupant and the home that can be used to define the market segment. A

common feature of this literature is that given the need for detailed information on house values and

characteristics, many of these studies examine either a single market area or a small number of

markets over a specific time period. In these cases it is not clear to what extent the findings are

generally true of low-valued market segments or are just unique to the individual markets over the

specific period studied.

Given these variations in the literature, it is hard to draw definitive conclusions about differences in

appreciation rates by housing value from any individual study. But taken as a whole, the literature

leads fairly convincingly to the conclusion that there is no consistent difference in appreciation rates

between low-income and high-income market segments. It is true that several studies have found that

lower-value homes or neighborhoods have experienced less appreciation (Poterba, 1991; Seward et

al., 1992; MacPherson and Sirmans, 2001; and Kim, 2000). But it is also the case that others have

found that lower-valued homes or neighborhoods have experienced more appreciation (Case and

Shiller, 1994; Case and Mayer, 1996; Archer et al., 1996; and Belsky and Duda, 2002b). Most

commonly, the results of these studies either find no significant difference in appreciation rates or the

results are mixed, finding that whether low- or high-valued homes or neighborhoods appreciate more

depends on the specific time period or the specific market studied (Kiel and Carson, 1990;

Pollakowski et al., 1992; Smith and Tesarek, 1991; Smith and Ho, 1996; Goetzmann and Spiegel,

1997; Li and Rosenblatt, 1997; Quercia et al., 2000; and Case and Marynchenko, 2002).

For example, Li and Rosenblatt (1997) examined price changes at the neighborhood level in three

different California metropolitan areas. Their analysis separates the period when house prices were

rising (1986 to 1990) from when they were falling (1990 to 1994). They find that while higher-value

homes appreciated more rapidly during the boom, they also fell more rapidly during the bust.

Importantly, they found that the direction, magnitude, and significance level of the explanatory

variables varied by time and by market area, indicating that there was no consistent relationship

between a neighborhood’s characteristics and house price trends. Case and Marynchenko (2002)

come to similar conclusions based on their examination of price trends for neighborhoods in Boston,

Chapter 4: Financial Benefits of Homeownership

Chicago, and Los Angeles over the period from 1983 through 1998. They find that low-valued

neighborhoods in Chicago experienced consistently higher price appreciation over this period, while

in both Boston and Los Angeles whether low- or high-valued homes experienced more or less growth

depended upon the specific time period considered.

In short, taken as a whole the literature indicates that there is no reason to believe that low-value

segments of the housing market will necessarily experience less appreciation than higher-valued

homes. In fact, at different points in time in different market areas, low-valued homes and

neighborhoods have experienced greater appreciation rates. Although the opposite is also true.

In comparison to studies of differences in appreciation rates across different submarkets defined in

terms of housing values, there is much less recent research examining differences in appreciation

rates by the race-ethnicity of the owner or the racial-ethnic composition of the neighborhood.

Given

the thinner literature on this topic it is harder to draw general conclusions. Kiel and Zabel (1996) link

data from the American Housing Survey for Chicago, Philadelphia, and Denver from the late 1970s

through 1990 to census data on neighborhood characteristics. Including controls for the housing unit,

the neighborhood, and the household, they find that the magnitude of the effect of neighborhood

racial composition varies significantly over time in the different market areas. At times, prices grew

faster in black areas, while at other times they grew more slowly. They conclude that the impact of

racial composition on house prices is not consistent either over time in the same market or across

market areas.

MacPherson and Sirmans (2001) examine changes in individual house prices in Orlando and Tampa

over the period from 1970 through 1997. They include measures of neighborhood racial and ethnic

composition in their model of house prices and find that in both markets a higher Hispanic share was

associated with greater price appreciation over the period, while the black share was not statistically

associated with price changes. They also include measures of the change in the Hispanic and black

population shares over time, and find that increases in the black share were associated with less

appreciation in both areas while increases in Hispanic share were associated with higher appreciation.

Two other studies come to more negative conclusions.

Quercia et al. (2000) examine price changes for individual homes in Miami over the period from 1972

to 1993. They find that homes in neighborhoods with a high concentration of minorities experienced

lower appreciation over the period than other neighborhoods. Kim (2000) modeled changes in house

prices at the neighborhood level in Milwaukee over the period from 1971 to 1993. The results

indicate that house price appreciation was much lower in neighborhoods that had high minority

shares. Finally, Coate and Vanderhoff (1993) use the American Housing Survey (AHS) for the

There is a literature from the 1970s and earlier that was concerned with the issue of whether racial

segregation of blacks resulted in blacks and whites paying different prices for comparable housing. These

studies generally focused on evaluating differences in the price of housing in a single market and at a single

point in time based on the racial composition of neighborhood and how the racial composition had been

changing. The general conclusion of this literature is that blacks paid a premium compared to whites for

housing in the 1960s, but as white suburbanization accelerated in the 1970s house prices in predominantly

black neighborhoods were lower than in white areas. See, for example, King and Mieszkowski (1973),

Schnare (1976), Schnare and Struyk (1977), and Yinger (1978).

Chapter 4: Financial Benefits of Homeownership

period 1974 to 1983 to analyze house price changes including the owner’s race as an explanatory

variable. They find that the owner’s race is not statistically significant in predicting house price

changes once other variables are controlled for.

Taken together, these few studies present a somewhat mixed picture. Only one study found that there

were periods when house prices rose more rapidly in neighborhoods with a high share black, although

two studies found no racial effect. On the other hand, the one study that included the Hispanic share

found that prices grew more rapidly in areas with a higher Hispanic presence. Based on these results,

it does appear that homes in mostly black areas may be less likely to experience appreciation, but this

conclusion is tempered by the small number of studies and the fact that these studies mostly analyzed

trends from the 1970s and 1980s, which may no longer be relevant. It is also important to bear in

mind that the literature reviewed in Chapter 2 suggests that minority first-time homebuyers are not

concentrated in predominantly minority neighborhoods, although they are moving to areas with

higher minority shares than the typical white homebuyer. Given the lack of research on variation in

appreciation rates by neighborhood racial composition, particularly for the period since the 1980s,

further research on the appreciation rates realized by minority homebuyers in the 1990s would make a

valuable contribution to the literature.

Variation in Appreciation by Structure Type

Another key factor to consider regarding differences in appreciation rates is the type of home

purchased. For the most part, studies examining both the financial returns to housing generally as

well as differences in appreciation rates for different segments of the housing market have focused on

price trends for single-family detached housing. Buts as shown in Chapter 2, over the 15 years from

1989 to 2003 nearly a quarter of low-income first-time homebuyers purchased a manufactured home,

compared to only 11 percent of moderate-income and 4 percent of high-income buyers. Black first-

time buyers are also somewhat more likely to purchase single-family attached homes (12 percent

versus 8 percent for all buyers). First-time buyers may also be more likely to purchase condominiums

in multiunit structures in high-cost markets as a more affordable way of attaining homeownership.

There is good reason to believe that structural differences exist in the supply and demand for different

types of homes, producing significant differences in appreciation rates. Unfortunately the literature in

this area is particularly thin so relatively little is known about differences in appreciation rates by

structure type.

Turning first to differences in appreciation rates among structure types other than manufactured

homes, Tong and Glascock (2000) claim that their study is the first designed to model drivers of

appreciation rates across unit types. They compare rates of price appreciation across three types of

housing units – single-family detached, townhouses, and condominiums – in three geographic areas in

Maryland—Baltimore City, Baltimore County, and Montgomery County. They note that two

previous studies provide incidental information on the issue. The analysis by Pollakowski et al.

(1992) included a dummy variable for single-family detached housing that was positive and

statistically significant, suggesting that these homes appreciate more rapidly than other structure

types. Similarly, Clapp, Giaccotto, and Tirtiroglu (1991) found that condos appreciated less rapidly

than single-family homes in the Hartford metropolitan area. In neither of these studies, however, did

the authors pay much attention to differences in appreciation rates by unit type.

Chapter 4: Financial Benefits of Homeownership

Looking at the issue directly, Tong and Glascock (2000) find substantial differences in the factors

explaining appreciation rates both across the three structure types and across the three geographic

areas studied. Not only do the explanatory variables differ, but they also found substantial differences

in appreciation rates by structure type. Over the entire 1973 to 1997 period, single-family detached

units in Montgomery County appreciated 22.4 percent in real terms, while townhouses fell 3.1

percent. In Baltimore County, townhouses appreciated most rapidly—24.8 percent versus 16.6

percent for single-family detached and a 10.8 percent decline for condos. In Baltimore City, single-

family detached home prices fell 10.7 percent, slower than the declines for both townhouses (-16.2

percent) and condos (-21.2 percent). The authors also show that condo prices are more volatile across

all three geographies. In short, the findings of Tong and Glascock indicate that there are important

differences in appreciation by structure type. Although the findings of this single study should not be

interpreted as an indication that structures other than single-family houses appreciate more slowly.

Data from the National Association of Realtors show that condominium prices surged much more

than even single-family home prices from 1997 through 2004. Hence, the appropriate conclusion is

that the correlation of price appreciation with structure type is sensitive to temporal and spatial

variations in the supply and demand for different structure types.

Given the relatively high share of manufactured housing among low-income first-time homebuyers

the question of whether these homes offer the same potential for appreciation is particularly

important. There are two different reasons why manufactured homes might not be expected to

produce the same returns. First, given the different building process and materials used to construct

these homes, there is a perception that they are likely to depreciate faster than site built homes

(Jewell, 2003). Second, about half of those who own manufactured homes do not also own the land

on which the homes are located. Since appreciation in house values largely reflects appreciation in

land values, those who do not also own the land on which their homes are located may be much less

likely to experience a growth in housing equity.

There are several studies that have compared rates of appreciation of manufactured homes with

conventionally built homes. The general conclusion of these studies is that in cases where owners

also own the land, manufactured homes generally appreciate at close to the same rate as other homes,

but that absent land ownership manufactured homes offer little opportunity for appreciation.

Stephensen and Shen (1997) analyze somewhat limited information on appreciation rates of

manufactured versus site built homes in four North Carolina counties with different housing market

characteristics. The study concludes that manufactured housing placed on permanent foundations and

titled as real property appreciated nearly identically to comparable site-built homes. However, they

also find that manufactured homes without land ownership depreciated in value over time.

Apgar et al. (2002) simulate changes in values of manufactured homes versus site built homes based

on changes in construction costs and real property values (which is used as a measure of land price

changes). Their simulation leads them to conclude that, because of improvements in manufactured

housing production processes over time that have lowered the costs of new units, existing

manufactured units themselves are likely to have appreciated a bit more slowly than site built homes.

However, they also find that since changes in land value play such an important role in overall house

price changes, the impact of slightly lower appreciation of the housing unit has only a slight impact

on overall appreciation rates. Specifically, they estimate that the real value of site built homes grew

Chapter 4: Financial Benefits of Homeownership

8.2 percent between 1990 and 2000, while manufactured homes where the land is also owned were

estimated to have increased in value by 7.5 percent. But, again, if owners of manufactured homes did

not own the land, they would not have realized any gains at all.

To assess differences in appreciation rates for manufactured homes, Jewell (2003) analyzes county

appraisal data for three Texas counties for the period from 1990 to 2002 as well as data from the AHS

for the period 1985 to 1999. His analysis of the appraisal data for the Texas counties finds no

significant difference in appreciation rates between site built and manufactured homes if land is

owned, although manufactured homes did exhibit greater variation in appreciation rates. However,

manufactured homes on leased land had much lower appreciation rates. His analysis of the AHS

came to very similar conclusions: while manufactured homes on owned land appreciated at rates

similar to site built homes, there was much greater variability in appreciation rates and manufactured

homes on leased land actually depreciated on average. A study by Boehm and Schlottmann (2004a)

using AHS data for 1991 to 2001 corroborates Jewell’s results.

There is a fairly clear consensus that emerges from these studies. Manufactured housing sited on

owned land on average appreciates at rates similar to site-built homes. However, there is greater

variation in appreciation rates so investments in these homes are somewhat more risky. Most

importantly, studies consistently find that manufactured homes that do not include land ownership do

not offer the potential for asset appreciation. As Boehm and Schlottmann concluded “manufactured

housing where the household does not own the lot is not an investment in any sense.” Thus, for the

half of manufactured home owners who do not own their lot there is little chance of realizing a

financial return.

Influence of the Timing of Purchase and Sale on Financial Returns

The findings from the literature reviewed above suggest that in general the rates of return make

homeownership a good investment and, importantly, there is no reason to believe that low-valued

homes are less likely to produce a positive return than higher-valued homes. However, these findings

do not mean that housing is a risk-free investment. As the literature examining housing appreciation

patterns makes clear, housing appreciation rates vary substantially both across markets and over time.

For example, an analysis of price trends in 163 market areas for which Freddie Mac produces price

indexes reveals that between 1990 and 1995 slightly more than a quarter of these areas experienced

declines in nominal home values while more than half had gains that did not keep pace with inflation.

But since 1995 strong house price growth has been widespread. Between 2000 and 2005 41 percent

of market areas had nominal price growth in excess of 50 percent and in virtually all markets house

price growth has outpaced inflation. The Freddie Mac price indexes also reveal that over the long-run

the periods of increases more than offset the periods of declining prices as house price growth

exceeded inflation for the period from 1975 to 2005 in all but 13 of the 163 markets. But since the

typical homeowner is unlikely to own a single home over the “long run,” whether a homebuyer

realizes a positive financial return depends critically on both what market they live in and when they

buy and sell their homes.

The analysis by Belsky and Duda (2002b) makes an important contribution to the literature by

evaluating the importance of the timing of the purchase and sale of a home in determining the

financial returns realized by individual owners. They analyze repeat sales data for single-family

Chapter 4: Financial Benefits of Homeownership

homes in four markets – Boston, Chicago, Denver, and Philadelphia – over the period from 1982 to

1999. Houses are divided into three cost categories at the time of sale. Low-cost homes are those

affordable to households at or below 80 percent of the area median income, moderate-cost homes are

those affordable to households between 80 and 120 percent of area median income, and high-cost

homes are those affordable to households above 120 percent of area median income. Belsky and

Duda estimate the mean return realized by homeowners by first deflating prices for inflation and then

subtracting six percent of the sales price for transaction costs. They find that the average return on

sale among low-cost homes was consistently large and positive, ranging from 23 percent in

Philadelphia to 54 percent in Boston. In contrast, the average return for moderate- and high-cost

homes was small or negative in most cases (with the one exception being moderate-cost homes in

Boston which increased in value by 19 percent on average).

But the key finding of Belsky and Duda was that a large share of owners at all price levels

experienced a financial loss after taking into account both general inflation and the transaction costs

of selling their homes. In three of the four markets a small majority of all owners experienced a loss,

while in the fourth 41 percent had a loss. Importantly, in all four areas, buyers of low-cost homes

were less likely to lose money, so homeownership appears to have been less of a gamble for those

who purchased low-cost homes in these four markets. But even among these buyers, between 21 and

42 percent were estimated to have had a real loss on their investment.

Belsky and Duda find that the better financial return experienced by buyers of low-cost homes

reflected better timing of purchases, as high-cost purchases were more likely to occur around peaks

and low-cost purchases made up a greater share of purchases during troughs. This result leads them

to question whether efforts to increase homeownership among low-income families during the

housing boom of the late 1990s might mean that these buyers will be more likely to experience losses

following the current boom.

However, there are several important caveats to the results of this study. First, as the authors make

clear, they are not reporting the shares of all buyers who experience losses, but rather shares of all

those who both bought and sold their homes within the period for which they had data. Since longer-

term owners are more likely to ride out the ups and downs of market prices, non-sellers – who are not

in the data set – may be more likely to have experienced gains. A large majority of the cases in their

data set consisted of owners who sold their homes within nine years of purchase. Based on data from

the AHS, the authors estimated that 35 percent of low-income homebuyers in the mid 1980s moved

within nine years as did 44 percent of high-income buyers.

Thus, when considering that about half

of all homebuyers who sold within nine years of buying their homes experienced a real financial loss,

it is important to bear in mind that these sellers actually account for less than half of all buyers.

Another caveat with their analysis is that while it is important to take inflation into account in

determining the financial return on homeownership, an owner may incur a loss in real terms but still

walk away from the sale with some cash in hand. This outcome is of no small importance as the cash

is likely to be important in providing the seller with funds to use for a downpayment on another

The shares of low-income buyers sustaining long spells of homeownership cited by Belsky and Duda are

higher than the shares reported in Chapter 3 because their analysis is not limited to low-income first-time

buyers, but instead includes all buyers of low-valued homes.

Chapter 4: Financial Benefits of Homeownership

home. Along those lines it is not unimportant that Belsky and Duda find that the vast majority of

owners sold their homes for more than they bought them and so may have had the financial means to

purchase another home. Along these lines, it is important to note that the study examines only single

spells of homeownership. It is possible that those who sold their homes near the bottom of a cycle

and experienced a real loss still managed to buy another home. In that case, they may have been well

positioned to benefit from the next upswing in prices.

Nonetheless, the findings of Belsky and Duda highlight the critical importance of sustaining home

ownership in order to ride out market downturns. The literature reviewed in Chapter 3 indicates that,

in fact, sizeable shares of low-income first-time homebuyers are not able to sustain homeownership

for at least five years. Belsky and Duda’s findings suggest that these sellers are fairly likely to

experience a loss in real terms. On the positive side, they do find that low-cost homes experienced

greater price appreciation and exposed buyers to less risk of loss compared to higher-cost homes—at

least in the markets and time period they studied. As Belsky and Duda (2002, 217-219) succinctly

observe: “Purchasing a home, especially on a single term of homeownership, is risky. The American

Dream of homeownership may turn out to be just that for millions of owners, but for large shares it is

not a fruitful investment unless sellers reenter the market and are able to ride one or more waves of

appreciation over their lifetime.”

Comparing the Costs of Owning and Renting

The previous section was concerned with the question of whether homeowners–and particularly low-

income and minority homeowners–are likely to realize a fair financial return on their investment

through appreciation in home values. But while homeownership may be likely to provide a fair

financial return, it is still possible that individuals could be financially better off by renting a home

and devoting their savings to other investments. In this section we review the literature that compares

the costs of owning and renting, with a particular emphasis on differences in these costs for low-

income or minority households.

The ongoing costs associated with owner occupied homes include those for mortgage interest,

property taxes, maintenance, and hazard insurance. In addition, there are transaction costs associated

with buying and selling a home and with originating a mortgage both at the time of purchase and

when the mortgage is refinanced. Offsetting these costs are the financial benefits associated with

appreciation in the value of the home and the potential deductibility of payments for mortgage interest

and property taxes. Thus, a complete accounting of the costs of homeownership must take into

account the ongoing costs of paying for the home, the annual tax benefits realized (if any), and the

one time costs and benefits of transaction costs and capital gains on sale of the home.

In comparison, the costs and benefits of renting are fairly straightforward. The costs include rent

payments and the transaction costs of signing a new lease (such as a realtor fee) and leaving an

Since there capital gains taxes are not paid on either the first $250,000 in gains for single-person owners or

$500,000 in gains for married couple owners, it is generally assumed that there are no capital gains paid on

housing appreciation by homeowners. While this treatment of capital gains is relatively new, the tax

treatment of capital gains on owner-occupied homes in the past also generally made it possible for most

owners to avoid these taxes.

Chapter 4: Financial Benefits of Homeownership

existing one (such as a loss of a security deposit for damage to the unit or fees for breaking a lease).

These costs are offset by the financial gain associated with investing funds that would otherwise be

used to support the purchase of a home. Alternatively, rather than count this return as a benefit of

renting, it can be counted as an opportunity cost associated with owning.

While the above description may make it seem as though a comparison of the costs of owning and

renting is fairly straightforward, in practice such a comparison is extremely complex given the large

number of factors that influence the costs of owning. Before turning to the literature that compares

the costs of owning and renting, it is helpful to begin by examining each of the components of these

costs and assessing how they might differ based on household income or race.

Variations in Owner and Renter Costs by Income and Race-Ethnicity

The most obvious way in which the cost of owning differs by income is as a result of the income tax

treatment of mortgage interest and property tax payments. Under federal tax rules, these costs are

both deductible from income. But an individual tax filer will only benefit from this provision of the

tax code if the value of their itemized deductions exceeds the standard deduction.

The value of this

benefit is equal to the amount of itemized deductions in excess of the standard deduction times the

marginal tax rate paid on the income that is sheltered by these excess deductions. Low-income

homebuyers are less likely to benefit from these tax provisions both because their lower housing costs

are less likely to exceed the standard deduction and, if they do itemize, because their lower marginal

tax rates reduce the value of this benefit.

Low-income married couples are particularly unlikely to realize these benefits, as their standard

deduction of $9,700 will often exceed interest and property tax payments on a modest home. Single

persons and, to a lesser extent, single heads of household are more likely to benefit as their standard

deductions are only $4,850 and $7,150, respectively. Even if married couples do itemize their

deductions, they are also penalized by the fact that they face lower marginal tax rates than single

persons or single heads of households. Follain and Ling (1991) and Capone (1996) both find that

low-income married couple households are unlikely to realize any tax benefits from homeownership.

To illustrate how the potential magnitude of income tax benefits relative to housing costs may vary by

income, Appendix A presents the results of a simulation comparing the value of tax benefits as a

percent of housing costs for different household types and income levels. Specifically, three

household types are considered: a married couple with two children, a single parent with one child,

and a single person. We examine each of these households at five different income levels: $15,000,

$30,000, $45,000, $60,000, and $75,000. At each income level, we estimate the maximum value of a

home that is affordable using fairly standard underwriting criteria. To put these incomes in

perspective, the typical low-income first-time homebuyer in the 2003 AHS had household income of

about $30,000, with incomes of $15,000 and $45,000 roughly representing lower and higher bounds

Other than mortgage interest and property tax payments the next most common itemized deduction are for

state and local taxes and charitable contributions. Other categories of itemized deductions include medical,

employment, and educational costs, but these are much less commonly claimed given the rules governing

what can be claimed as a deduction in each of these categories.

Chapter 4: Financial Benefits of Homeownership

on this group. Incomes of $60,000 and $75,000 may be taken as more typical of moderate- and

higher-income first-time buyers.

Exhibit 14 presents the results of this exercise. The value of tax benefits is expressed as a percent of

total housing costs paid. As shown, at the lowest income level considered ($15,000) none of the three

types of households realize any tax benefits from homeownership. This result reflects the fact that the

value of homes that are affordable at this income are too low to produce itemized deductions that

exceed the standard deduction. But as household income rises to $30,000, both single persons and

single-parent households begin to realize some tax benefits from homeownership as itemized

deductions exceed the standard deduction. Even at this income level married couples still do not

realize any benefits given the higher level of their standard deduction. The magnitude of benefits

realized by single persons and single parents is also fairly small at seven and four percent,

respectively. Given that the average income in 2003 for first-time low-income buyers was $30,000,

this simulation suggests that the tax benefits realized by these typical first-time buyers does little to

reduce their out-of-pocket housing costs.

Exhibit 14

Value of Tax Benefits from Homeownership as Percent of Total Housing Costs by

Gross Household Income and Household Type

25%

20%

18%

17%

16%

15%

10%

0% 0% 0%

$15,000 $30,000 $45,000 $60,000 $75,000

Gross Household Income

Single Person Single Parent, One Child Married Couple, Two Children

Note: See text for assumptions used in estimating the value of tax benefits.

It is possible that in many markets the assumptions used here underestimate the actual housing costs

incurred by buyers due to higher house values, higher property taxes, or higher interest rates. In these cases

the tax benefits would be higher—but the housing costs burdens would also be much higher.

Chapter 4: Financial Benefits of Homeownership

As incomes rise above $30,000 homebuyers become more likely to realize tax benefits from owning,

although benefits remain lower for single parents and, particularly, married couples. Tax benefits

only exceed 15 percent of housing costs for singles at $45,000 in income, for single parents at

$60,000 in income, and for married couples at $75,000 in income. This example illustrates how the

value of tax benefits can greatly reduce the costs of homeowning, but that low-income households,

and particularly married couples, are unlikely to realize these benefits. Thus, low-income households

will be more likely to have the costs of owning exceed the costs of renting.

Another aspect of owner’s costs that are likely to vary by income and race is the mortgage interest

rate. There has been tremendous growth in higher-cost subprime lending over the last decade, with

these loans particularly common among low-income and minority owners. While the analysis of the

AHS presented in Chapter 2 did not find any significant difference in the interest rate on home

purchase mortgages by income or race, it may be that the increased use of subprime loans for home

purchase was not yet evident in the period from 2001 to 2003 covered by most recent survey.

However, the information presented in Chapter 3 does indicate that low-income and minority owners

are less likely to benefit from refinancing their mortgages to lower interest rates when such

opportunities arise. Minorities, in particular, are both less likely to refinance when interest rates drop

and, when they do refinance, they are found to pay interest rates that are about one percentage point

higher than whites. In short, differences in mortgage choices, both at purchase and over time, are

likely to contribute to differences in the cost of owning relative to renting for low-income and

minority owners.

As will be discussed more below, a particularly challenging and important aspect of the comparison

of the costs of owning and renting is determining the market rent for a home of a given value. One

approach used is to make an assumption about the ratio of annual rent to the home’s value. Most

studies do not address the question of whether there might be variations in the ratio between rent and

house values. However, findings from Linneman and Voith (1991) and Capone (1996) suggest that

there might be a higher ratio of rents to values among low-priced homes. Linneman and Voith base

this conclusion on an analysis of survey data for Philadelphia while Capone’s conclusions is derived

from the results of interviews he conducted with real estate professionals around the country. One

argument for this result is that the lower income tax benefits realized by low-income households

reduces the amount they are willing to bid for these homes. To the extent that rents are a larger share

of house values among low-priced homes, then low-income owners will find it more likely that

owning is preferred to renting.

A final way in which the costs of owning may vary by income and race is due to differences in the

length of time that owners occupy a home. Given the substantial transaction costs of buying and

selling homes, renting will generally be a better financial choice if a household is expected to move

within a few years of buying.

The studies reviewed in Chapter 3 found that low-income and

minority owners were more likely to move within five years. This shorter tenure in the home will

tend to make renting more attractive compared to owning.

In fact, as will be discussed below, the focus of several of most studies that compare the costs of owning

and renting is in estimating the length of time that a household must stay in the home in order for owning to

be preferred to renting.

Chapter 4: Financial Benefits of Homeownership

Other aspects of owner costs are less likely to vary by income or race. Neither transaction costs nor

the rate of return available on alternative investments should vary by income. As discussed above,

there is also no reason to believe that there are systematic differences in the appreciation rates

experienced by low-income owners. While available studies are less conclusive regarding differences

in appreciation rates experienced by minorities, there is also not strong evidence that appreciation

rates are lower for these owners. Nonetheless, house price trends do vary significantly over time and

across markets. In short, where and when you buy a home will play a significant role in determining

whether owning would be preferred to renting, but this variability will be evident among both low-

and high-income buyers.

Studies Comparing the Costs of Owning and Renting

There are surprisingly few studies that actually compare the cost of owning and renting. The few that

exist differ in important ways in terms of how detailed their estimates of costs are, the methods used

to determine market rents, whether they assess the importance of variations in price trends and other

market factors over time, how tax issues are handled, and other factors. Given the variation in

methods and assumptions used in these studies it is not surprising that they reach different

conclusions about whether owning is financially preferred to renting. While in many situations

owning is found to be financially preferred to renting, it is also true that under a variety of plausible

assumptions about owners’ circumstances and market conditions, renting can be found to be cheaper.

Most studies also conclude that the value of federal income tax benefits can be decisive in making

owning financially preferable to renting, with the implication that low-income households who do not

realize these tax benefits are often better off renting. Nonetheless, the conclusion about whether

owning is preferred to renting largely rests on assumptions about the level of rents compared to house

values. Since there is little reliable information on this relationship or how it might vary across price

segments of the market, it is difficult to draw any definitive conclusions about whether owning is

financially preferable to renting. Nonetheless, as will be discussed in the final section of this chapter,

studies do consistently find that homeowning households are more likely to accumulate wealth over

time than renters.

Mills’ (1990) compares the costs of owning and renting by first developing equations to estimate

these costs and then making assumptions about the values of key parameters. He then uses these

equations to test the importance of various components of homeowners’ costs in determining under

what circumstances owning is cheaper than renting. The key variables he focuses on are the income

tax treatment of interest payments and property taxes, the rate of house price appreciation, the

mortgage interest rate, and rent levels.

In all of the scenarios he considers, homeowner costs start

Assumptions about the rate of house price appreciation account for the benefits of financial leverage

offered by homeownership, where owners realize increases in the value of the entire house and not just on

their equity stake. Related to this, another important assumption in comparisons of the cost of owning and

renting is the opportunity cost of funds invested in the home; that is, the financial return that would be

realized if the downpayment and accumulated home equity were invested in the next best investment

alternative. This foregone return is part of the cost of owning (or, alternatively, the benefit of renting).

None of the studies treat the rate of return earned on other investments as a variable to identify how

important this variable is in determining whether owning is better than renting. In general, the more

Chapter 4: Financial Benefits of Homeownership

out being higher than renting due to the high transaction costs of buying and selling a home. But as

the length of time that the home is occupied increases, the importance of these transaction costs is

reduced. Given this pattern, the focus of his analysis is in determining the number of years the home

must be owned for owning to be preferred to renting.

Under his initial assumptions, he finds that a household would need to stay in the same home for 7.5

years in order for owning to be cheaper than renting. He explores a number of variations in

assumptions regarding average housing appreciation rates (increasing these rates from a base

assumption of 5 percent in nominal terms to 6 or 7 percent) and transaction costs (lowering the costs

of buying a home from 5 percent of its value to 4 percent and the costs of selling a home from 10

percent of its value to 9 percent). By making these more optimistic assumptions he finds that the

break even holding period is reduced to a range of 3.3 to 6.7 years. The basic conclusion that

emerges from these simulations is that under most scenarios a household needs to occupy a home for

between five and seven years for owning to be a better financial choice.

However, Mills’ analysis finds that the two most important factors in determining whether owning is

preferred to renting are the ratio of rent to property value and the household’s marginal tax bracket.

In his initial analysis, Mills sets annual rent levels at 7 percent of house values. He bases this

assumption on a comparison over time in the value of rent in national income and product accounts to

the value of dwellings from national capital accounts. Using these data, he finds that rent levels

fluctuated on an annual basis between 4.1 and 9.1 percent of house values over the previous half

century. His assumption of 7 percent is slightly above the mid-point of this range reflecting recent

upward trends in this measure. His results, however, indicate that whether renting is preferred to

owning is highly dependent on this assumption. If rent levels are assumed to be 8 percent of house

values rather than 7, the break even holding period falls from 7.5 years to 4.4 years. On the other

hand, if rents are assumed to average 6 percent of house values, the break-even holding period is

increased more than three fold to 23.5 years. There is no other variable that is found to be as

important in producing changes in the breakeven holding period.

Another key conclusion from Mills’ analysis is that the degree to which households are able to benefit

from the deductibility of mortgage interest and property taxes is of fundamental importance in

determining whether owning is preferred to renting. In his base case scenario, he assumes that the

household is in a 28 percent marginal tax bracket. When he lowers the marginal tax rate to 15 percent

he finds that owning is never preferred to renting. Mills probably even underestimates the importance

of tax benefits since he ignores the standard deduction and simply assumes that all owners can claim

the full value of mortgage interest and property tax payments as deductions. In fact, as demonstrated

above, many low-income owners receive no tax benefits from owning because their housing costs do

not exceed the standard deduction.

An analysis by Goetzmann and Spiegel (2002) of the importance of differences in the tax treatment of

residential property between owners and investors supports the conclusion reached by Mills regarding

the limited financial appeal of homeownership for those in low tax brackets. Goetzmann and Spiegel

start by assuming that in equilibrium, the rent earned by property investors must just equal their costs,

conservative the assumption about the return on the next best alternative investment, the more attractive

homeownership will be.

Chapter 4: Financial Benefits of Homeownership

including a fair return on their capital investment. They then compare the costs incurred by owner

occupants to those of investors and show how the question of whether owning is preferred to renting

hinges on a comparison of the tax brackets of owners and investors. Based on assumptions about the

marginal tax rates likely to be incurred by investors in rental properties, they find that owner

occupants must have a marginal income tax rate of 33 percent in order to be financially better off

owning. Since this rate is higher than that faced by low-income households, they conclude that

renting is financially preferred to owning for these households. While the analysis by Goetzmann and

Spiegel is highly stylized, ignoring, for example, the transaction costs of buying and selling homes,

their analysis does make it clear that income tax benefits are an important factor in determining

whether owning is cheaper than renting.

Capone (1996) extends the work by Mills in two way: first, by exploring the validity of his

assumptions about the relationship between rents and house values; and second, by examining in

more detail the importance of the tax treatment of housing costs in the comparison of renting and

owning. To test the validity of Mills assumptions about rents, Capone develops a detailed equation of

the cost to investors of residential property and then solves this equation for the rent level (as a

percent of the property value) that must be charged for investors to break even. Capone finds that a

key assumption in the calculation of investors costs is their expected holding period, with longer

holding periods resulting in lower rent to value ratios. His simulations find that rent to value ratios

range from a low of 4.4 percent given an investor holding period of 20 years to a high of 10.9 percent

for holding periods of just five years. In fact, this range is fairly close to the range estimated by Mills

using a different methodology.

However, Capone cites two sources in support of a higher assumed ratio of rents to values for low-

income households. First, he cites a study by Linneman and Voith (1991) that used the AHS for the

Philadelphia metropolitan area in 1982 to estimate rent-to-value ratios as a function of both housing

and occupant characteristics. While this analysis finds a market average rent to value ratio of 10

percent, it also finds that these ratios are consistently higher for lower income households—generally

exceeding 12 percent for those with incomes of $15,000 or less. Linneman and Voith argue that the

higher capitalization rates found for lower-income households are likely a result of the lower tax

benefits from homeownership for these households. In essence, they argue that there is segmentation

of the housing market by income, with the tax benefits realized by the marginal buyer in different

segments of the market being capitalized into home prices. Goodman and Kawai (1982) provide

further support for this view that differences in tax benefits by income are reflected in housing prices.

Capone also justifies his assumption of higher rent to value ratios on the basis of interviews he

conducted with real estate professionals to gather their assessment of the ratio of annual rents to

property values for single-family homes. These interviews lead him to conclude that rent to value

ratios for low-valued homes are generally in the range of 10 to 15 percent of value. He argues that his

simulations of rent-to-value ratios required of investors are consistent with the findings of both

Linneman and Voith and his interviews if more conservative assumptions are made about investors’

expectations regarding vacancy periods between tenants.

Capone then compares the costs of owning and renting assuming that rents range from 10 to 15

percent of home values. He finds that owning is preferred to renting as long as owners stay only

between two and three years. Given Mills findings about the importance of the rent-to-value

Chapter 4: Financial Benefits of Homeownership

assumption in determining whether owning is preferred to renting, it is perhaps not surprising that

Capone would reach a much more optimistic conclusion about the advantages of owning given his

much higher assumptions about rent levels. But countering his more generous assumptions about rent

levels, Capone also makes much more restrictive assumptions about the tax benefits of

homeownership. One of the purposes of his study is to assess the extent to which federal income tax

provisions help support homeownership among moderate-income households. To do this, he assumes

a much lower cost home than Mills ($56,000 rather than $200,000) and a lower marginal tax rate for

the homebuyer (18 percent rather than 28 percent). His estimate of owners’ costs also takes into

account the level of the standard deduction and only includes the value of tax deductions that exceed

the standard deduction. As a result of using a much lower priced home and factoring in the level of

the standard deduction, in Capone’s base case scenario owners receive essentially no income tax

benefits. So his conclusion that owning is preferred to renting for stays of as short as two to three

years is notable for not relying on income tax benefits.

Capone uses his equations to examine the relative importance of changes in income tax treatment and

assumptions about rent levels on the estimated breakeven residence in the home. He finds that the

higher that rents are assumed to be, the less important tax benefits are in determining the minimum

period that the home needs to be occupied before owning is preferred to renting. For example, if

rents are assumed to be 12 percent of value, then the breakeven period for owning is two years

regardless of whether any tax benefits are realized. However, if rents are set at eight percent of house

values then the breakeven holding period is as short as three years for those who benefit substantially

from tax benefits to as long as five years for those who do not benefit at all.

In short, the key extension to Mills conclusions that comes from Capone’s work is to show that if

rents are assumed to be as high as eight percent of house values, then low-income households may be

better off owning even if they do not obtain any tax benefits.

By assuming a constant relationship between house prices and rents, one issue that Mills and Capone

do not shed light on is how fluctuations in the relative levels of house prices and rents over time affect

the estimation of whether owning is financially preferred to renting. As Goodman (1998) notes, the

assumption that rents and house values remain in equilibrium over time ignores the fact that housing

prices may be slow to adjust to changes in market conditions. For example, Blackley and Follain

(1996) find that investors’ costs are much more volatile than market rents. Also, Gallin (2004) has

found that while rents and house prices do tend to move together in the longer run, in the short run

they may exhibit divergent trends.

To address this concern, Goodman uses actual national trends in home prices and rents between 1985

and 1995 to estimate the costs of owning and renting over this period. Using the American Housing

Survey from 1985, he estimates the rent on a prototypical single family home by a hedonic regression

model, with the value of the home based on a tabulation of owner reported home values. His

estimated annual rent is 6.9 percent of the house value, remarkably close to the assumption of 7

percent used by Mills. He then trends these values over time using the Freddie Mac index of home

prices and the Consumer Price Index for rental housing. However, the data presented by Goodman

shows that over this period rents and house prices grew by nearly an identical average annual growth

rate. As a result, his rent to value ratio hardly deviates from the starting assumption of 6.9 percent so

Chapter 4: Financial Benefits of Homeownership

his results do not offer a valid test of how sensitive this analysis is to an assumption of a stable

relationship between rents and house values.

Given the similarity of his assumption about the relationship between rents and values to that used by

Mills it is not surprising that his conclusion differs little from Mills. He finds that owning is preferred

to renting only as long as the home is occupied for about nine years. This is somewhat longer than

Mills estimate because Goodman makes somewhat more conservative assumptions about the value of

tax benefits by factoring in the value of the standard deduction. However, one interesting aspect of

Goodman’s results is that he shows an annual estimate of the ratio of owner and renter costs. After

the third year, owners’ costs are never more than three percent higher than renters’ costs. This result

is also consistent with the findings of Mills and Capone that with somewhat more favorable

assumptions about owners’ costs the breakeven holding period for owning can be as short as three

years.

Belsky, Retsinas, and Duda (2005) provide a more thorough test of the importance of the actual

volatility of house prices and rents in determining whether owning is preferred to renting. Like the

other studies reviewed, they construct an equation to estimate total housing costs for owners that can

be compared to the cost of renting. Like Goodman, rather than assume a constant relationship

between house prices and rents, they use price indexes to incorporate actual trends in these relative

prices over time. However, they improve upon Goodman’s analysis by analyzing price trends in four

different markets over an 18-year period. The four markets were chosen for study as they represent

different degrees of house price growth and volatility. The focus of their analysis is to estimate how

often owner costs are less than renter costs assuming holding periods of 3, 5, or 7 years. Given the

length of their data series on rents and house values, they can identify 16 different 3-year holding

periods, 14 5-year holding periods, and 12 7-year holding periods. They then report the share of these

different holding periods where owner costs were lower than renters’ costs.

Reflecting their concern with the question of whether homeownership is financially appealing for

low-income households, their starting home value is half of the market median value. They use data

from the Bureau of Labor Statistics on owners’ estimates of the rental value of their properties and

compare this to home values based on tabulations of the American Housing Survey for these markets.

The ratio of rents to values is then used to estimate a rent level for homes at half the median home

value. They then apply the market-specific Freddie Mac house price indexes and the consumer price

indexes for rental housing to these initial values. While not reported in their paper, the authors

indicate that the starting ratio of rents to values is on the order of five to seven percent across the

markets studied.

Belsky, Retsinas, and Duda note that the two factors that are most likely to affect the cost of owning

for low-income households is whether they are able to realize any tax benefits from owning and the

costs of mortgage finance. They then present a series of scenarios to test the impact of these factors

on the relative costs of owing and renting. To test the importance of tax benefits on the costs of

owning, as a starting assumption they assume that households benefit from mortgage interest and

property tax payments only to the extent the value of these deductions exceed the standard deduction.

They then assume that owners do not realize any tax benefits and see how this affects the likelihood

that the cost of owning will exceed the cost of renting. To text the impact of mortgage choices, they

Chapter 4: Financial Benefits of Homeownership

then increase the level of initial interest rates to simulate subprime mortgage rates and also examine

what happens if opportunities for refinancing to a lower cost mortgage are missed.

Pooling their results from the four markets, in their base case scenario where full tax benefits are

realized they find that in 53 percent of the possible 3-year holding periods owning would be preferred

to renting. If the holding period is extended to 5 or 7 years, the share of cases where owning is

preferred to renting rises to 63 percent of all possible holding periods. Thus, in only a little more than

half of the possible holding periods in the four markets is owning preferred to renting. Perhaps not

surprisingly given that their analysis focuses on low-valued homes, their results indicate that there is

little impact of tax benefits on the likelihood that the costs of owning are lower than the costs of

renting. They do, however, find that having an interest rate that is 2 percentage points higher, which

simulates the impact of moderately-higher interest rates from a subprime loan, reduces the likelihood

that the costs of owning are less than renting by between 8 and 16 percentage points. Further

increasing the interest rate obtained to be 5 percentage points above prevailing rates lowers the

likelihood that owning is preferred to renting to only between 15 and 22 percent. In short, their

analysis indicates that very high cost subprime loans make it unlikely that owning would be cheaper

than renting.

In assessing the conclusion reached by Belsky, Retsinas, and Duda that in a little less than half of all

cases examined renting would have been preferred to owning it is important to bear in mind that their

analysis is based on a fairly low ratio of rents to values, which tends to favor renting. They do not

attempt to assess whether these ratios vary by price segment of the market or what impact this

variation would have on their conclusions. Nonetheless, their analysis does show that volatility in

both house prices and rents is an important factor in determining whether owning is cheaper than

renting. Under their baseline assumptions, in only a slight majority of cases do they find that owning

will be cheaper than renting given a stay of between three and seven years. However, this conclusion

is as likely to hold for high-income households as it is for low-income households as they find that

whether a household realizes any tax benefits or not has little impact on their results. But they do find

that the higher mortgage rates associated with subprime lending, which are more likely to be incurred

by low-income and minority owners, can have a significant impact of whether owning is cheaper. In

particularly, if mortgage interest rates are five percentage points above prime rates, in less than 20

percent of cases is owning cheaper.

Conclusions Regarding the Costs of Renting and Owning

Looking solely at the literature that has compared the cost of owning and renting it is difficult to

reach any definitive conclusions about whether low-income or minority households are better off

owning or renting. Whether owning is financially preferable to renting is very sensitive to

assumptions about the level of rents compared to house values. Unfortunately, there is very little

empirical information on how rents compare to values for comparable housing units. While one

The authors also examine the impact of missing refinance opportunities, but find that there is very little

affect on the likelihood that owning is financially preferred compared to renting. However, this result

likely reflects the fact that they assume that there are only two years that offered refinance opportunities

(1993 and 1998), so that only a small subset of all possible holding periods were affected by these missed

opportunities.

Chapter 4: Financial Benefits of Homeownership

study found that the ratio of rents to values is higher in the low-cost segments of the market, since the

results are based on a single market at a single point in time there is not enough evidence to support

this claim. Given the importance of this issue to the question of whether owning is cheaper than

renting, more research on differences in rent levels relative to house values across different segments

of the market and at different points in time would be very informative.

The literature also suggests that under certain assumptions, the value of tax benefits can be quite

important in determining whether owning is a better deal than renting. The fact that low-income

households, particularly married couples, receive fewer tax benefits from owning means that these

households are more likely to be better off renting than households headed by unmarried individuals.

But Capone’s analysis also indicates that as long as rents are high relative to values, owning will

clearly be preferred to renting for stays as short as three years whether tax benefits are realized or not.

While the results of these studies suggest that there are a number of situations where low-income

households would have lower housing costs if they rented, it is clear that a large share of low-income

households still chose to own. In part, this may reflect the fact that there are many reasons for

wanting to own a home beyond purely financial considerations, including greater security of tenure,

the ability to tailor the home to fit one’s tastes, and greater privacy. It may also be that even if

homeownership is more costly it may still be more likely that owner households will accumulate

greater wealth as a result of the forced savings associated with paying off a mortgage over time. The

next section provides some support for this view by reviewing literature that has found that

homeownership is strongly associated with wealth accumulation even for low-income households.

Homeownership and Long-Run Wealth Accumulation

As the issues raised in Chapter 3 and in the sections above make plain, there is a complex web of

factors that play a role in determining whether owners realize the financial benefits of

homeownership. These include the degree to which house prices increase, whether owners are able to

sustain homeownership, the timing of buying and selling a home relative to housing price cycles, the

degree to which owners can take advantage of the income tax benefits of owning, and the choices

owners make along the way regarding financing, maintaining, and improving their homes. The

literature that has been reviewed in each of these areas provides both good and bad news regarding

the likelihood that low-income and minority owners are likely to benefit financially from

homeownership. On the plus side, analyses of the overall financial return from homeownership

suggest that ownership is a prudent financial investment. There is also no reason to believe that low-

valued homes are less likely to appreciate than higher-valued homes. In fact, research suggests that in

the past low-income buyers were more likely to benefit from buying near the bottom of housing

cycles. On the negative side, most low-income owners realize few, if any, of the tax benefits of

homeownership. Also, as discussed in Chapter 3, low-income first-time buyers, particularly minority

buyers, are more likely to have early exits from homeownership. Additionally, the surge in home

buying by low-income and minority households at what now appears close to the top of a price cycle

may eliminate the advantage they had in previous periods of buying at the bottom of the cycle.

Finally, minority owners are also more likely to miss out on valuable refinancing opportunities and

may be less likely to maintain their homes.

Chapter 4: Financial Benefits of Homeownership

In this section we review literature that takes a more holistic approach to evaluating the financial

returns to homeownership in an attempt to account for this complex web of factors. Rather than

attempt to isolate the impact of individual factors, these studies use longitudinal surveys of

households to observe their tenure choices over time and to assess how these tenure choices relate to

overall changes in household wealth. By following households over a long period and observing

changes in wealth, these studies implicitly account for the combined effect of housing price changes,

the timing of purchase and sale, the ability to take advantage of tax benefits, and the impact of

choices regarding refinance, maintenance, and improvement. An important caveat with these studies

is that they do not incorporate any controls for selection bias in who becomes an owner, and so it is

not clear whether the observed differences in wealth accumulation reflect solely the impact of

homeownership or whether the result is due to other unobserved differences between owners and

renters. Despite this important shortcoming these studies do shed light on whether homeownership is

as likely to support wealth accumulation among low-income and minority households.

Perhaps the most thorough of these studies is Di et al. (2004), which uses the Panel Survey of Income

Dynamics (PSID) to examine the relationship between homeownership and changes in household

wealth over the period from 1984 to 2001. The basic approach used in this paper is to estimate a

regression model that predicts the level of wealth in 2001 as a function of demographic

characteristics, starting wealth in 1984, and measures of the length of time spent as an owner over the

period since 1984. The results indicate that, as would be expected, average income and starting

wealth are the single most important predictors of wealth in 2001. But the next most important

factors are the length of time spent as an owner. Owning a home for between 1 and 5 years was

associated with an increase in wealth of more than $50,000, with each additional year owned up to 17

years adding about $6,700 a year to wealth on average. Those who owned homes for the entire 18-

year period between 1984 and 2001 had wealth levels that were $156,000 higher than those who

rented the whole period. Interestingly, a variable measuring the number of home sales during the

period was not significant, so the results do not provide any indication that more moves lowers wealth

as might be expected given the high transaction costs of moving.

Di et al. also examine whether the factors predicting wealth levels differ for low-income owners.

When they estimate a separate model predicting 2001 wealth for those with average incomes over the

study period in the lowest quintile of the sample they reject the hypothesis that the same factors

predict wealth levels for both low- and upper-income households. This result indicates that the

process for accumulating wealth through homeownership for low-income owners is distinct from that

experienced by higher income owners. In fact, the results suggest that homeownership is arguably

more important in predicting wealth accumulation for low-income households. Among low-income

households, those owning for the entire 18-year period have nearly 8 times the average wealth of

those who rented for the entire period, while for higher-income households owning for the entire

period is associated with only having twice as much wealth as those who always rent. That said, it is

also true that low-income households accumulate much less wealth than higher-income households

with or without homeownership. Among low-income households, owning a home for 18 years

increases average wealth to $49,700 compared to $5,700 for those who rented the entire period. But

among higher-income households, wealth increases to $260,000 for those who owned the whole time

compared to $84,000 for those who only rented.

Chapter 4: Financial Benefits of Homeownership

Finally, Di et al. also investigate the issue of whether holding wealth in stocks rather than housing

helps to increase wealth. This question relates to the issue of whether holding wealth in non-housing

assets rather than housing would result in higher wealth levels. That is, even though homeownership

is found to increase wealth, it might still be the case that devoting savings to stocks or other financial

assets would create even greater wealth. Di et al. indicate that their interest in this question in part in

response to findings by Hurst et al. (1998) who found that the level of housing wealth in 1989 was

negatively associated with overall wealth in 1994. However, Di et al. argue that this result was

probably due to the time period studied, which coincided with a period when house prices declined in

many markets.

To test the impact of holding stocks on changes in wealth between 1984 and 2001, Di et al. include a

variable that indicates whether the household held a greater share of their 1984 wealth in stocks than

in housing equity. Since the model already includes a measure of the overall level of wealth, this

variable provides an indication of whether less concentration of wealth in housing produced greater

financial returns given the same starting wealth. The estimated coefficient on this variable is both

positive and statistically significant, indicating that holding more wealth in stocks did, in fact, result

in higher wealth levels. But the results also indicate that the gains in wealth due to homeownership

were much larger than the gains due to holding more wealth in stocks. For example, a household that

rented for the entire 18-year period but held stock wealth in 1984 is estimated to have average wealth

in 2001 of $47,857. In comparison, if the household did not hold stock in 1984 but owned a home for

only one to five years, their wealth is estimated to be $80,899 on average, or nearly 70 percent more

than if they had just invested in stocks. Thus, while holding more of wealth in stocks does increase

wealth in the long run, it still the case that being a homeowner is associated with greater increases in

wealth than if the household just held stocks. This result provides support for the argument that there

is a “forced savings” aspect of homeownership that is nontrivial. Homeownership is thought to force

savings by requiring owners to make payments toward mortgage principal each month. This accrual

of equity in the home due to mortgage amortization can be a significant source of wealth over the life

of a mortgage.

This result is similar to the findings of Krumm and Kelly (1989) who explored the impact of

homeownership on total savings levels. They used data from the 1976 Survey on Consumer Credit to

examine the relationship between household characteristics and both non-housing and total wealth.

They explored several different modeling approaches to control for the fact that the level of savings

and the level of non-housing wealth are endogenous with respect to the decision to own. They found

that while the magnitude of their results were somewhat sensitive to the approach used, in general the

findings were consistent across approaches. Specifically, they found that the level of non-housing

savings was either no different or slightly lower when owners were compared with renters. This

suggests that owning either has little impact on non-housing savings or actually reduces it. But,

importantly, they found that total wealth levels were always higher for owners compared to renters as

the accumulated wealth in home equity made up for lower non-housing savings. The differences in

wealth levels were also substantial between owners and renters, with owners’ wealth anywhere from

95 to 180 percent greater depending upon the modeling specification used.

Another recent study that uses the PSID to examine the association between homeownership and

changes in wealth over time is Reid (2004). While Reid does not model the factors associated with

wealth levels, she is able to use the panel nature of the PSID to identify consistently low-income

Chapter 4: Financial Benefits of Homeownership

households that remained renters throughout the period from 1976 through 1994. Reid defines low-

income households as those whose income remained below 80 percent of the area median income in

every year they were surveyed up until the year they purchased a home (or throughout the entire

period if they never bought a home). Reid compares wealth levels in 1994 of low-income households

who never purchased a home with those who did become owners. She finds that while both low-

income whites and minorities who were always renters had essentially no wealth in 1994, those who

had become owners had roughly $25,000 to $30,000 in wealth on average, with a large majority of

this wealth in the form of home equity. It is also the case that she found wealth levels among low-

income owners to be much less than for higher income owners. Thus, consistent with the findings of

Di et al., Reid finds that low-income owners do not accumulate nearly the same amount of wealth as

higher-income owners, but they nonetheless accumulate more wealth than renters.

Finally, Boehm and Schlottmann (2004c) provide further evidence of the importance of

homeownership to overall wealth creation for low-income households, defined in their study as

households with income below the area median income when they are first observed in 1984. Like

the two previous studies discussed, they also use the PSID, this time covering the period from 1984 to

1992, although their approach is somewhat different. Their analysis begins by estimating a hazard

model to predict movements into and out of homeownership over the nine-year period studied. They

then model the value of homes purchased by movers. They combine the results of these models with

information on average annual house price appreciation rates in the census tracts where respondents

lived based on changes in median house values between the 1990 and 2000 decennial censuses. More

specifically, their approach is to estimate the probability that a specific household will own a home in

a given year and, if they do own, how much the home will be worth. They then apply the average

annual change in house prices in the census tract where they live to predict the change in housing

equity in that year. By mapping out the probability of different tenure paths over the nine year

period, the combined models yield an estimate of the overall appreciation each individual would be

likely to experience over the time span studied.

There are several points worth noting about their approach. Their estimates of housing wealth ignore

the transaction costs of moving and so they may overstate the gains from homeownership. On the

other hand, their basic analysis also ignores growth in equity from amortization of mortgage debt.

However, they do present some estimates of the potential value of amortization, which illustrate that

these gains over a nine-year period are likely to be small. Perhaps the most interesting aspect of their

analysis is the fact that they track the neighborhood choices of these households over time. Thus,

they provide some important insights into how differences in neighborhood choices of owners by

income level and race can affect the financial returns to homeownership.

Their results indicate that the financial returns to homeownership are, in fact, greater for higher

income groups and whites, but that low-income and minority owners are still likely to experience

significant positive financial gains from homeownership. Over the period they study, the median

high-income white household is estimated to accrue about $4,500 annually in capital gains from

homeownership, compared to $3,400 for high-income minorities, $2,700 for low-income whites, and

$1,700 for low-income minorities. The larger gains for whites and higher-income households reflects

both the greater propensity for these households to be owners in any given year and the fact that they

own more expensive houses and so have larger absolute gains from increases in house values. They

also simulate the potential annual increase in housing wealth from amortization and find that this

Chapter 4: Financial Benefits of Homeownership

would add between $200 and $400 a year to the financial gains realized, or roughly 10 percent of the

wealth gained through appreciation.

But as the authors note, while the gains realized by low-income and minority owners are smaller than

for other owners, the gains are still positive and non-trivial. To put these gains in perspective, Boehm

and Schlottmann also use the panel nature of the PSID to estimate the annual gain in non-housing

wealth for each of these household types over the period from 1984 to 1992. They find that the

median low-income minority household did not accrue any non-housing savings over the period,

while low-income whites only gained $300 a year on average. They conclude that not only is

homeownership an important means of wealth accumulation for low-income families, but for the

majority of these households it is the only form of wealth accumulation.

Summary and Conclusions

One of the principal factors underlying the strong support from policy makers for efforts to improve

homeownership opportunities among low-income and minority families is the strong association

between homeownership and the creation of household wealth. One of the key determinants of the

financial gains to homeownership is appreciation in house values over time. Studies that have

examined the financial return to homeownership taking into account average rates of appreciation as

well as the value of housing services provided by the home, the income tax benefits of

homeownership, and the use of financial leverage by the typical homebuyer have found that

homeownership does belong as part of an optimal portfolio of household investments.

But given that low-income and minority homebuyers are buying homes in different market segments

than higher-income and white households, it is possible that they would be less likely to realize

financial gains from appreciating values. Looking collectively at a broad range of studies that have

analyzed differences in housing appreciation rates across different market segments defined in terms

of house values, it is clear that there is no systematic tendency for low-income areas or low-valued

homes to appreciate less. Indeed, there have been a number of periods when low-valued homes in

individual markets have appreciated more rapidly than higher-valued homes. Thus, it is not the case

that low-income owners are less likely to experience real gains in home values.

But while there have been a large number of studies examining differences in appreciation rates by

house value, there has not been nearly as much attention paid to differences in housing appreciation

by either the race-ethnicity of the neighborhood or of individual owners. Among the studies that have

been done two suggest that areas with a greater concentration of blacks have had lower rates of

appreciation in housing values, while a single study has found that higher Hispanic shares were

associated with greater price appreciation in two Florida markets. However, one study that examined

differences in appreciation rates by the minority share in the neighborhood over a longer period of

time and across several markets found that whether minority areas appreciated faster or slower than

whites varied over time and across markets. In short, there is no evidence that house prices in

minority areas systematically appreciate more slowly, but given the small number of studies on this

topic it is not possible to draw general conclusions. Further study of the degree to which minority

homebuyers experience housing appreciation would make a valuable addition to the literature.

Chapter 4: Financial Benefits of Homeownership

Another factor to consider in evaluating differences in appreciation rates is whether there are

differences in appreciation rates for different structure types. The vast majority of studies of

appreciation rates have focused exclusively on single-family attached housing. But in the 1990s

manufactured housing emerged as an important choice for low-income first-time buyers, accounting

for nearly a quarter of homes bought by this group. Several studies have compared appreciation rates

of manufactured homes using different methodologies and covering different geographic areas and

time periods. But despite these differences the conclusions of these studies are consistent.

Manufactured homes where the land is owned appreciate at rates close to that of site built homes,

although there is greater variability in price trends. However, in cases where owners do not own the

land along with their home, which is true of about half of manufactured home owners, there is

essentially no appreciation in value. While owners of these homes may well experience other

benefits of homeownership due to security of tenure and control over their home, they are very

unlikely to realize any of the financial benefits of homeownership. To the extent that manufactured

homes are seen as providing an affordable entry point into homeownership, policy makers should

strongly encourage developments that give buyers the opportunity to own the land.

While manufactured homes are the most important alternative structure type for low-income buyers,

it is also the case that in some market areas blacks are more likely to occupy single family attached

housing, while in high cost markets first-time buyers are more likely to occupy condominiums in

multiunit structures. There are very few studies that have analyzed differences in appreciation rates

by structure type, but those that have generally conclude that rates of appreciation do differ. While

there are too few studies to draw general conclusions, it does appear that the variation in house prices

is higher for condominiums. Thus, first-time buyers in this type of home may face greater risk.

One important issue highlighted by the literature assessing differences in appreciation rates is that

investments in homes are clearly not without risk. While over the longer run homes in the vast

majority of markets do appreciate in real terms, in the short run there can be substantial fluctuations

in home prices. A study by Belsky and Duda (2002) found that while homes affordable to low-

income buyers experienced much greater appreciation than those affordable to either moderate- or

high-income buyers, it was nonetheless the case that a large share of low-income buyers lost money

in real terms on their housing investments during the 1980s and 1990s. Losses resulted both because

of soft housing markets during the early 1990s and because of the high transaction costs associated

with selling a home. This finding highlights the pitfalls of homeownership when owners are forced to

sell homes into a down market. In order for owners to realize gains in house values it is important

that they are able to sustain homeownership through periods of weak home prices. Belsky and Duda

did find that homebuyers in the low-income segment were less likely to incur losses because buyers

were less likely to buy during periods of peak home prices. But the authors speculate that support for

low-income homeownership during the period of booming prices over the last few years could mean

that these buyers will be more exposed to financial losses when housing prices next soften. The

findings of Belsky and Duda provide further support for the need to ensure that low-income buyers

are able to sustain homeownership in order to realize any financial benefits.

But while it may be that low-income and minority homeowners are as likely as other owners to

experience appreciating home values, it is still possible that individuals could be financially better off

by renting a home and devoting their savings to other investments. There are several reasons why

renting might be a better deal for low-income and minority households compared to white and higher-

Chapter 4: Financial Benefits of Homeownership

income owners. First, low-income households are less likely to realize tax benefits of any

significance, which may tip the scales in favor of renting. Second, since low-income and minority

owners are more likely to use subprime lending, the higher cost of owning may make it less attractive.

On the other hand, one study found that rents are higher relative to house values in the low-cost

segment of the market, which helps to tip the scales in favor of owning. However, there is very little

research examining the issue of how the ratio of rents to house values differs across different

segments of the housing market, so it is not clear whether this relationship holds up over time and

across markets.

It is difficult to draw any definitive conclusions from studies that have compared the costs of owning

and renting. It is fairly clear that given the high transaction costs associated with buying and selling a

home, owning is rarely a wise choice if the household is unlikely to stay in the home for at least three

years. Beyond that, the conclusions of these studies are highly sensitive to assumptions about the

level of rents relative to house prices. When higher rents are assumed, low-income households are

found to be better off owning even when they do not realize any tax benefits. However, at lower rent

levels, the lack of any tax benefits from owning can make renting a better deal for low-income

owners. Further research on the difference in rents relative to values for different segments of the

housing market would help to clarify whether owning is cheaper than renting for low-income

households. Nonetheless, the research that does exist indicates that it is likely that in situations where

house prices are growing slowly or declining, households of all income levels may be financially

better off renting.

In short, it is difficult to conclude whether low-income and minority households are likely to realize

financial benefits from homeownership given the complex web of factors that contribute to the

outcomes, including the location and timing of purchase and sale, the ability to sustain

homeownership over time, the availability of tax benefits, and the choices made about financing,

maintaining, and improving homes. One approach to account for all of the ways in which

homeowners differ in all of these areas based on income and race-ethnicity is simply to observe

changes in wealth of individual households over time and how this correlates with their tenure

choices. In fact, there have been several studies that have used longitudinal survey data to undertake

this type of analysis. Given their nature, these studies inherently account for the multifaceted array of

factors that contribute to wealth accumulation through homeownership over time. An important

caveat with these studies is that they do not incorporate any controls for selection bias in who

becomes an owner, and so it is not clear whether the observed differences in wealth accumulation

reflect solely the impact of homeownership or whether the result is due to other unobserved

differences between owners and renters. Despite this important shortcoming these studies do shed

light on whether homeownership is as likely to support wealth accumulation among low-income and

minority households.

The general conclusion that emerges from these studies is that spells of homeownership are strongly

associated with higher levels of wealth, with longer periods spent as an owner associated with ever-

greater wealth. In contrast, those who remain renters for long periods of time accumulate much less

wealth, even after accounting for income and starting wealth. It is the case that low-income owners

accumulate much less wealth than higher income owners, both because low-income households own

lower-valued homes and because they spend less time as owners. But on average low-income owners

nonetheless accumulate non-trivial amounts of wealth through homeownership. Importantly, these

Chapter 4: Financial Benefits of Homeownership

studies also find that there is little or no accumulation of non-housing wealth among low-income

owners or renters. Thus, homeownership is often the only form of wealth accumulation for these

households.

The general conclusion that emerges from the review of this literature is that homeownership can be a

good investment for low-income and minority households—but it can be risky. While many

households will experience financial losses, particularly over short spells as owners, on average low-

income owners do accumulate wealth over time. But the literature once again highlights the

importance of sustaining homeownership over time. Those who can sustain homeownership through

periods of weak prices are extremely likely to realize real returns on their investments. As a result, it

is vitally important that policy makers ensure that low-income owners have the support they need

either to weather the crises that threaten their ability to keep their homes, or, if giving up

homeownership temporarily is their best option, to regain homeownership in the future.

However, even if the financial benefits of homeownership are small or non-existent, low-income and

minority households might still prefer to own given the appeal of various non-financial benefits of

homeownership. To assess this issue, in the next chapter we review the literature that has assessed

the degree to which homeowners are likely to experience positive outcomes in terms of impacts on

their children, increased social involvement, and improved psychological and physical health.

Chapter 4: Financial Benefits of Homeownership

Chapter Five:

Social Impacts of Homeownership

While the association between owning a home and wealth creation is an important part of the appeal

of homeownership as a policy goal, policy makers are also quick to cite a variety of non-financial

benefits as justification for efforts to increase homeownership. Some of these non-financial benefits

were touched upon in Chapter 2, including greater satisfaction among owners with their homes and

neighborhoods. But there are other non-financial benefits—generally referred to as social benefits—

that go beyond merely being more satisfied with one’s home. These include positive impacts on

children growing up in owner-occupied homes, increased involvement in community affairs by

owners with potentially positive impacts on surrounding communities, and improved psychological

and physical health among owners. This chapter reviews the literature that has examined

homeownership’s impacts in each of these spheres.

Before turning to these specific topics, there are several broad issues about this literature to note. One

significant challenge that plagues research on the social impacts of homeownership is the fact that

people who choose to become owners are, on average, likely to be different from renters in important

ways that may not be apparent from available data. This is largely because households with certain

propensities self select into homeownership. For example, given the high transaction costs associated

with buying and selling a home, households expecting to stay longer in one home are more apt to

elect to own. This reduced residential mobility, rather than homeownership itself and the behaviors

homeownership may evoke, may be in large part responsible for the impacts associated with

homeownership, such as positive impacts on children and greater social involvement. Another reason

why people may be more drawn to homeownership is that they prefer to live a more home-centered

life, and so are motivated to invest in a larger, higher quality home. The increased quality of the

house and the focus on spending more time at home as a family may also contribute to some of the

impacts associated with homeownership. Finally, to the extent that homeowners tend to cluster in

neighborhoods—and they do in many cases—there may be spillover benefits from living in areas

where residential mobility is lower and where household incomes are higher. But again, to the extent

that reduced residential mobility and greater income mixing yield positive social impacts, it may be

possible to produce these conditions by means other than promoting homeownership (as discussed by

Apgar (2004)).

While some of these factors might be captured by observable characteristics of the household, the

home, or the neighborhood, many of these factors may not be easy to capture with survey data. As a

result, one issue that plagues research on the impacts of homeownership, particularly with regard to

these social impacts, is how to account for the unobservable factors that both lead people to choose

whether to own and rent and are also likely to affect the outcomes of interest. In social science

research, the issue of selection bias is ideally addressed through studies that employ an experimental

design where participants in the study are randomly assigned into treatment and control groups. But

for a variety of reasons, this approach has not been used to study the impacts of homeownership. As

a fallback, researchers employ statistical techniques to try to account for this selection bias. The most

common approach is to first estimate the likelihood of homeownership for a household using

observable factors, at least some of which would not be expected to influence the social outcome of

Chapter 5: Social Impacts of Homeownership

100

interest. This estimate of whether a household is likely to be a homeowner is then used to test the

influence of homeownership on the outcome of interest. While not a perfect solution for the problem

of selection bias, such estimation techniques provide at least a partial test of whether

homeownership’s impacts are likely due to selection bias.

In the review that follows, studies that

include such tests are regarded as providing greater evidence of homeownership’s likely impacts.

Another concern with the existing literature is that many studies do not include measures of the

confounding factors that may help produce the outcomes associated with homeownership, most

notably residential stability and housing quality. Studies that do include measures of these factors

provide a better test of homeownership’s independent effect on social outcomes as well as the

mechanism by which homeownership may produce the outcomes of interest.

Finally, the focus of this study is in assessing whether there are any differences in the likelihood of

realizing the benefits of homeownership by the income or race-ethnicity of the owner. Much of the

literature on the social impacts of homeownership is aimed at assessing whether there is an

association generally between homeownership and the outcomes of interest and so shed little light on

whether there are differences by income or race.

Nonetheless, there have been a few studies with a

particular focus on assessing outcomes among low-income homeowners. These studies are given

particular attention in our review. Virtually no studies have assessed differences in social outcomes

by the owner’s race-ethnicity, and so this issue is not addressed in this review.

Impacts on Children

Homeownership is purported to have a variety of positive impacts on children, including higher

educational attainment, greater success in labor markets, fewer behavioral problems, and higher rates

of homeownership as adults. Synthesizing the various theories presented in the literature, Harkness

and Newman (2002) identify four pathways by which homeownership may produce these positive

impacts on children. To begin with, there is evidence that homeownership may be associated with a

more stimulating and emotionally supportive home environment. In support of this view, Haurin,

Parcel, and Haurin (2002) find that there is a statistically significant positive association between

homeownership and indicators of a more nurturing home environment even after controlling for a

variety of household characteristics and employing statistical controls for selection bias in who

becomes an owner. What is not clear is exactly why homeownership would lead to a more supportive

home environment. One hypothesis is that owners have greater life satisfaction and self-esteem,

which helps foster this environment. Another argument is that owners are more likely to make

investments in their home to tailor it to fit their tastes, which supports a more home-centered life.

Another way in which homeownership may have a positive impact is by providing a better physical

environment for children. Better physical conditions may improve children’s physical health by

For a thorough discussion of the issue of selection bias as it relates to the social benefits of homeownership

and the statistical techniques available to address this problem, see Dietz and Haurin (2003).

Two recent excellent reviews of this literature in general are Rohe, McCarthy, and Van Zandt (2002) and

Dietz and Haurin (2003).

Chapter 5: Social Impacts of Homeownership

101

reducing the risk of illness or injury due to such factors as improperly functioning heating and cooling

systems, infestations of insects or rodents, or exposure to hazards such as lead paint. Improved

physical health in turn may contribute to better performance in school and to greater ability to interact

socially with others. Furthermore, to the extent that owner-occupied homes tend to be larger single-

family units, children may also benefit from having greater physical space and privacy to do school

work or pursue other interests.

A third way in which homeownership may help produce positive outcomes for children is by helping

to promote residential stability by insulating the family from the need to move at a property owner’s

discretion. Owners may also be more reluctant to move because of the higher transaction costs

associated with moving. Residential stability has been found to be associated with better educational

outcomes (Hanushek, Kain, and Rivkin, 2004), and may help foster greater social connections that

enhance a child’s self-esteem and provide greater opportunities for social engagement.

Finally, it is also hypothesized that the greater wealth accumulation associated with homeownership

may confer a variety of benefits both by providing a financial cushion that can be used in times of

need to provide a more stable home environment and by making it feasible to invest in education.

One of the issues that arises in the literature assessing the impact of homeownership on children is

whether homeownership ought to be given credit for the benefits associated with characteristics of the

living environment that could be achieved without being an owner. For example, while

homeownership may be associated with better physical housing conditions, it should be possible to

obtain high quality rental housing as well. Some of the literature addresses this issue by including

controls for housing quality and assessing the effect of including these controls on the estimated

homeownership impact.

Perhaps more challenging is the question of whether homeownership ought to be credited with the

benefits associated with reduced residential mobility, which is consistently found to play an important

role in a variety of outcomes. Homeownership does insulate households from the need to move as a

result of choices made by the property owner about the use of the home. On the other hand, because

of the high moving costs faced by owners, households that do not expect to move will be more

attracted to homeownership. As a result, it is hard to identify the extent to which homeownership

produces residential stability from the extent to which expected residential stability leads to the

choice to own. In addition, some authors argue that to the extent that reduced mobility is the

mechanism by which the benefits are realized, other policies to promote residential stability could

produce the same result. There are differences of opinion in the literature regarding whether the

benefits of reduced mobility ought to be attributed to homeownership. For example, while Harkness

and Newman (2002) do not include controls for residential mobility in their analysis so that the tenure

variable will capture this effect, Aaronson (2000) explicitly attempts to isolate the effects of

residential stability from the effects of homeownership. To the extent that studies include controls for

residential mobility the importance of this issue can be separated from other ways in which

homeownership influences educational outcomes.

In the sections that follow we discuss in turn findings from the literature regarding the impacts of

homeownership on children’s educational outcomes, success in labor markets, behavioral problems,

and homeownership.

Chapter 5: Social Impacts of Homeownership

102

Educational Outcomes

There have been a number of high quality studies that have investigated the impacts of

homeownership on the educational attainment of children. The studies differ in the types of

educational outcomes examined, the data sets used, and the methodological approaches employed.

Yet, these studies universally conclude that the children of homeowners have better educational

outcomes than the children of renters even after controlling for a wide variety of other household

characteristics and employing statistical methods to account for selection bias in who becomes an

owner.

Among the first studies to address this question was Green and White (1997). Using data from the

Panel Study of Income Dynamics, they estimate the probability that 17 year olds were either still in

school or had graduated from high school. The explanatory variable of interest is whether the child’s

parents were homeowners, but they also control for race, household income, parental education, and

other household characteristics. They attempt to control for selection bias in who becomes a

homeowner by estimating a bivariate probit model of the joint outcomes of housing tenure and

educational outcomes for children. Green and White find that the 17-year old children of owners are,

in fact, more likely to be in school than the children of renters. Importantly, they also find that the

impacts of homeownership on the probability of being in school are larger for low-income families.

Children in homeowner households with incomes below $10,000 are found to be 19 percentage points

more likely to be in school than the children of renters, while among owner households with incomes

above $40,000 the difference between owners and renters is only 12 percentage points.

Green and White also test their results by examining another data source, the 1990 decennial census,

and produce results that are similar to those found using the PSID. In order to test whether the

homeownership effect found using the PSID could be attributed to homeowners living in higher

quality housing or having longer duration of residence in a given location, their analysis using census

data also incorporates measures of housing quality (as captured by house value or rent) as well as

length of time residing in the house. Even after adding these additional control variables the results

indicate that homeownership has a statistically significant independent effect on increasing the

likelihood of being in school at age 17. However, they do not attempt to control for selection bias in

who becomes an owner in their analysis of decennial census data.

In his analysis of data from the PSID, Aaronson (2000) attempts to extend the work of Green and

White in two ways. First, he introduces a much broader set of control variables in order to explore

the specific mechanism by which better educational outcomes are obtained. Second, he employs a

different methodology to control for potential selection bias in who becomes a homeowner.

Aaronson begins by estimating a model that includes a similar set of explanatory variables as used by

Green and White, including controls for family composition, income, and parental education. The

results indicate that children of homeowners have a likelihood of graduating from high school that is

10 percentage points higher compared to the children of renters. He then examines whether this

effect is the result of greater household stability by incorporating measures of change over time in

employment, marital status, and household composition. He finds that adding these controls does not

affect the estimated homeownership impact. He also employs a different data set to examine whether

Chapter 5: Social Impacts of Homeownership

103

a parental inclination toward education might account for homeownership’s impacts. In this analysis

he includes measures of parental IQ, membership in parent-teacher organizations, and whether

parents read newspapers or have a library card. But none of these measures is found to diminish the

estimated impact of homeownership on high school graduation rates.

When Aaronson adds measures of residential mobility, including both how often the household

moved when children were between ages 7 and 16 and the maximum duration of continual residence

during those years, he finds that the estimated impact of homeownership on high school graduation

rates is halved, from 10 percentage points to 5. Aaronson concludes that a good deal of the impact of

homeownership is, in fact, attributable to greater residential stability that is correlated with owning.

Finally, Aaronson adds further controls to account for differences in household wealth, including the

amount of housing equity. Including housing wealth in the estimated model is found to further reduce

the estimated impact of homeownership on high school graduation by about half, with greater levels

of housing equity associated with a greater likelihood of graduation. While Aaronson hypothesizes

that this result reflects the association between wealth and other household characteristics that affect

well being, he also notes that non-housing wealth does not have the same positive association with

graduation rates. This result is consistent with the argument made by others that housing wealth is

indicative of larger and higher quality homes, which may support better educational outcomes.

Nonetheless, while Aaronson finds that including controls for residential mobility and wealth reduces

the impact of homeownership, it is still the case that an independent and statistically significant

association between homeownership and high school graduation remains. This is consistent with the

results obtained by Green and White in their analysis of decennial census data with similar controls.

In two related studies, Harkness and Newman (2002, 2003) make several important contributions to

the existing literature. First, they focus their analysis specifically on low-income households (defined

as those with income less than 150 percent of the federal poverty definition) to examine whether low-

income households are as likely as higher income groups to realize the benefits of homeownership.

Second, they introduce controls for neighborhood characteristics in order to examine the extent to

which the realization of benefits of homeownership may vary depending upon the socioeconomic

status of the neighborhood. Their analysis of the PSID finds that by age 20 the children of

homeowners have on average completed a half year more of schooling, are 13 percentage points more

likely to have graduated from high school, and are 6 percentage points more likely to have obtained

some postsecondary education.

Harkness and Newman (2003) also compare the magnitude of homeownership’s effects between low-

and higher-income households. They find that homeownership’s positive impacts are consistently

larger in low-income families. Aaronson’s results also support this conclusion. While Aaronson does

not sort his sample into low- and high-income households, he does estimate separate models for low-

and high-income neighborhoods. He finds that the benefits of homeownership are statistically

significant in low-income areas but not in high-income areas. Since low-income owners are more

likely to live in low-income areas, this result is consistent with the findings of Harkness and Newman.

Harkness and Newman (2003) also test the sensitivity of their results to the use of four different

instrumental variables to control for selection bias in who chooses to own a home. They find that for

Chapter 5: Social Impacts of Homeownership

104

three of the four instruments the homeownership effect is still statistically significant with regard to

educational outcomes for low-income children. This result leads them to conclude that the findings

are robust even when using a variety of controls for selection bias. However, they find that for

higher-income families, the use of these instruments results in a loss of significance for the

homeownership variable. Thus, for higher-income households there is less evidence of an impact of

homeownership once controls are implemented for the selection bias in who becomes an owner.

While the basic model presented in Harkness and Newman (2002) does not include controls for

residential mobility or housing equity, they do test for the impact of including these variables on the

estimated homeownership effect. Consistent with Aaronson’s results, the inclusion of measures of

residential mobility does diminish the homeownership effect, but it still remains positive and

statistically significant. Housing equity, on the other hand, is not statistically significant. Boehm and

Schlottmann (1999) also use the PSID to estimate the impact of homeownership on children’s

educational attainment. They find that housing equity is not statistically significant in predicting

graduation from high school, but it is significant in predicting graduation from college. This result is

consistent with the hypothesis that the wealth generated through homeownership may make it

financially feasible to attend college.

The other principal contribution made by Harkness and Newman is to incorporate measures of

neighborhood socioeconomic status, as captured by the share of residents in their homes for five years

or more, the poverty rate, and the homeownership rate. Their results indicate that the effect of

neighborhood characteristics on educational outcomes is weak, with only neighborhood stability

being marginally statistically significant. However, when they interact the individual’s tenure status

with neighborhood characteristics they find that neighborhood characteristics have a greater impact

on owners compared to renters. In particular, greater neighborhood stability is found to have more of

an impact on owners’ children. This is consistent with findings by Aaronson that the positive impacts

of homeownership on high school graduation rates are larger in neighborhoods with low mobility.

On its face, Harkness and Newman’s finding that the children of low-income owners are more

sensitive to neighborhood stability would suggest that homeownership in unstable communities

would have more deleterious effects on owners than renters.

However, they find that when the

positive effects of homeownership itself are considered, the children of owners living in unstable

neighborhoods are still found to have higher educational outcomes than renters’ children in these

areas. In short, they conclude that homeownership is beneficial to low-income families even if they

live in neighborhoods with low socioeconomic status.

Haurin, Parcel, and Haurin (2002) take a somewhat different approach to evaluating the impact of

homeownership on educational outcomes. Rather than look at the level of educational attainment,

they use data from the National Longitudinal Survey of Youth (NLSY) to examine the association

between homeownership and results on math and reading achievement tests. They find that

homeownership has a positive and significant effect on test results for the children of owners—on

average raising math scores by 9 percent and reading scores by 7 percent. The positive influence of

homeownership remains even when controls are incorporated to account for sample selection bias.

The fact that there are statistically significant, positive impacts of residential stability on educational

outcomes implies that a lack of stability will have negative impacts on these outcomes.

Chapter 5: Social Impacts of Homeownership

105

Thus, in addition to evidence that owners’ children are more likely to stay in school longer, Haurin,

Parcel, and Haurin provide evidence that the academic achievement of these children is also higher.

Employment, Earnings, and Welfare Use

Two studies have assessed the impact of homeownership on the labor market outcomes realized by

the children of homeowners, including the wage rates they achieve as young adults, the likelihood

that they will be idle at age 20 (that is, neither employed or in school), and whether they are more

likely to receive welfare as young adults. The results suggest that homeownership is associated with

at least moderately positive outcomes for children in labor markets. Harkness and Newman (2002)

find that homeowners’ children are 7 percentage points less likely to be idle at age 20, have average

wage rates that are $0.70 per hour higher between the ages of 24 and 28, and are 9 percentage points

less likely to receive welfare between the ages of 24 and 28. However, when controls for

neighborhood characteristics are introduced, the impacts on idleness and wage rates are no longer

statistically significant, although owners’ children are still less likely to receive welfare. When they

employ instrumental variables to control for selection bias in who becomes an owner (Harkness and

Newman, 2003), they do not find a statistically significant impact on idleness, but there is a

significant and positive association between homeownership and both wages and reduced welfare use.

In their analysis of the PSID, Boehm and Schlottmann (1999) examine the association between

homeownership and children’s average wages 10 years after leaving their parents’ home. While their

results do not find a statistically significant direct effect of homeownership on children’s earnings,

they note that homeownership does have an indirect effect on wage rates through its statistically

significant association with increased educational attainment. Using the results of their models, they

find that the increase in educational attainment that is associated with growing up in an owner-

occupied home produces an increase in average annual earnings of $7,500.

Teenage Pregnancy and Behavioral Problems

Several studies have investigated the association between homeownership and the incidence of

teenage pregnancy or behavioral problems. Green and White (1997) use the High School and Beyond

survey to evaluate whether the daughters of homeowners are less likely to have had a child by age 18.

While they do find a positive impact of homeownership the magnitude is fairly small, reducing the

likelihood of having a child by only 2 percentage points. Using the PSID, Harkness and Newman

find that the children of homeowners have about a 3 percentage point lower chance of having a child

by age 20, but this difference is not statistically significant in any of the specifications tested.

Haurin, Parcel, and Haurin (2002) evaluate the association between growing up in an owner-occupied

home and an index of behavioral problems as measured by mothers’ responses to 28 questions in the

National Longitudinal Survey of Youth about the prevalence of behaviors such as acting out, having a

strong temper, demanding attention, and being depressed or anxious. Their results indicate that

homeownership is associated with a 3 percent reduction in the incidence of problematic behaviors,

but the effect is not statistically significant.

Taken together these papers suggest that homeownership may have some positive impact on

children’s behaviors, but if so the magnitude of this impact is fairly small.

Chapter 5: Social Impacts of Homeownership

106

Homeownership

One last outcome that has been examined in the literature is whether the children of homeowners are

more likely themselves to become owners. Boehm and Schlottmann (1999) use the PSID to examine

the homeownership rates of children 10 years after leaving their parents home. They find that even

after controlling for the usual predictors of homeownership, such as income and marital status, the

children of owners have homeownership rates that are 25 percentage points higher than the children

of renters. This much greater tendency to own may reflect some combination of a greater preference

for homeownership among those who have experienced it, greater comfort and familiarity with what

is entailed in being a homeowner, or greater parental wealth that can be tapped to help achieve

homeownership. While Boehm and Schlottmann do not attempt to control for the selection bias in

who chooses to become an owner, the rather substantial increase in the propensity to own a home

among those who grew up in owner-occupied housing would seem likely to remain even if such

controls were employed.

Conclusions Regarding Impacts on Children

The literature that has examined homeownership’s impacts on children is among the most convincing

work on the social impacts of homeownership. A series of studies have examined this issue using

different data sets and employing a variety of methods to control for selection bias in who becomes an

owner. These studies have also attempted to isolate the effect of homeownership from the impact of

reduced residential mobility and neighborhood attributes. The results strongly suggest that

homeownership has a significant, positive effect on children’s educational attainment even after

controlling for all of these potentially confounding factors. Importantly, the literature suggests that

these positive impacts are larger for low-income families and outweigh any negative impacts

associated with owning in distressed communities.

There is also somewhat weaker evidence that homeownership is associated with better labor market

outcomes and reduced behavioral problems, although the magnitude of these impacts is small enough

that the statistical significance of the results is sensitive to the controls employed. Finally, one study

has also found a significant association between growing up in an owner-occupied home and the odds

of becoming a homeowner as a young adult.

Impacts on Social Involvement

One of the frequently touted benefits of homeownership is that owners are more engaged in efforts to

improve the community. Thus, increases in homeownership are thought to create more stable and

healthier neighborhoods. There are a number of arguments put forward in the literature to explain

why homeowners are thought to be more likely to be engaged in efforts to improve their communities

(Cox, 1982; Baum and Kingston, 1984; Rohe and Stegman, 1994; DiPasquale and Glaeser, 1999;

Rohe, McCarthy, and Van Zandt, 2002; Dietz and Haurin, 2003). Since neighborhood conditions

have an effect on housing values, owners have a strong financial incentive to work to improve their

communities. In addition to their financial stake, owners are also likely to have an emotional stake in

their homes and a pride of ownership that will motivate them to improve the surrounding community.

Owners also face higher moving costs than renters, so they may be more motivated to work to solve

Chapter 5: Social Impacts of Homeownership

107

neighborhood problems since it is harder for them to move out. Finally, owners’ longer duration of

residence in a neighborhood may also increase the strength and number of relationships they have in

the community, which increases both their willingness and ability to engage in efforts to improve the

neighborhood.

The existing literature has examined several of dimensions of social involvement. One aspect is the

likelihood that a household will be engaged in political affairs as evidenced by how frequently they

vote or whether they know the names of elected officials. Another measure of social involvement is

the degree to which individuals participate in local organizations and institutions. A final dimension

is the extent to which households are familiar and interact with neighbors.

However, much of the existing literature suffers from either a failure to account for selection bias or

does not attempt to evaluate whether homeownership’s impacts differ with either the income or race

of the owner. Also, studies on this topic are much less likely to include controls for residential

duration to separate the effects of homeownership from the effects of residential stability. The most

important study on this topic is DiPasquale and Glaeser (1999) because they not only introduce

controls for selection bias in who becomes an owner but they also investigate whether there are

differences in homeownership’s impacts between low- and high-income owners and assess the impact

of residential duration on estimated homeownership effect. Also of some importance are a series of

studies by Rohe and various colleagues (Rohe and Stegman, 1994a; and Rohe and Basolo, 1997)

based on surveys of participants in a low-income homeownership program in Baltimore and a

pseudo-control group of low-income renters from the same geographic area. These studies are

important because they focus explicitly on low-income households and because they examine

changes in household behavior following a move to homeownership.

Aside from these studies, there is a variety of other research that examines the relationship between

homeownership and measures of social involvement. While the lack of adequate controls for

selection bias and other confounding factors makes it difficult to draw definitive conclusions, we will

also briefly review these studies as they provide some indication of homeownership’s potential

impacts. In the sections that follow we examine findings regarding impacts on voting and political

participation, involvement with local organizations, and the extent of involvement with neighbors.

Voting and Political Participation

A number of studies have found that homeowners are more likely than renters both to vote in

elections (Kingston and Fries, 1994; Gilderbloom and Markham 1995; Rossi and Weber, 1996) and to

be knowledgeable about and engaged in the political process as evidenced by knowing the names of

elected officials, engaging in lobbying activities, or attending meetings about public affairs (Rossi and

Weber, 1996; Cox, 1982; Lyons and Lowery, 1989). However, none of these studies attempted to

include any controls for the selection bias in who opts to become a homeowner and, with one

exception, they do not investigate whether there are differences in owners’ political engagement by

income. The one exception is Gilderbloom and Markham (1995), who analyze data from the General

Social Survey regarding voting in the 1992 presidential election. They find that while owners with

incomes above the median had a greater likelihood of voting, owners with incomes below the median

were no more likely than renters to have voted.

Chapter 5: Social Impacts of Homeownership

108

DiPasquale and Glaeser (1999) make a significant contribution to the literature both by introducing

controls for the selection bias in choosing to be an owner and by assessing whether there are

differences in political activity by owner income. Their analysis of the General Social Survey for

1987 finds that homeowners are 13 percent more likely to know the name of the head of local

schools, 11 percent more likely to know the name of their U.S. Representative, and 16 percent more

likely to vote in local elections. When they add controls for duration of residence, they find the

impact of homeownership is roughly cut in half, but remains statistically significant. When they

introduce statistical controls for selection bias all of these impacts are still positive, although the

likelihood of knowing the name of the U.S. Representative is no longer statistically significant.

However, most importantly for our purposes, they also investigate whether there differences in the

impacts across different income groups. They find that compared to renters, owners with incomes in

the bottom quartile are only about 5 percent more likely to know the local school head, 7 percent

more likely to know their U.S. Representative, and 2 percent more likely to vote in local elections,

with none of these differences being statistically significant. Thus, while DiPasquale and Glaeser

find fairly convincing evidence that homeowners generally are more likely to be engaged in the

political process, they also find that owners in the bottom quartile of the income distribution are only

slightly more engaged than renters, if at all. This result is consistent with the findings of Gilderblum

and Markham regarding participation in the 1992 presidential election.

Involvement with Local Organizations and Institutions

Several studies have found that owners are more likely to be involved with local organizations such

as neighborhood associations, social organizations, school associations, nationality groups, or

churches, (Cox, 1982; Baum and Kingston, 1984; Guest and Oropesa, 1996; and Rossi and Weber,

1996). Again, for the most part these studies do not attempt to control for selection bias or assess

whether low-income owners behave differently than higher income owners.

The study by DiPasquale and Glaeser is once again important in this area. Analyzing data from the

General Social Survey for 1986 to 1994 they find that ownership is associated with greater

participation in non-professional organizations. This result is maintained even when instrumental

variables are used to control for selection bias. However, most of the homeownership effect is

eliminated when controls for residential duration are included. More importantly, they once again

find that owners in the bottom quartile of the income distribution are no more likely than renters to

belong to organizations. However, they also examine church attendance as another measure of

involvement with local institutions. They find that owners are more likely to go to church, and this

difference is statistically significant even when controlling for selection bias and length of residence.

Also, while low-income owners are less likely to attend church than high-income owners, they are

still more likely to attend church than renters.

Another measure of local involvement analyzed by DiPasquale and Glaeser is whether owners are

more likely to work to solve local problems, a question that was asked in the 1987 General Social

Survey. They find that owners are 10 percent more likely to be engaged in these efforts than renters,

although when statistical controls are introduced for selection bias the result is not statistically

significant. Interestingly, they do find that owners in the bottom quartile of the income distribution

are more likely than renters and no less likely than other owners to work to solve local problems.

Chapter 5: Social Impacts of Homeownership

109

However, the failure of this effect to be evident when controls are used for selection bias suggests that

the association between homeownership and working to solve local problems is at best weak.

As noted above, the studies by Rohe and Stegman (1994a) and Rohe and Basolo (1997) also make an

important contribution on this topic both because of their explicit focus on low-income households

and because they gather information on the behaviors of households both before and after becoming

owners. When matched with a comparison group of households who are otherwise similar but who

did not become owners, such a longitudinal study may be better able to identify household changes

that are related to the move to homeownership. Specifically, these studies track participants in a low-

income homeownership program in Baltimore from the early 1990s. As a comparison group, Rohe

and Stegman identified a group of households receiving rental housing vouchers in the city who had

demographic characteristics that were similar to participants in the homeownership program. The

comparison group provides a measure of the degree of change in behaviors and attitudes that might

have been expected over time due to such factors as aging of the respondents or changes in

circumstances in Baltimore during this period. Both groups were surveyed prior to the move to

homeownership to gather baseline information on their social and political involvement and measures

of self-esteem, perceived control over their life, and life satisfaction. The households were then

surveyed again approximately 18 months (as reported in Rohe and Stegman, 1994a) and 3 years after

buying a home (as reported in Rohe and Basolo, 1997) to gauge whether the move to homeownership

had been associated with any changes in these behaviors or attitudes that could be attributed with the

move to homeownership.

Before turning to their findings, several important caveats regarding the study are worth noting. First,

participants in the Baltimore homeownership program represent a unique group of low-income

homebuyers both because they were provided with fairly substantial subsidies (including a forgivable

loan worth nearly a quarter of the home’s value) and because they purchased newly constructed

homes that are clustered with other participants of this program. Another issue to note is that the

sample sizes available for study are fairly small and ultimately represent somewhat small shares of

those identified for study. Of the 171 homebuyers who participated in the program, 143 completed

the baseline survey, 125 completed a survey 18 months after purchase, and 90 completed a survey 3

years after purchase. Among the 202 households selected as a comparison group, 140 completed the

baseline survey, 101 completed a survey 18 months later, and 65 completed a survey 3 years later.

Thus, ultimately only about half of all homebuyers and a third of all renters were represented in the

survey 3 years after purchase. The small sample sizes will make it more difficult to detect an effect

of homeownership if magnitude of the impact is not large.

The studies by Rohe and his colleagues assess the degree of involvement with local organizations by

counting the number of school associations, political organizations, neighborhood associations,

church groups, and social organizations that respondents belonged to as well as the number of

meetings of each type they attended. Using multivariate techniques to control for other differences

between the two groups, they find that homebuyers were more likely to belong to and attend meetings

of neighborhood associations but were no more likely to be involved in other groups. This effect was

evident both 18 months and 3 years following purchase. On the other had, they also found that the

comparison group of renters were more likely to be involved with church groups 3 years after the

time of purchase.

Chapter 5: Social Impacts of Homeownership

110

Involvement with Neighbors

Another aspect of social involvement that has been assessed in the literature is the degree to which

homeowners interact with their neighbors. In general, there have fewer studies that have investigated

this issue and the results have been less favorable. Baum and Kingston find that owners were more

likely to have social relationships with neighbors. Rossi and Weber (1996), however, found that

renters were more likely to socialize with neighbors at night and to go out to bars, although measuring

the degree of interaction with neighbors as the extent of socializing at night is a somewhat limited

view of neighborly attachment. However, neither of these studies examine whether there are

differences between low- and upper-income owners or attempt to control for selection bias in who

becomes an owner.

Rohe and his colleagues again provide an important contribution on this topic. Their survey of

Baltimore homeowners and renters also investigated the degree of “neighboring” as measured by

summing responses to a series of questions asking how many of their neighbors respondents

recognized, knew by name, had social interactions with, or thought of as friends. Using multivariate

techniques to control for other differences between the two groups, they find that owners scored

lower on this measure of neighboring 18 months after purchase. However, the authors note that this

result might simply reflect the fact that their sample of owners had, be definition, recently moved,

which was not necessarily true of renters. In fact, when they re-survey the group 3 years after

purchase they find a positive association between neighboring and being an owner, although this

difference is not statistically significant.

Conclusions Regarding Impacts on Social Involvement

In general, there is not a rich literature examining the association between homeownership and social

involvement by low-income families. Studies by DiPasquale and Glaeser (1999), Rohe and Stegman

(1994a), and Rohe and Basolo (1997) stand out as making particularly important contributions on this

subject. On the whole, the findings from these studies indicate that there is at best only a slight

tendency for low-income owners to be more socially involved. With regard to voting and other

indicators of engagement in the political process, while homeowners generally are more likely than

renters to be engaged in political activities, this does not appear to be the case for low-income owners.

DiPasquale and Glaeser do find that low-income owners are more likely than renters to work to solve

local problems, but in general the association between homeownership and such efforts is fairly weak

as it is not statistically significant when controls for selection bias are employed. The studies by

Rohe and his colleagues do provide some indication that low-income owners are more likely to be

involved with neighborhood associations, but since this study concerns participants in a low-income

homeownership program that provided newly constructed homes in Baltimore, it is not clear how

generalizable these results are. Finally, there is a scarcity of studies investigating differences in

interactions with neighbors by tenure. Those that have been done provide little evidence that low-

income owners, in particular, are more involved with neighbors.

Impacts on Psychological and Physical Health

Another purported benefit of homeownership is a positive impact on both the psychological and the

physical health of owners. Psychologically, homeownership is thought to increase self-esteem, the

Chapter 5: Social Impacts of Homeownership

111

perception of control over one’s life (or self-efficacy), and overall life satisfaction. There are a

variety of mechanisms by which homeownership is believed to contribute to these outcomes. Higher

self-esteem can result from the greater social status associated with homeownership and from the

achievement of what is often an important personal goal of purchasing a home. Owners are thought

to have greater perceived control over their lives because they have greater control over their living

situation. Finally, the wealth created through homeownership both can contribute to a greater sense

of financial security and can help provide more of life’s comforts. However, on the other hand, it

may also be that the greater responsibilities of maintaining a home and meeting its financial

obligations can produce higher levels of stress for some households. In particular, households facing

a potential foreclosure may experience a significant degree of emotional strain leading to a loss of

self-esteem, a feeling of having no control over one’s life, and reduced life satisfaction.

Greater physical health may in part be attributable to greater psychological health based on the

premise that lower stress and a better outlook on life will have positive repercussions for physical

health. In addition, improved physical health could result from better quality of living conditions

associated with enhanced home maintenance by owner-occupants. Finally, increases in wealth

associated with homeownership may also improve physical health by supporting better access to

health care.

In general, the literature examining the impacts of homeownership on psychological and physical

health is not of high quality. With one notable exception, none of the existing studies employ any

attempts to control for what is likely to be a fairly significant self-selection of individuals with better

psychological and physical health to be more likely to choose to become a homeowner. Similarly,

with one exception, none of the existing studies attempt to evaluate whether the impacts of

homeownership on health differ between low- and high-income owners. In fact, many of the existing

studies use homeownership as a proxy for overall socioeconomic status in the absence of measures

for household income, wealth, or education, all of which would be expected to exert an important

independent effect on health. Finally, ideally studies would include controls for housing quality and

wealth to attempt to isolate whether an association between homeownership and positive

psychological and physical health might result from improvements in these areas. However, few

studies include controls for these factors.

The primary exception to these failures in the literature are the studies discussed above by Rohe and

Stegman (1994b) and Rohe and Basolo (1997). Their analysis of survey data for participants in a

Baltimore low-income homeownership program provide insights specifically on outcomes among

low-income households and provide some controls for selection bias by observing changes in the

households circumstances following a move to homeownership. However, the contributions of this

study are with regard to the psychological benefits of homeownership. There is no work of similar

quality that has assessed homeownership’s impacts on households’ physical health.

In the two sections that follow, we first describe the literature that has assessed impacts on

psychological outcomes and then turn to studies that have assessed impacts on physical health.

Chapter 5: Social Impacts of Homeownership

112

Impacts on Psychological Health

As noted above, the three main aspects of psychological health that are associated with

homeownership are increases in self-esteem, perceived control over life or self-efficacy, and overall

life satisfaction. Interestingly, while it seems fair to say that these impacts of homeownership are

widely taken for granted, there has been little quantitative work that has assessed the validity of these

claims. Qualitative research generally finds that participants in low-income homeownership

programs express very positive views of homeownership. One oft-cited study is Balfour and Smith

(1996) who conducted focus groups with participants in a lease-purchase program in Cleveland. The

authors summarize the feedback from the focus groups as indicating that the opportunity to work

toward homeownership “elevates their status in society and contributes to personal security and self-

esteem.” However, since the program entailed a lengthy period of time (15 years) as renters prior to

obtaining ownership, the focus-group participants also noted that during the lease period the lack of

control over maintenance to their homes and the need to follow rules established by property

managers was a source of irritation and frustration—which actually highlights the value owners attach

to having greater control over their homes. Herbert et al. (2003) also conducted focus groups with

participants in a low-income homeownership program in Michigan, who generally reported that

becoming an owner enhanced their self-esteem and the feeling of control over their living

environment.

Quantitative studies assessing homeownership’s psychological impacts are few. Attempting to fill

this void in the literature, Rossi and Weber (1996) use data from the National Survey of Families and

Households from 1988 to examine the association between homeownership and a range of indicators

of psychological health, including degree of self-satisfaction, feelings of competence, degree of

happiness and optimism, and feelings of depression. Their results confirm that homeowners do fare

better than renters in all of these dimensions, although the magnitudes of these differences are not

large. However, the conclusions of this study are fairly tentative as the study also only includes

controls for the respondents’ age and social status (based on an index that incorporates information on

income and education). The results do not shed any light on whether these outcomes vary with the

income of the owner. Nor do the authors attempt to account for any selection bias in who decides to

become an owner.

The most thorough research on this topic once again comes from Rohe and Stegman (1994b) and

Rohe and Basolo (1997). As described previously, these studies track participants in a low-income

homeownership program in Baltimore over time and compare changes in measures of psychological

health to changes observed over the same time in a comparison group of low-income renters.

Specifically, surveys were conducted prior to purchasing a home as well as 18 months and 3 years

after purchase. Survey questions were designed to evaluate the respondents’ self-esteem, perception

of control over their life, and overall life satisfaction. By examining changes in individual owners

over time, the study provides some control for potential selection bias in who chooses to buy a home.

The focus of these studies specifically on low-income households is also unique.

The authors measure self-esteem using an index based on responses to five questions asking the

respondent to give a self-assessment of their skills and ability. Their analysis of survey responses

both 18 months and 3 years after purchase does not find a statistically positive association between

homeownership and self-esteem. Interestingly, the only variable that is significant is the self-

Chapter 5: Social Impacts of Homeownership

113

assessment of housing condition, which is positively associated with self-esteem in both periods. A

comparison of simple mean differences in self-esteem reveals that owners had greater gains over time

in self-esteem than renters. Thus, the modeling results indicate that the one factor that explains this

difference is difference in housing conditions between owners and renters. This result suggests that

to the extent that homeownership affects self-esteem in may be as a result of increases in housing

quality. However, this result suggests that a move to higher quality housing—whether owned or

rented—could produce the same result.

These same studies assess the association between homeownership and the perception of having

control over one’s life based on the responses to nine questions related to respondents perceived

ability to have an impact on events in their life. They do not find a statistically significant association

between owning and the perception of control either 18 months or 3 years after purchasing.

However, they find few factors that are associated with perceived control over one’s life. But once

again survey results from 3 years after purchasing do find a positive association between housing

condition and perceived control over life, although this effect was not evident 18 months after

purchase. This suggests that homeownership’s impacts may again be related to helping to increase

housing quality, but this conclusion is weakened by the fact that housing quality does not have a

consistent relationship over time with perceived control over life.

Finally, these studies also ask respondents to rate their overall satisfaction with their life. The results

for this variable are the most consistent of the various psychological measures they test. Both 18

months and 3 years after purchase they find a strong positive and significant association between

owning and life satisfaction. Interestingly, housing condition is also positive and significant in both

surveys. Thus, the results of their analysis of the factors associated with life satisfaction suggest that

homeownership may have both a direct impact in increasing satisfaction and an indirect effect by

increasing housing quality.

Impacts on Physical Health

Compared to studies assessing impacts on psychological health, many more studies have examined

the relationship between homeownership and physical health. In part, the literature is broader on this

topic because in many cases studies of factors associated with physical health have used

homeownership as a proxy for a respondent’s income, wealth, and education in data sets where this

information is not available. But since these studies do not include controls for these factors, it is not

possible to tell whether homeownership has an independent effect on physical health or if the

association found is due to these other unaccounted for factors. These studies use a variety of

indicators of health, including indexes for overall health (Baker and Taylor, 1997; Kind et al., 1998;

Rossi and Weber, 1996; and Fogelman, Fox, and Power, 1989), frequency of visits to doctors (Baker

and Taylor, 1997; and Carr-Hill, Rice, and Roland, 1996), mortality (Sundquist and Johansson, 1997),

and degree of neurotic disorders (Lewis et al., 1998). These studies all find some significant

association between homeownership and positive health outcomes. But given the lack of controls

generally for economic status they shed little light on whether these outcomes are truly associated

with homeownership.

A few studies do incorporate controls for income and other key characteristics that are likely to affect

health outcomes, such as education. In an analysis of the American’s Changing Lives data set, Robert

Chapter 5: Social Impacts of Homeownership

114

and House (1996) include controls for respondents’ income and education and find that

homeownership is still associated with better functional health, although it is not associated with the

incidence of chronic conditions or self-rated health. In an analysis of a Scottish health survey,

Macintyre et al. (1998) add controls for income and find that homeownership is still associated with a

range of positive health outcomes, including overall responses to a general health questionnaire,

respiratory function, the number of long-standing illnesses, and the number of symptoms of poor

health. Interestingly, the authors also add a measure of self-esteem as a control variable but find that

this factor has only an imperceptible effect on the association between homeownership and health

outcomes. This result suggests that improvements in self-esteem may not be an important mechanism

by which homeownership affects physical health.

However, neither of these studies provide any indication of whether health outcomes might very with

the owner’s income level. One study with a particular focus on moderate-income households is Page-

Adams and Vosler (1997). These authors surveyed nearly 200 workers from two automobile plants

that were affected by the economic recession of the early 1990s, including one that was closed. The

survey sought to assess the incidence among these workers of signs of economic strain, depression,

and problematic alcohol use. The study also assessed whether workers had access to a social support

network as evidenced by having someone they could turn to for understanding and advice other than

their spouse. Incorporating controls for respondents’ education, income, and financial assets, they

study examined the association between homeownership and these outcomes. They find that

homeowners were less likely to suffer signs of economic strain, depression, or problematic alcohol

use, although they were not more likely to have access to a social support network. The results

suggest that homeownership may help ease the psychological stress of a financial crisis, in part by

helping to reduce the economic strain caused by lost or reduced employment. However, an important

caveat is that the study does not make any attempt to control for potential selection bias in which

workers chose to become homeowners.

While the findings of Page-Adams and Vosler suggest that homeownership can provide financial and

psychological support in times of economic crisis, in an analysis of survey data from Great Britain

from two time periods in the early 1990s, Nettleton and Burrows (1998) find that homeowners who

have difficulty making mortgage payments are more likely to experience indicators of poor health.

Specifically, these owners are found to lower scores on a general health questionnaire and to be more

likely to have visited a doctor. The results of this study highlight the potential negative effects of

homeownership on physical health should the owner experience difficulty in making mortgage

payments and face loss of their home. However, one weakness of the study by Nettleton and Burrows

is that they do not attempt to compare the experience of owners and renters who experience financial

hardship. While it is undoubtedly true that owners who face foreclosure will experience stress that is

sufficiently severe to affect physical health, it is not clear from these results whether these impacts

would be any worse for renters. The findings of Page-Adams and Vosler suggest that

homeownership may provide some protection against these stresses. In general, the results of these

two studies serve to highlight how little is known about differences in the impacts of financial stress

on owners and renters.

Chapter 5: Social Impacts of Homeownership

115

Conclusions Regarding Impacts on Psychological and Physical Health

In general, the literature assessing the impacts of homeownership on the psychological and physical

health of homeowners is too thin to draw any firm conclusions, particularly with regard to whether

these impacts may differ with household income. The studies by Rohe and Stegman (1994b) and

Rohe and Basolo (1997) represent the most compelling work on the issue of psychological impacts.

Their results suggest that homeownership may have a positive impact on overall life satisfaction of

low-income owners, but they find little support for the hypothesis that homeownership increases self-

esteem or perceived control over life. However, they do find that improved housing conditions are

associated with increase self-esteem and perceived control over life, which is consistent with the idea

that homeownership indirectly influences these outcomes by helping to improve housing quality. But

since efforts to improve the quality of rental housing might have the same result, if this is the

mechanism by which homeownership improves these outcomes, it may be possible to achieve the

same results by means other than promoting homeownership. With regard to physical health, there

are some indications that owners do enjoy better health, but since most studies do not employ

adequate controls for other aspects of a households’ socioeconomic status or for housing quality, it is

not possible to firmly conclude that this association exists. These studies generally do not shed light

on whether these impacts vary with homeowners’ income. In short, the question of whether

homeownership has an impact on physical health is very much an open question and one that requires

further research.

Summary and Conclusions

The most convincing work on the social impacts of homeownership has been concerned with impacts

on children. A series of studies have examined a variety of outcomes using different data sets and

different approaches to control for selection bias in who becomes an owner. Many of these studies

have also included controls to determine whether the association of positive outcomes with

homeownership may be due to reductions in residential mobility or differences in neighborhood

attributes. These studies have universally concluded that homeownership does have a positive and

statistically significant impact on children even after controlling for a variety of potentially

confounding factors. Importantly, the literature suggests that these positive impacts are actually

larger among low-income families and are large enough to outweigh any negative influences of living

in distressed communities. By far the most significant impacts are found with regard to educational

outcomes. The most common educational outcomes are that homeownership increases the number of

years of schooling completed, particularly that children are more likely to graduate from high school.

But one study also found positive impacts in terms of scores on reading and math achievement tests.

There is also some evidence that homeownership is associated with better labor market outcomes,

reduced behavioral problems, and increased homeownership for the children of owners, but these

results are not as strong as those for educational outcomes.

As far as the impact of homeownership on the owners themselves, there is less convincing evidence

that ownership is associated with increased social involvement among low-income owners. While

owners generally have been found to be more likely to vote and more likely to know the names of

elected officials, the few studies that have examined whether this impact differs by income have

found that this result is not evident among low-income owners. There is some indication that low-

Chapter 5: Social Impacts of Homeownership

116

income owners are more likely to work to solve local problems and to be involved in neighborhood

organizations, but given the small number of studies assessing these issues and the unique

circumstances of these studies, this conclusion is by no means definitive.

Research on homeownership’s impacts on owners’ psychological and physical health is particularly

weak. Only one series of studies of a low-income homeownership program in Baltimore provide any

convincing information on the psychological impacts of homeownership. These studies find that

while owners do exhibit higher life satisfaction, there is no evidence that they have higher self-esteem

or perceived control over their lives. However, these studies do find that improved housing

conditions are associated with higher self-esteem and greater perceived control over life, which may

be an indirect effect of homeownership. Unfortunately, there are no studies that provide any

convincing evidence about the impacts of homeownership on physical health.

Taken as a whole, the literature on social impacts suggests that there may be few significant social

benefits aside from impacts on children. However, the impacts on children alone may be important

enough to provide significant support for low-income homeownership. Nonetheless, further research

on the social impacts of homeownership is needed to come to a more definitive conclusion about

whether low-income households are likely to realize these benefits. Whether low-income owners are

more socially involved than renters is an important question to the extent that homeownership is

justified on the basis of benefits that accrue to the broader community. Given the lack of solid

research on the impacts of homeownership on a household’s psychological and physical health,

further well-designed research in this area would make a significant contribution to the literature.

Within this area, a particular interesting question is whether during times of economic crisis the

negative impacts on an owner’s psychological and physical health due to the stress of facing the loss

of a home through foreclosure outweigh any positive impacts from the accumulated wealth and

security of tenure that comes with homeownership. Perhaps more importantly, differences in

homeownership’s impacts by the race-ethnicity of the owner has essentially been completed ignored

across all spheres of this literature, creating a significant gap in our knowledge.

Chapter 5: Social Impacts of Homeownership

117

Chapter Six:

Summary of Findings, Policy Implications, and

Areas for Further Research