Mortgage default risk, foreclosures, and local house price ...

Mortgage default risk, foreclosures, and local house price dynamics

Damian S. Damianov,* Cheng Yan and Xiangdong Wang

11 September 2018

Abstract

We examine the dynamic interdependence between (two measures of) mortgage default risk and (top and bottom tier Zillow) house prices in 92 MSAs located in 25 U.S. states. The mortgage default risk index, which reflects household Google search behavior for foreclosure help (Chauvet, Gabriel, and Lutz, 2016) has a stronger negative impact on the prices of high tier homes, while the percentage of homes foreclosed has a stronger negative impact on the prices of low tier homes. Shocks to high tier home values have a stronger negative impact on default risk in non-recourse states than in recourse states, while shocks to low tier home values have about the same impact on default risk in recourse and non-recourse states. Our findings suggest that owners of high-value homes are more financially sophisticated and strategic: they are less likely to default in states where lenders can pursue a deficiency judgment against them.

[Word count: 149]

Keywords: Foreclosure; mortgage default risk; house prices. JEL classification: D12, D14, E51, G21, G33, L85, R31.

*Corresponding author. Damian S. Damianov, damian.damianov@durham.ac.uk, Durham University Business School, Millhill Lane, Durham, UK, DH1 3LB. Cheng Yan, cheng.yan@dur.ac.uk, Durham University Business School, Millhill Lane, Durham, UK, DH1 3LB. Xiangdong Wang, xiangdong.wang@dur.ac.uk, Durham University Business School, Millhill Lane, Durham, UK, DH1 3LB. We thank Alla Koblyakova, Lok Man Michelle Tong, Gianluca Marcato, and participants at the 2018 AREUEA International Conference, the 25th Annual Conference of European Real Estate Society, the 5th Money, Macroeconomics and Finance (MMF) PhD conference, and the MMF and Cass Business School conference on Investment, Funding and Industrial Policy for their comments.

1

1. Introduction One of the main lessons of the subprime mortgage crisis is that the financial concerns of homeowners are of first-order importance for the stability of the U.S. financial system and economy (Mian and Sufi, 2014). In the post-crisis period, a voluminous literature has developed that aims to shed light on a key relationship in the run-up to the financial crisis: the interdependence between rising mortgage default rates and downward spiraling house prices. A better grasp of this issue was a matter of urgency during the housing market downturn as policymakers evaluated initiatives to reduce foreclosure activity and help `underwater' homeowners to stay in their homes (Calomiris, Longhofer and Miles, 2013; Foote, Gerardi and Willen, 2008). Yet, the topic remains high on the public policy agenda as it lays bare the tension between housing affordability and financial stability, and carries implications for mortgage market design and macroprudential regulation.

In this paper, we study the interaction between house prices and mortgage default risk via a disaggregated analysis along two dimensions: house price segments and default risk indicators. To capture the spatial and market segment differences in house price dynamics, we use Zillow's tiered home price indices. These indices have a relatively broad coverage across U.S. Metropolitan Statistical Areas (MSAs) and account for the differences in appreciation rates between house price segments within the same geographical area. The differences in the dynamics of the two house price tiers account for the behavioral differences between owners of starter homes and trade-up homes as well as their impact on equilibrium outcomes.

2

To disaggregate mortgage default risk, we use two measures that cover the time span from the moment where an individual household starts seeking mortgage foreclosure help online to the moment where the home is foreclosed. One of the indicators is the Mortgage Default Risk Index (MDRI, hereafter) constructed by Chauvet, Gabriel and Lutz (2016) which measures the default risk of households divulged through online Google searches. The MDRI is derived from search information by mortgage borrowers and can be perceived as an index measuring the household fear of default and foreclosure. The other indicator is the number of Homes Foreclosed per 10,000 homes (HF, hereafter) for each of the studied MSAs. These two indicators account for both household sentiment and real economic activity. The approach of disaggregating mortgage default risk and housing market segments helps develop a better understanding of the relative size of the strategic effects related to the interaction between homeowners and lenders versus the macroeconomic equilibrium effects related to house price adjustments.

Recent research has established that differences in the legal system across U.S. states have important implications for borrower default behavior (Ghent and Kudlyak, 2011; Mian, Sufi and Trebbi, 2015). We therefore further examine how the relationship between mortgage default risk and house prices differ across recourse and non-recourse states and in judicial and non-judicial states. In recourse states, lenders can pursue a deficiency judgment against borrowers if the sale of the foreclosed property is not sufficient to cover the outstanding mortgage debt. In non-recourse states, in contrast, the lender is not able to collect on mortgage debt beyond the amount raised through the foreclosure sale proceedings. In judicial

3

states, the lender has to file a lawsuit in court in order to foreclose on a delinquent borrower

while in non-judicial states court order is not required.1

We examine how local house price dynamics interact with mortgage default risk and

actual default behavior of households. To account for the heterogeneity across local

residential areas, we use a Panel Vector AutoRegressive (PVAR) specification in which we

treat MSAs as cross-sectional units (Abrigo and Love, 2015; Glaeser, Gyourko, Morales, and

Nathanson, 2014). This specification allows us to exploit the variation in local market

conditions across different MSAs in order to identify the linkages among mortgage default

risk and housing market variables without imposing restrictions (Calomiris, Longhofer, and

Miles, 2013; Canova and Ciccarelli, 2013).

We find that, for the entire sample and in recourse (judicial) states, shocks to the MDRI

index affect both tiers but have a stronger impact on high-value homes, while shocks to the

HF have a stronger impact on low-value homes. However, in non-recourse (non-judicial)

states, shocks to the two default risk indicators have similar impacts on bottom-tier and

top-tier house prices. One possible explanation is that homeowners of high-value homes

search more intensely online for foreclosure help, yet they are able to avert foreclosure when

they default on their mortgages. Non-foreclosure methods to settle mortgage debt such as a

short sale2 or a deed-in-lieu of foreclosure3 were often used by homeowners during the

1 We attribute states to the Recourse and Non-Recourse categories following the categorization presented in Ghent and Kudlyak (2011). Further, states are placed into the Judicial and Non-Judicial categories following the categorization presented at . A state is placed into the Non-Judicial state category only when the categorization therein is "Non-Judicial only". For more details on the differences between Judicial and Non-Judicial states see Mian, Sufi and Trebbi (2015). 2 In a `short sale', with the approval of the lender, borrowers sell their home to a third party and use the proceeds to settle their mortgage debt. The seller agrees to a less than full repayment of the outstanding mortgage balance in order to avoid the larger losses that would be incurred in a foreclosure procedure. 3 In a `deed-in-lieu of foreclosure' borrowers transfer the deed to the property to the lenders and the house becomes a real estate owned property.

4

financial crisis and allowed them to walk away from their investment without repaying their debt in full.

Examining the reverse direction of causality, we find that a shock to top-tier house prices has a stronger effect on mortgage default risk in non-recourse (non-judicial) states. By contrast, a shock to bottom-tier house prices has about the same effect in both recourse (judicial) and non-recourse (non-judicial) states. The higher sensitivity of mortgage default risk to high-tier house prices across recourse (judicial) vis-?-vis non-recourse (non-judicial) states is indicative of strategic default behavior at the upper end of the market. That is, affluent homeowners take advantage of the put option4 inherent to their mortgage contract. We find no evidence for this effect at the lower end of the market.

Our work complements the findings by Calomiris, Longhofer, and Miles (2013), who estimate a vector autoregressive model in which state-level macroeconomic variables such as employment, building permits and home sales interact with house prices and foreclosures. They find that the effect of house prices on foreclosures is substantially greater than the effect of foreclosures on prices. That is, the relationship between house prices and mortgage defaults is predominantly based on the endogenous adjustment of foreclosures to prices rather than by the downward pressure that foreclosure-related excess supply (or other types of externalities, e.g., disamenity effects) exerts on house prices.5

4 According to the option theory of default, homeowners prefer to default on their mortgage loans when they are sufficiently deep `underwater' even if they are able to afford their mortgage payments (Kau, Keenan and Kim, 1994; Deng, Quigley, and Van Order, 2000). 5 Calomiris, Longhofer and Miles, (2013) argue that if the relationship between house prices and foreclosures is dominated by the strategic reaction of households to declining house prices, rather than by house adjustments to market equilibrium forces, initiatives to limit foreclosures will not stem the decline in house prices.

5

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download