220 Week 3 F /For WIZARD KIM

Housing Market Dynamics: Disequilibrium, Mortgage Default, and Reverse Mortgages

Timothy Jones1 & Dean Gatzlaff2 & G. Stacy Sirmans3

Published online: 24 June 2016 # Springer Science+Business Media New York 2016

Abstract The mortgage market collapse which followed the burst of the housing bubble in 2007 led to unprecedented levels of mortgage default, mortgage foreclosure, and housing vacancies. The dramatic increase in mortgage terminations prompted extensive discussions among academics, industry practitioners, regulators, and policymakers. These discussions included such issues as the underlying causes of the mortgage market collapse and the impact of declining house prices on mortgage default and foreclosure. This special issue results from the 2013 FSU-UF Critical Issues Symposium entitled BHousing Market Issues: Initiatives, Policies and the Economy^ where the impact of the recent market collapse was examined on topics such as the negative spillover effect of foreclosure, the determinants of mortgage default and prepayment, disequilibrium models of the housing market, and the growth of the reverse mortgage market. This introductory paper serves to set the stage by relating the papers in this special issue to the previous literature. It provides both a clear picture of the state of knowledge in these critical areas and the significant contributions of these recent studies.

J Real Estate Finan Econ (2016) 53:269–281 DOI 10.1007/s11146-016-9567-y

* Dean Gatzlaff [email protected]

Timothy Jones [email protected]

G. Stacy Sirmans [email protected]

1 Department of Finance, Williams College of Business, Xavier University, Cincinnati, OH 45202, USA

2 Center for Real Estate Education & Research, College of Business, The Florida State University, Tallahassee, FL 32309-1110, USA

3 Department of Insurance, Real Estate, and Legal Studies, College of Business, The Florida State University, Tallahassee, FL 32309-1110, USA

Keywords Mortgage market . Mortgage default . Market disequilibrium . Spillover effects . Unemployment rates . Reverse mortgage

Introduction

After the collapse of the housing and mortgage markets in the 2000s decade, discussions on residential mortgage default, mortgage foreclosure, and housing vacancies increased dramatically in academic and industry-related circles. Just prior to the collapse, the U.S. homeownership rate peaked at nearly 70 %; however, from 2004 to 2009 the national default rate increased from approximately 1 % to 5.7 % and residential mortgage foreclosures increased from 1.2 % to 4.3 %.1 Additionally, the U.S. Census Bureau’s Current Population Survey/Housing Vacancy Survey reports that homeowner vacancy rates nearly doubled from nearly 1.5 % to 2.8 % over the same time period.

The dramatic increases in mortgage default, mortgage foreclosure, and housing vacancies are generally attributed to the rising popularity of aggressive mortgage lending (including subprime lending) and declining house prices. This resulted in millions of homeowners being underwater, i.e., owing more on their property than the property’s value. Discussions on issues such as the underlying causes of the housing market collapse and the impact of declining house prices on mortgage default and foreclosure led to the 2013 FSU-UF Critical Issues Symposium entitled BHousing Market Issues: Initiatives, Policies and the Economy .̂ Papers from that symposium resulted in this special issue of the Journal where the impact of the market collapse on topics such as the negative spillover effect of foreclosure, the determinants of mortgage default and prepayment, disequilibrium models of the housing market, and the growth of the reverse mortgage market is discussed.

The studies in this special issue seek to better identify the underlying factors of mortgage termination and explore the impacts of declining house prices on conven- tional and non-conventional mortgages. Specifically, the papers are devoted to analyz- ing the impact of the recent housing market collapse on negative spillover effects of foreclosure, the determinants of mortgage default and prepayment, disequilibrium models of housing markets, and reverse mortgage markets. We will set the stage in our introduction by relating the studies in this special issue to the previous literature.

Negative Spillover Effects of Foreclosure

As the steep rise in U.S. foreclosure rates prompted a thorough examination by both policymakers and academics, policymakers were especially interested in determining whether government intervention was appropriate and, if so, how to develop and implement programs that would effectively limit the occurrences of foreclosure. Academics, mean- while, sought to gain a better understanding of the determinants of foreclosure, the

1 Data on the U.S. mortgage default rate is collected from the S&P/Experian’s First Mortgage Default Index, and data on the U.S. foreclosure rates comes from the Mortgage Bankers Association (MBA) and Haver Analytics.

270 T. Jones et al.

magnitude of its impact on the selling price of foreclosed residential properties, and the spillover (or contagion) effects on nearby non-foreclosed properties.

In examining the spillover effect of foreclosure, the most common methodology has been the hedonic pricing model (see, for example, Immergluck and Smith (2006); Rogers and Winter (2009), and Daneshvary and Clauretie (2012)). In the hedonic model, home sale prices are regressed on property-specific characteristics (e.g., square footage, lot size, number of bedrooms and bathrooms, etc.) and neighborhood and demographic character- istics (e.g., population density, income, and race). The negative spillover effect was typically captured by adding the number of foreclosed properties within a certain distance as an explanatory variable. A second method, the repeat sales approach, has been used less often due to data restrictions (multiple sales dates per property and reduced sample size). However, the repeat sales approach benefits from reducing the omitted variable problem of hedonic models, controlling for time-invariant heterogeneity across properties, and separately iden- tifying the effects of price trends and the negative spillover effect of foreclosures (Harding et al. (2009); Gerardi et al. (2012)).

Table 1 summarizes the extant literature on the magnitude of the negative spillover effect related to foreclosures. Several empirical studies, beginning with Immergluck and Smith (2006), have estimated this spillover effect with results varying across different locations and time periods.

The studies in Table 1 consistently find evidence supporting the existence of a negative spillover effect related to foreclosure. This effect typically diminishes with distance and time. However, the underlying causes of the negative spillover effect are not as clearly defined. Three mechanisms have been proposed as explanations for the negative spillover effect: (1) the disamenity effect, (2) the comparable valuation effect, and (3) the competitive effect. The disamenity effect argues that foreclosed properties attract increased levels of crime and vandalism given the neglected maintenance. This, in turn, creates a negative externality for neighboring homes. Immergluck and Smith (2006); Leonard and Murdoch (2009), and Anenberg and Kung (2012) are examples of studies that partially attribute negative spillover to the disamenity effect.

The comparable valuation effect, on the other hand, argues that foreclosed properties lower the value of nearby, non-distressed properties since the former group sells at a significant discount which factors into the valuation of the latter group (see Lin et al. (2009)). The competitive effect also argues that foreclosed properties increase the supply of housing (relative to demand) at the neighborhood-level which reduces the selling price of nearby homes, other things equal. Anenberg and Kung (2012) find evidence that points to the competitive effect of foreclosure as the primary transmission mechanism.

Although these mechanisms are generally considered to be the underlying causes of the negative spillover effect, few studies have explicitly modeled them in theoretical or empirical frameworks. Anenberg and Kung (2012) empirically disentangle the disamenity and competitive effects for a sample of over 240,000 single-family property listings in Chicago, Phoenix, San Francisco, and Washington DC between January 2007 and January 2009. They argue that any spillover effect that emerges while the foreclosed property is for sale represents a competitive effect because Bthis is the time period when the foreclosure is competing with neighboring listings for buyers.^ The disamenity effect, by comparison, should emerge prior to the listing date when the distressed borrower has less incentive to maintain the property. Their evidence supports the existence of both competitive effects and disamenity effects but the disamenity effect is only significant for high density, low value

Housing Market Dynamics: Disequilibrium, Mortgage Default 271

T ab

le 1

S ur ve y of

E xt an t L it er at ur e on

th e N eg at iv e S pi ll ov er

E ff ec ts of

F or ec lo su re

A ut ho r( s) an d Y ea r

L oc at io n

Y ea r

S am

pl e

M od el

S pi ll ov er

D is ta nc e

Im m er g lu ck

an d S m it h (2 00 6)

C h ic ag o , IL

1 99 9

9 60 0 h o m e sa le s

H ed on ic

−0 .9 0 %

1/ 8t h m il e

−0 .3 0 %

1/ 8t h an d 1 /4 th

m il e

M ik el b an k (2 0 08 )

C o lu m bu s, O H

1 99 0 – 2 00 0

9 04 6 h o m e sa le s

H ed on ic

−3 .1 0 %

0 to

2 50

ft

−1 .8 0 %

25 0 to

5 00

ft

−2 .1 0 %

0 to

2 50

ft

−1 .1 0 %

25 0 to

5 00

ft

S ch ue tz et al . (2 0 08 )

N ew

Y o rk

C it y, N Y

2 00 0 – 2 00 5

5 7, 0 00

ho m e sa le s

H ed on ic

−1 .8 0 %

1/ 20 th

m il e

C am

pb el l et al . (2 0 09 )

M as sa ch us et ts

1 98 7– 2 00 8

1 .8

m il li o n h om

e sa le s

H ed on ic

−1 %

1/ 20 th

m il e

H ar di ng

et al . (2 00 9)

37 M S A s

1 98 9 – 2 00 7

4 00 ,0 0 0 re p ea t- sa le s pa ir s

R ep ea t sa le s

−1 .3 0 %

1/ 20 th

m il e

−0 .6 0 %

1/ 20 th

an d 1/ 10 th

m il e

L eo na rd

an d M u rd oc h (2 0 09 )

D al la s C o un ty , T X

2 00 6

2 3, 0 00

ho m es

sa le s

H ed on ic

−0 .5 0 %

0 to

2 50

ft

−0 .1 0 %

> 1 00 0 ft

L in

et al . (2 0 09 )

C hi ca go , IL

(P M S A )

1 99 0– 2 00 6

1 4, 4 27

ho m e sa le s

H ed on ic

−9 .7 0 %

< 3 00

ft

−5 .3 0 %

0. 1 km

- 0 .2

k m

−1 .2 0 %

0. 8 km

- 0 .9

k m

R o g er s an d W in te r (2 00 9 )

S t. L o ui s C o un ty , M O

2 00 0 – 2 00 7

9 8, 8 28

ho m e sa le s

H ed on ic

−1 .1 0 %

20 0 y ar ds

−0 .3 7 %

60 0 y ar ds

−0 .7 7 %

60 0 y ar ds

W as sm

er (2 0 11 )

S ac ra m en to , C A

Ja n . 2 0 08

- Ju n e 2 00 9

3 6, 0 00

ho m e sa le s

H ed on ic

−0 .6 1 %

1/ 10 th

m il e

−0 .1 0 %

1/ 10 th

an d 1/ 4t h m il e

D an es hv ar y an d C la ur et ie (2 0 12 )

L as

V eg as , N V

Ja n . 2 0 08

- Ju n e 2 00 9

3 6, 0 93

ho m e sa le s

H ed on ic

−1 .4 0 %

1/ 10 th

m il e

−1 %

1/ 10 th

to 1 /4 th

m il e

G er ar di

et al . (2 0 12 )

15 M S A s

2 00 1 – 2 01 0

9 58 ,5 1 3 re p ea t- sa le s pa ir s

R ep ea t sa le s

−1 .2 0 %

1/ 10 th

m il e

−0 .8 0 %

1/ 10 th

to 1 /4 th

m il e

A n en be rg

an d K u ng

(2 0 12 )

4 M S A s

Ja n . 2 0 07

- Ja n . 20 09

2 40 ,0 0 0 h o m e sa le s

H ed on ic

−1 .6 0 %

1/ 10 th

m il e

272 T. Jones et al.

census tracts. In particular, the marginal impact of the competitive effect is roughly 1.6 % for one additional REO within one-tenth of a mile of the subject property, while the disamenity effect is roughly 1.5 %.

The two papers in this special issue by Ihlanfeldt and Mayock) and Li make valuable contributions to the literature on the negative spillover effect of foreclosures. Specifi- cally, both studies examine the transmission mechanism through which foreclosed properties reduce nearby property values, emphasizing the disamenity effect. Ihlanfeldt and Mayock argue that the increased number of investor-acquired REOs has shifted the composition of housing stock from owner-occupied properties to rental properties. Assuming that rental properties are less maintained than owner-occupied properties, an increased number of investor-acquired REOs reduces neighborhood amenity thereby lowering neighborhood property values. The status of the property (investor-acquired versus owner-occupied) is identified using the post-sale homestead status reported in county-level property tax rolls. Their sample consists of residential homes sold between 1996 and 2011 in a set of Florida counties.

The authors find a statistically and economically significant negative spillover effect for REO properties and this effect is generally larger for investor-acquired REOs relative to owner-occupied REOs. The magnitude of the negative spillover effect varies considerably across each county, but the effect is pervasive across all ten counties in the sample. This finding supports the hypothesis that the disamenity effect is the underlying cause of the negative spillover effect. The hypothesis is further supported by the diminishing effect of the negative externality as the distance from the REO property increases. We note that no direct evidence on the relative maintenance expenditures of investor-acquired REOs and owner-occupied REOs is reported.

Filling a gap in the existing literature, Li provides direct evidence of the transmission mechanism using a unique dataset of capital expenditures and physical quality mea- sures for Milwaukee properties sold between 2005 and 2009. Information on capital expenditures and physical quality of the subject property was obtained from the Milwaukee tax assessor’s office. Li hypothesizes that individual homeowners reduce capital expenditures when home prices fall and the probability of foreclosure increases. This results in lower neighborhood amenity because reductions in capital expenditures on foreclosed properties incentivizes nearby homeowners to spend less money on their respective home improvements. This further reduces home prices and exacerbates the under-investment problem. Leonard and Murdoch (2009) had previously argued that reductions in neighborhood quality are the primary transmission mechanism of foreclosed properties. Since they employ only property-level transaction data and neighborhood characteristics rather than tax assessor’s data they did not explicitly test for changes in neighborhood quality.

Li also uses a more robust empirical methodology that considers foreclosure out- comes, capital expenditures, time on market (TOM), and sale prices as endogenous variables. Prior studies have acknowledged that foreclosures and selling prices may suffer from an endogeneity problem. For example, Immergluck and Smith (2006) argue that the negative relationship between foreclosures and property values could be capturing the effect of property values on the probability of nearby foreclosures. However, they do not address this problem because they are unable to identify appropriate instrumental variables. Li does directly address the endogeneity problem by estimating a three-stage least square model which controls for the likelihood of

Housing Market Dynamics: Disequilibrium, Mortgage Default 273

foreclosure in the first stage, estimates capital expenditures and TOM in the second stage, and estimates the determinants of the sale price in the third stage.

Consistent with prior studies, Li provides evidence of a negative spillover effect related to foreclosure. There is a 7.3 % price discount for homes within 200 m of a property that foreclosed in the prior month and this effect diminishes to 0.7 % within five years after the foreclosure. The magnitude of the negative spillover effect is larger than the results reported in the existing literature but is similar to the results reported by Lin et al. (2009). Li further shows that, controlling for capital expenditures and TOM, the own-price discount related to foreclosure is only 10 %. This is significantly lower than the 20 % commonly reported in previous studies. However, this is consistent with Clauretie and Herzog (1990) who control for the physical quality of the property and the relationship between TOM and price.

Default and the Unemployment Rate

Mortgage default studies have largely focused on the underlying determinants of default and have highlighted the effect of loan characteristics, borrower attributes, local housing market and macroeconomic conditions, and trigger events (e.g. unemployment and divorce).2 The unemployment rate seems especially important as a potential determinant of mortgage default given its steep rise during the recent financial crisis. According to the Bureau of Labor Statistics, the national unemployment rate rose from less than 6 % to roughly 10 % between 2004 and 2009, respectively. The default rate similarly increased from 1 % to nearly 6 % over the same time period, based on estimates from S&P/Experian’s First Mortgage Default Index. At first glance, there appears to be limited support for a positive relationship between default and the unemployment rate.

Several studies have examined the relationship between default and the unemploy- ment rate in a multivariate setting. Theoretical predictions, such as the ability-to-pay theory, argue that default and the unemployment rate are positively related. Homeowners have greater difficulty meeting their contractual debt obligations in periods of high unemployment, which results in an increased likelihood of mortgage default and foreclosure. The evidence on the influence of unemployment rates is mixed, however. Clauretie and Herzog (1990); Hendershott and Schultz (1993); Springer and Waller (1993); Capozza et al. (1997); Deng (1997); Ambrose and Capone (2000); Deng et al. (2000), and Noordewier et al. (2001) find a positive relationship between default and the unemployment rate, consistent with the theoretical predictions.3 Conversely, Cunningham and Capone (1990) find that regional unemployment rates are negatively related to the default rate for a sample of fixed-rate mortgages (FRMs) and adjustable- rate mortgages (ARMs) originated in Texas between 1962 and 1985. They are unable to explain the observed negative relation; however, they argue that Beven if the unemploy- ment rate increases, resulting in lost jobs and increased payment-to-income ratios, borrowers should not default if positive equity can be recovered by selling the property.^

2 Quercia and Stegman (1992) provide a comprehensive survey of the residential mortgage default literature from the early 1960s to the early 1990s. 3 The unemployment rate is generally measured at the state or regional level, and the results appear to be robust to the choice of measurement.

274 T. Jones et al.

Therefore, the unemployment rate should play a secondary or tertiary role to loan characteristics and borrower attributes in default. Deng et al. (1996) and LaCour-Little (2004) find that the unemployment rate is not significantly related to default. LaCour- Little (2004), in particular, suggests that the presence of junior liens, not unemployment, is largely responsible for the increased default rates of the mid-1990s.

More recent studies using data that cover the mortgage market collapse of the 2000s also produce mixed results. Elul et al. (2010) and Quercia et al. (2012) find a positive relationship between default and the unemployment rate that is more pronounced for high-LTV loans. Ghent and Kudlyak (2011) also find a positive but not statistically significant relationship for their full sample of loans. Looking only at those loans originated between 2005 and 2008, the sign of the coefficient becomes negative although it remains not statistically significant. Archer and Smith (2013) find conflict- ing evidence of both positive and negative relationships between default and unemployment depending on the model specification. In short, the relationship between default and the unemployment rate differs significantly by location and time period. In a recent paper, Tian et al. (2015) find both household unemployment and the local unemployment rate to be important predictors of mortgage default. They also find the presence of precautionary savings and the duration of unemployment benefits to be significant moderators of mortgage default.

In this special issue, Quercia et al. make a significant contribution to the literature on unemployment rates and mortgage default by examining different types of unemployment. They argue that the differential impacts of unemployment rates on mortgage default and prepayment may be due to differences in cyclical (or short-term) and structural (or long- term) unemployment. Cyclical unemployment arises from temporary labor market condi- tions while structural unemployment results from mismatches between labor supply and demand.

The authors separately identify cyclical and structural unemployment rates using the Beveridge-Nelson decomposition and the Hodrick-Prescott filter decomposition alternatives.

Their sample consists of a national sample of community reinvestment loans originated by low- or moderate-income families in low-income or minority areas. Controlling for loan and borrower characteristics, they find that all local unemployment rates except cyclical unemployment are positively related with default and negatively related with prepayment. These findings support theoretical predictions and provide evidence that long-term unem- ployment rates are more important determinants of mortgage termination than short-term unemployment rates. Interaction terms further suggest that both cyclical and structural unemployment are highly sensitive to borrower FICO scores at origination.

Disequilibrium Models of the Housing Market

Hedonic pricing models and general equilibrium models generally assume that the housing market is in equilibrium. Specifically, house prices are assumed to adjust nearly instantaneously so that the supply of housing is perfectly offset by the demand for housing at any point in time. The validity of this assumption, however, has been questioned following the realization that housing markets adjust slowly to changing market conditions (Case and Shiller (1989) and changing demographics (Mankiw and Weil (1989)). The dramatic rise in mortgage default, mortgage foreclosure and housing

Housing Market Dynamics: Disequilibrium, Mortgage Default 275

vacancies has shed further doubt on the market clearing mechanism in the housing market given excessive and persistent vacancy rates.

DiPasquale and Wheaton (1994) show that the housing market is subject to extended periods of disequilibrium and they develop a theoretical model explaining this disequi- librium. Their partial adjustment housing market model simultaneously considers the stock (or long-run supply) of housing and the flow (or net investment) of housing. The stock-flow component of the model allows for gradual adjustment in prices via gradual changes in the supply of housing and it assumes that a rise in prices results in a temporary rise in the flow of new housing. Market disequilibrium may result from shifts in supply conditions, demand conditions, or both.4

DiPasquale and Wheaton estimate house prices using two-stage least squares re- gression for a national sample of U.S. homes. Controlling for supply- and demand-side disturbance, they find that lagged prices indicate a gradual response to exogenous shocks, while the lagged stock of housing variable indicates a gradual adjustment to the flow of new housing. Both results provide evidence supporting disequilibrium in the housing market. Changing demographics, however, do not appear to significantly impact the demand for housing.

Riddel (2004) later develops a similar housing market disequilibrium model that separately identifies supply-side and demand-side disturbances. He extends the partial- adjustment model of DiPasquale and Wheaton by employing a multiple error- correction approach. This accounts for both the magnitude and underlying source of the supply-side adjustment. Estimating the model for a nationwide sample of single- family homes between 1967 and 1998, he also finds a gradual adjustment to changes in the housing stock. The price adjustment, on the other hand, responds relatively quickly to the changes in the supply of housing. These results further support the existence of disequilibrium in the housing market, driven primarily by the housing stock being slow to adjust to supply-side disturbances.

Hwang and Quigley (2006) also model the relationship between housing demand, housing supply, and vacancies using a nationwide sample of single-family homes across 74 metropolitan statistical areas (MSAs) between 1987 and 1999.5 Their study is the first to explicitly model vacancies in the local housing market, which represents differences in the market clearing demand and supply of homes. Prior studies do not account for changes in vacancy rates that may impact house prices and construction of new properties. Further, the structural models are tested on local housing markets, rather than using aggregate, national data.

Their three-equation model shows that housing prices (or the inverse-demand) are negatively related to contemporaneous measures of the housing stock and the vacancy rate. MSA-level macroeconomic conditions, household income, and employment all have a positive relationship with house prices. Supply is negatively related to labor and material costs, financing costs, and stringent regulations. Changes in house prices have a positive effect on the housing stock while changes in vacancy rates have a negative

4 Changes in construction wages, building material costs, short-term lending rates, or the restrictiveness of housing regulations can generate supply-side disturbances. Fluctuating macroeconomic conditions, changes in borrower income or wealth, changes in market demographics, and changes in the price of rental properties can generate demand-side disturbances. 5 The supply of housing (or supplier activity) is approximated by the number of building permits in a specific MSA, rather than using the stock of housing.

276 T. Jones et al.

effect on the housing stock, although both relationships are only marginally significant. Finally, vacancy rates are positively related to the supply of housing and negatively related to house prices, but the macroeconomic variables are not statistically significant. These results provide evidence on the underlying causes of supply-side disturbances that are not well understood.

Grimes and Aitken (2010) examine land costs as a factor in housing market disequilibrium. They analyze the relationship between housing supply elasticities, land costs, and house price dynamics. Previous studies such as Glaeser and Gyourko (2006) and Glaeser et al. (2008) have examined house price dynamics but neither study explicitly incorporates in a measure of land costs. Grimes and Aitken provide a couple of important conclusions. First, they confirm the rela- tionship between housing supply and house price dynamics (with increasing demand, increased housing supply relative to population reduces the long-run price impact). Second, they find that housing supply (house prices) will be more (less) responsive to demand shocks where the supply of land is elastic. By using panel data for 73 regions of New Zealand and a model that explicitly includes land costs along with construction costs, they find that land costs constitute approxi- mately 30 % of total development costs. Their finding that regional supply elasticities are inversely related to dynamic regional price adjustments validates the importance of including land costs when modeling housing supply.

In this special issue, Zabel advances the literature on housing market disequilibrium models. Extending the analysis of Hwang and Quigley (2006), he develops a disequilib- rium model where lagged vacancies are governed by an error correction model. Specif- ically, the two error correction mechanisms, which capture supply and demand shifts, measure the difference between observed vacancy rates and Bnatural^ vacancy rates.6

Zabel’s model differs from the Hwang and Quigley’s model with the use of MSA and time fixed effects rather than random effects. Foreclosures and their associated impacts on vacancy rates are also explicitly modeled. Using a sample of single-family houses at the MSA-level between 1991 and 2011, his findings on the determinants of foreclosure are largely consistent with extant literature. Changes in house prices have a negative impact on the likelihood of foreclosure. This confirms the importance of negative equity in the default-foreclosure decision. The unemployment rate is positively related to the likelihood of foreclosure, consistent with theoretical predictions. According to model estimates, nearly 10 % of properties that are seriously delinquent (90+ days) will become vacant within one year while a 10 % increase in employment reduces vacancies by 6 %. In the house price equation, the significance of lagged house prices provide further evidence that the housing market gradually adjusts to changing market conditions. House prices are inversely related to house prices in periods of excess demand; however, prices do not respond to increasing vacancy rates in periods of excess supply. Finally, lagged foreclosure rates are positively related to house prices in periods of excess demand or excess supply.

6 The Bnatural^ vacancy rate implies that there is no excess supply or demand for housing, which represents Bthe outcome of the residential search process that involves households with stochastic preferences matching with a heterogeneous housing stock^ (Arnott (1989) and Wheaton (1990)).

Housing Market Dynamics: Disequilibrium, Mortgage Default 277

The Growth of the Reverse Mortgage Market

Fluctuations in housing demand and supply and the associated impact on house prices have affected the demand for all types of mortgages, including reverse mortgages. A reverse mortgage is used by senior homeowners to convert home equity into income. Repayment is not required unless one of the following conditions occurs: (1) the borrower dies, (2) the home is sold, (3) the borrower permanently moves out of the house, or (4) some other component of the contract is violated (such as nonpayment of property taxes or insurance). In the U.S., the Home Equity Conversion Mortgage (HECM) is the most popular type of reverse mortgage and accounts for more than 90 % of the reverse mortgage originations.7

Borrowers of HECMs must be 62 years of age or older, must occupy the property as their primary residence, and must either own the home or have small mortgage balance. Unlike conventional mortgages, reverse mortgages are generally not subject to stringent under- writing standards such as minimum income levels or creditworthiness. HECMs are particularly attractive to Bhouse-rich, cash-poor^ elderly homeowners.

The popularity of reverse mortgages increased substantially in the mid-2000s, just prior to the mortgage market collapse. A growing body of literature has analyzed the underlying cause of growth in the reverse mortgage market. In particular, empirical studies have examined the extent to which loan and property characteristics, borrower attributes, and demographic characteristics explain the increased take-up rate of HECMs. Case and Schnare (1994) examine the preliminary results of the HECM program following its implementation in 1987 and find, relative to the general elderly population, that HECM borrowers tend to be house-rich but cash-poor. They loosely define these borrowers as Belderly homeowners whose incomes are below $15,000 and whose homes are valued at more than $50,000.^ According to the 1989 American Housing Survey (AHS), this definition applies to more than 3.5 million elderly homeowners. They further show that HECMs are more popular among older homeowners, non-minorities, and single females.

Two studies, Rodda et al. (2000) and Redfoot et al. (2007), provide additional evidence on the popularity of HECMs. Rodda et al. (2000) find that only 1 % of eligible homeowners have reverse mortgages and nearly one-third of the outstanding reverse mortgages were originated in a single year, 1997. Similar to Case and Schnare (1994), they find that the take-up rate for reverse mortgages is higher for older homeowners, female-headed households, and lower income households. They further observe that relatively better-educated homeowners tend to exhibit higher HECM take-up rates. Redfoot et al. (2007) also find that HECM take-up rates are higher for older and better educated homeowners, female-headed households, and lower income households.

Shan (2011) examines data on reverse mortgage loans for more than 380,000 HECMs originated from 1989 to 2007. The analysis focuses on explaining reverse mortgage take- up rates across ZIP codes, differences between recent and earlier borrowers, and the

7 Created by Congress in 1987, the HECM’s primary objectives are (1) to permit the conversion of home equity into liquid assets to meet the special needs of elderly homeowners, (2) to encourage and increase participation by the mortgage markets in converting home equity into liquid assets, and (3) to determine the extent of demand for home equity conversion and the types of home equity conversion mortgages that best serve the needs of elderly homeowners. The program is administered by the U.S. Department of Housing and Urban Development (HUD).

278 T. Jones et al.

growth in the reverse mortgage in the mid-2000s. Results are largely consistent with prior studies. In particular, reverse mortgages are higher in zip codes with better educated residents and lower income households with higher house prices, consistent with the notion of HECMs being more attractive to cash-poor, house-rich elderly homeowners.

Unlike Case and Schnare (1994), Shan finds that zip codes with higher fractions of blacks and Hispanics are associated with higher take-up rates, suggesting that demographic characteristics of borrowers have changed considerably since the introduction of the HECM program in the late 1980s and early 1990s. Comparing recent and earlier reverse mortgage borrowers, recent borrowers tend to be younger, single males, or couples who can access a larger fraction of their housing equity. Finally, Shan argues that nearly one- third of the expansion of the reverse mortgage market can be attributed to house price appreciation, which allows elderly homeowners to borrow more against their homes.8

Haurin et al. in this special issue, make a significant contribution to the literature on HECM take-up rates by examining changes across states and over time. Their work is closely related to that of Shan (2011) in that house prices are assumed to play a primary role in explaining inter-temporal and spatial variations in reverse mortgage take-up rates. Their major hypothesis is that variations in HECM take-up rates depend on current house prices relative to historical average prices and the level of house price volatility. Specifically, elderly households in states with volatile house prices and where house prices are high compared to historical averages are more likely to take out reverse mortgages. Conversely, HECM take-up rates decrease if house prices are below the mean in a state with a history of volatile prices. Further results show that HECM take-up rates are higher for older borrowers (aged 74 to 77), Hispanics, and Asians. Education has no significant effect.

Summary and Conclusions

This special issue presents theoretical and empirical studies on the impact of the recent mortgage market collapse. These studies make significant contribution in several different areas including (a) the determinants of mortgage default, prepayment, and foreclosure, (b) the magnitude and causes of the negative spillover effects of foreclo- sure, (c) the differential impact of short- and long-term unemployment rates on the likelihood of default, (d) disequilibrium models of the housing market, and (e) the growth of the reverse mortgage market. Results show that declining house prices (which produced reduced or negative home equity) and long-term unemployment rates at least partially explain the increased occurrence of mortgage default and foreclosure and housing vacancies, while the negative spillover effect of foreclosure can be primarily attributed to the disamenity effect. Furthermore, given the perceived imbal- ance between housing demand and supply, disequilibrium models are becoming more prevalent in the literature. These models explain most of the variation in demand-side disturbances; however, supply-side disturbances and changes in vacancy rates are less

8 Shan (2011) also predicts that lower interest rates increase the attractiveness of reverse mortgages since it allows borrowers to cash out a larger fraction of their home equity at the time of loan origination. However, it is impossible to separately identify the effect of lower interest rates on the demand for reverse mortgages since yields on Treasury bills do not vary across locations.

Housing Market Dynamics: Disequilibrium, Mortgage Default 279

clearly defined. Finally, the growth in the reverse mortgage market can partially be explained by house price appreciation prior to mortgage market collapse. Given the severity and wide-reaching effects of the mortgage market collapse, there are promising opportunities for future research related to the above topics and many others.

References

Ambrose, B. W., & Capone, C. A. (2000). The hazard rates of first and second defaults. Journal of Real Estate Finance and Economics, 20(3), 275–293.

Anenberg, E., and Kung, E. (2012) Estimates of the size and source of price declines due to nearby foreclosures, Working paper, Duke University Department of Economics.

Archer, W. R., & Smith, B. C. (2013). Residential mortgage default: the roles of house price volatility, euphoria and the borrower’s put option. Journal of Real Estate Economics and Finance, 46(2), 355–378.

Arnott, R. (1989). Housing vacancies, thin markets, and idiosyncratic tastes. Journal of Real Estate Finance and Economics, 2(1), 5–30.

Campbell, J. Y., Giglio, S. and Pathak, P. (2009) Forced sales and house prices, No. w14866, National Bureau of Economic Research.

Capozza, D. R., Kazarian, D., & Thomson, T. A. (1997). Mortgage default in local markets. Real Estate Economics, 25(4), 631–655.

Case, B., & Schnare, A. B. (1994). Preliminary evaluation of the HECM reverse mortgage program. Real Estate Economics, 22(2), 301–346.

Case, K. E., & Shiller, R. J. (1989). The efficiency of the market for single family homes. American Economic Review, 79, 125–137.

Clauretie, T. M., & Herzog, T. (1990). The effect of state foreclosure laws on loan losses: evidence from the mortgage insurance industry, Journal of Money. Credit and Banking, 22(2), 221–233.

Cunningham, D. F., & Capone, C. A. (1990). The relative termination experience of adjustable to fixed-rate mortgages. Journal of Finance, 45(5), 1687–1703.

Daneshvary, N., & Clauretie, T. M. (2012). Toxic neighbors: foreclosures and short-sales spillover effects from the current housing-market crash. Economic Inquiry, 50(1), 217–231.

Deng, Y. (1997). Mortgage termination: an empirical hazard model with a stochastic term structure. Journal of Real Estate Finance and Economics, 14(3), 309–331.

Deng, Y., Quigley, J. M., & Van Order, R. (1996). Mortgage default and low downpayment loans: the costs of public subsidy. Regional Science & Urban Economics, 26, 263–285.

Deng, Y., Quigley, J. M., & Van Order, R. (2000). Mortgage terminations, heterogeneity and the exercise of mortgage options. Econometrica, 68(2), 275–307.

DiPasquale, D., & Wheaton, W. C. (1994). Housing market dynamics and the future of housing prices. Journal of Housing Economics, 35(1), 1–27.

Elul, R., Souleles, N. S., Chomsisengphet, S., Glennon, D., & Hunt, R. M. (2010). What Btriggers^ mortgage default? American Economic Review, 100(2), 490–494.

Gerardi, K., Rosenblatt, E., Willen, P. S. and Yao, V. (2012) Foreclosure externalities: Some new evidence, No. w18353, National Bureau of Economic Research.

Ghent, A. C., & Kudlyak, M. (2011). Recourse and residential mortgage default: evidence from US states. Review of Financial Studies, 24(9), 3139–3186.

Glaeser, E., & Gyourko, J. (2006). Housing dynamics, NBER working paper 12787. Cambridge, MA: National Bureau of Economic Research.

Glaeser, E., Gyourko, J., & Saiz, A. (2008). Housing supply and housing bubbles. Journal of Urban Economics, 64(2), 198–217.

Grimes, A., & Aitken, A. (2010). Housing supply, land costs and price adjustment. Real Estate Economics, 38(2), 325–353.

Harding, J. P., Rosenblatt, E., & Yao, V. W. (2009). The contagion effect of foreclosed properties. Journal of Urban Economics, 66(3), 164–178.

Hendershott, P. H., & Schultz, W. R. (1993). Equity and nonequity determinants of FHA single-family mortgage foreclosures in the 1980s. Real Estate Economics, 21(4), 405–430.

Hwang, M., & Quigley, J. M. (2006). Economic fundamentals in local housing markets: Evidence from US metropolitan regions. Journal of Regional Science, 46(3), 425–453.

280 T. Jones et al.

Immergluck, D., & Smith, G. (2006). The external costs of foreclosure: the impact of single-family mortgage foreclosures on property values. Housing Policy Debate, 17(1), 57–79.

LaCour-Little, M. (2004). Equity dilution: an alternative perspective on mortgage default. Real Estate Economics, 32(3), 359–384.

Leonard, T., & Murdoch, J. C. (2009). The neighborhood effects of foreclosure. Journal of Geographical Systems, 11(4), 317–332.

Lin, Z., Rosenblatt, E., & Yao, V. W. (2009). Spillover effects of foreclosures on neighborhood property values. Journal of Real Estate Finance and Economics, 38(4), 387–407.

Mankiw, N. G., & Weil, D. N. (1989). The baby boom, the baby bust, and the housing market. Regional Science and Urban Economics, 19(2), 235–258.

Mikelbank, B. A. D2008] Spatial analysis of the impact of vacant, abandoned and foreclosed properties, Federal Reserve Bank of Cleveland.

Noordewier, T. G., Harrison, D. M., & Ramagopal, K. (2001). Semivariance of property value estimates as a determinant of default risk. Real Estate Economics, 29(1), 127–159.

Quercia, R. G., & Stegman, M. A. (1992). Residential mortgage default: a review of the literature. Journal of Housing Research, 3(2), 341–379.

Quercia, R. G., Pennington-Cross, A., & Tian, C. Y. (2012). Mortgage default and prepayment risks among moderate- and low-income households. Real Estate Economics, 40(1), 159–198.

Redfoot, D. L., Scholen, K., & Brown, S. K. (2007). Reverse mortgages: niche product or mainstream solution?, AARP Public Policy Institute. DC: Washington.

Riddel, M. (2004). Housing-market disequilibrium: an examination of housing market price and stock dynamics 1967—1998. Journal of Housing Economics, 13(2), 120–135.

Rodda, D. T., Herbert, C., & Lam, H. K. (2000). Evaluation report of FHA’s home equity conversion mortgage insurance demonstration. Cambridge, MA: Apt Associates.

Rogers, W. H., & Winter, W. (2009). The impact of foreclosures on neighboring housing sales. Journal of Real Estate Research, 31(4), 455–479.

Schuetz, J., Been, V., & Ellen, I. G. (2008). Neighborhood effects of concentrated mortgage foreclosures. Journal of Housing Economics, 17(4), 306–319.

Shan, H. (2011). Reversing the trend: the recent expansion of the reverse mortgage market. Real Estate Economics, 39(4), 743–768.

Springer, T. M., & Waller, N. G. (1993). Termination of distressed residential mortgages: an empirical analysis. Journal of Real Estate Finance and Economics, 7(1), 43–54.

Tian, C. Y., Quercia, R. G., & Riley, S. (2015). Unemployment as an adverse trigger event for mortgage default. Journal of Real Estate Finance and Economics, 51(1), 1–22.

Wassmer, R. W. (2011). The recent pervasive external effects of residential home foreclosure. Housing Policy Debate, 21(2), 247–265.

Wheaton, W. C. (1990). Vacancy, search, and prices in a housing market matching model. Journal of Political Economy, 98(6), 1270–1292.

Housing Market Dynamics: Disequilibrium, Mortgage Default 281

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.