We no longer support this browser. Using a supported browser will provide a better experience.

Please update your browser.

Close browser message

RESEARCH Falling Behind

Bank Data on the Role of Income and Savings in Mortgage Default

Findings

Download

Introduction

For many American households, buying a home represents one of their largest lifetime expenditures. And because most homeowners finance their home purchase with a mortgage, buying a home is also one of their largest sources of debt. In our report Mortgage Modifications after the Great Recession: New Evidence and Implications for Policy, we measured the impact of mortgage payment and principal reduction on default and consumption. We found that a 10 percent mortgage payment reduction decreased default rates by 22 percent, whereas for borrowers who remained underwater, principal reduction had no effect on default or consumption. This finding implies that short term liquidity was a key factor driving mortgage default.

In fact, we found that for homeowners who defaulted on their mortgage, default was correlated with a drop in income. This is illustrated in Figure 1, which is reproduced from Mortgage Modifications after the Great Recession and shows the relationship between income loss and default for a sample of de-identified Chase mortgage customers who had a Chase deposit account and defaulted on their mortgage.1,2 Figure 1 shows the change in the path of monthly income (where income is defined as all checking account inflows) and mortgage payment made over the 12 months before and after default relative to a baseline period (12 months before default).3 Income dropped in the five months leading up to default and, a few months after the initial drop in income, mortgage payments also declined until borrowers defaulted.4,5

In this follow-up research to our Mortgage Modifications after the Great Recession report, we further examined the relationship between income shocks and mortgage default. We found that the relationship between negative income shocks and mortgage default illustrated in Figure 1 held for homeowners across all levels of home equity and regardless of income level or total debt-to-income ratio (DTI) at origination. Deeper and longer duration negative income shocks were associated with increasing delinquency, whereas to the extent their income recovered quickly, homeowners promptly resumed making their mortgage payments. Homeowners with savings used their financial buffer to delay mortgage default following a negative income shock. Finally, we examined the relationships between financial buffers, income, payment burden, and default rates. Homeowners with larger financial buffers had lower default rates regardless of their income level or payment burden.

Taken together, these findings suggest that providing borrowers with an incentive to build and maintain a post-purchase financial buffer may be a more effective approach to default prevention than underwriting standards based on meeting ability-to-repay rules at origination.

Using a sample of de-identified Chase customers who had both a Chase mortgage and a Chase deposit account, we analyzed the relationships between negative income shocks, savings, and mortgage default.

Findings

  • For borrowers who defaulted on their mortgage, default closely followed a negative income shock regardless of their level of home equity.
  • For borrowers who defaulted on their mortgage, default closely followed a negative income shock regardless of their income level or payment burden.
  • Recovering from mortgage default was associated with recovering from a negative income shock; homeowners who experienced deeper and longer duration drops in income became increasingly delinquent.
  • Homeowners with larger financial buffers used their savings to delay mortgage default following a negative income shock.
  • Default rates for homeowners with small financial buffers were higher regardless of income level or payment burden. Therefore, building and maintaining a financial buffer may be a more effective tool to help borrowers avoid default than meeting total DTI standards at origination.

Figure 1: For homeowners who defaulted, a substantial negative income shock preceded their default.

Implications

From the preceding analyses, we draw implications for policy along three dimensions: default prevention, helping homeowners facing a negative income shock, and mortgage design.

Default prevention

Establishing and maintaining a financial buffer was an important component of avoiding default in the face of a negative income shock, even for higher income homeowners. The default rate for borrowers with less than the one mortgage payment equivalent held in reserve (2.54 percent) was seven times higher than the default rate for borrowers with at least four mortgage payments in reserve (0.36 percent).

The public and private sector should consider ways to provide new borrowers with an incentive to build and maintain a reserve fund associated with their mortgage that could be drawn down in the face of a negative income shock to avoid default. Designed carefully, such a program could provide a method of reducing defaults that aligns interests across the relevant mortgage stakeholders (borrower, servicer, investor, GSEs, and insurer).

Previous research has shown that liquidity is a more important determinant of default than debt level. As such, some lenders may want to consider the trade-offs between down payment size and residual cash reserves at origination. For example, a buyer who purchases a $250,000 home and makes a 21 percent down payment but is left with no savings may be more vulnerable to default compared to a buyer who puts down 20.2 percent, has a monthly mortgage payment that is $10 higher, and is left with the equivalent of 2 mortgage payments that can be held as a financial buffer and used in the event of a temporary negative income shock to avoid default.25 Paired with a reserve fund as described above, some additional flexibility around down payment and LTV limits at origination could lead to lower default rates.

A policy based on maintaining a minimum post-purchase financial buffer may be a better approach to default prevention than underwriting standards based on measuring the borrower’s static ability-to-repay at origination. Meeting the ability-to-repay rule requiring that a borrower’s total DTI at origination not exceed 43 percent to satisfy the Qualified Mortgage rule was not enough to help borrowers faced with a negative income shock avoid default. While total DTI measured at origination may have some predictive power for default, the considerable heterogeneity in housing costs and incomes makes it difficult to find a single level of total DTI that indicates affordability across all households and regions. Research has shown evidence that had the 43 percent total DTI limit been in effect after 2004, it would have resulted in a minimal reduction in five-year default rates for mortgages originated between 2005 and 2008. Thus there is little evidence that potential homeowners just below the 43 percent total DTI threshold should be treated differently than potential homeowners just above the 43 percent total DTI threshold.

A financial buffer that is maintained through the early life of the mortgage (when lower home equity levels leave homeowners who face financial difficulty with fewer choices) may prove more effective at avoiding default relative to meeting an ability-to-repay minimum standard that is only observed at origination. Placing a limit on total DTI as part of the underwriting process is inherently limited as a default prevention tool precisely because (1) total DTI will likely change over the life of the mortgage and (2) it creates no additional incentive for the homeowner to build savings to counter a negative income shock. In contrast, a policy that provides an economic incentive for borrowers to build and maintain a financial buffer of a few mortgage payments could be a more effective default prevention approach, precisely because the reserves would be available in the event of a negative income shock.

Helping homeowners facing a negative income shock

As discussed in our previous research, mortgage modification programs that aim to reduce default rates should focus on providing homeowners who are struggling to make their monthly mortgage payments with material payment reduction, regardless of their previous income level or home equity. Furthermore, the amount of payment reduction should not be predicated on reaching a predetermined affordability target.

The results in this brief further highlight the important connection between a negative income shock and default and underscore the importance of early intervention. By responding to missed mortgage payments earlier, mortgage stakeholders and homeowners can together estimate to the best of their ability the depth and potential duration of their financial difficulty and arrive at a solution before arrearages build and default ensues. Temporary forbearance could be used to mitigate the impact of transitory income shocks, whereas more permanent modifications may be appropriate for longer duration drops in income.

Considering the broad definition of income we used in our analysis, there may be many reasons why a homeowner with a mortgage suffers the type of negative income shock that we see associated with default, including job loss, a reduction in hours within a job, or a loss of transfers. For the homeowners who are experiencing a job loss, unemployment insurance can act as a complement to traditional housing policy programs. Research has shown that increasing unemployment benefits after the Great Recession reduced mortgage defaults and foreclosures. In contrast, the Federal programs designed to help homeowners with a mortgage who were facing financial difficulty and/or negative equity suffered from low take-up rates, suggesting these programs may have been less impactful. Closer coordination between housing policymakers and the federal and state policymakers charged with determining the amount and duration of unemployment insurance benefits could better ensure that the marginal tax dollar aimed at helping homeowners facing a negative income shock avoid default is invested efficiently.

Research estimates that increases in unemployment insurance payment amounts or duration helped avoid 1.3 million mortgage foreclosures between 2008 and 2013 and therefore acted as an automatic stabilizer for the housing market during this period.26 This result is consistent with our findings regarding the impact of mortgage payment reductions on default rates, and can be generalized—policies that either reduce monthly payments or replace monthly income for homeowners struggling to make their monthly mortgage payments will reduce subsequent default rates.

In contrast, the various federal programs implemented to help unemployed or underemployed homeowners with a mortgage suffered from relatively low subscription rates and therefore had a much smaller impact on foreclosures. The Hardest Hit Fund (HHF) was established in 2010 to help homeowners hit hardest by the economic and housing market downturn and offered mortgage payment assistance for unemployed or underemployed homeowners. By the end of 2016, only 292,000 homeowners had received assistance through the HHF. The Home Affordable Unemployment Program (UP) was introduced in 2010 to provide assistance to homeowners who were unable to make their mortgage payments as a result of unemployment. As of the end of 2016, only 46,485 homeowners were participating in the UP program.27

Similarly, the mortgage modification programs (e.g., the Home Affordable Modification Program or HAMP) designed to aid homeowners in the post-Great Recession period suffered from low uptake and had a smaller impact on foreclosures. Between March 2009 and June 2010 about 55 percent (almost 675,000) of HAMP trial modifications were cancelled because homeowners could not provide the requisite income verification documentation. By April 2015, more than one million homeowners had been denied a HAMP modification because they did not provide the financial and/or hardship verification documentation required to complete the evaluation of their request in a timely manner.28 Research estimates that the HAMP program helped avoid 600,000 foreclosures between March 2009 and December 2012.

Research has also shown that the refinancing programs introduced during the same period (e.g., the Home Affordable Refinance Program) suffered from frictions that reduced their uptake to less than 50 percent of eligible borrowers, and resulted in modest reductions to foreclosure rates. Similarly, changing refinancing programs to require income verification and upfront payment of closing costs reduced refinancing rates by 50 percent.

While state-level unemployment insurance programs have varying eligibility requirements, unemployment benefits directly target households that would be ineligible for the various mortgage programs that require income documentation.29 Unemployment insurance has the added advantage relative to mortgage programs in that it offers relief to homeowners that have lost a job without requiring input from mortgage servicers, investors, various government agencies (e.g., FNMA, FHLMC, FHA, etc.) or second lien holders. Taken together, these facts suggest that unemployment insurance can act as a complement to traditional housing policy programs to help homeowners facing a negative income shock.

Mortgage design

Our analysis has implications for housing policymakers as they consider the trade-offs between fixed-rate mortgages (FRMs) and adjustable-rate mortgages (ARMs). The connection between negative income shocks and default suggests that ARMs, which automatically adjust their interest rates and monthly payments in accordance with the Federal Reserve's interest rate target range, can serve as a way to help stabilize the economy during a recession.30 Additional empirical research is needed to better understand the impact of ARMs on consumer spending and default as policy rates normalize to higher levels.

As discussed in our previous research, accommodative monetary policy is automatically transmitted to homeowners with ARMs, while the transmission channel to homeowners with FRMs is harder to activate and requires the borrower meet various pre-conditions that limit its effectiveness. For borrowers with an FRM, this is true for both the refinancing process and the mortgage modification process.

With respect to finding the optimal mortgage design, research suggests that ARMs that include a feature that allows the borrower to reduce or defer payments in the face of a negative income shock or during a recession would reduce defaults and stabilize consumption across business cycles. These ideas should be weighed in the broader context of providing borrowers with mortgages that are simple and easy to understand.

Once armed with an assessment of the impact of ARMs on default and consumption during an economic expansion, housing policymakers can make the determination as to whether to consider the promotion and standardization of ARMs for the appropriate set of borrowers (given demographic and other characteristics).

Data Asset

In this research brief, we analyzed a sample of de-identified Chase customers who had both a Chase mortgage and a Chase deposit account. For each specific finding and figure, we focused on a slightly different part of this population as described below:

Table 1. Sample Requirements and Sizes.

 

Finding Base Requirements Figure Additional Requirements Sample Size
Introduction
  • Had one Chase mortgage and Chase deposit account(s) between October 2012 and August 2015
  • Defaulted on mortgage (see additional requirements for specifics); default date is used to define t=0
  • No more than one negative mortgage payment
  • Adequate mortgage and deposit data for us to follow them for 12 months before and 12 months after default (we include homeowners who drop out of our mortgage data after default due to foreclosure, but we drop those who prepay their loanfootnote 31)
1
  • Default = 90 days delinquent
Over 11,500
1 2
  • Default = 90 days delinquent
  • Had observed LTV at default
Over 11,200
2 3
  • Default = 90 days delinquent
  • Had observed verified income at origination
Almost 4,100
4
  • Default = 90 days delinquent
  • Had observed total DTI at origination
Over 8,200
3 5
  • Default = 90 days delinquent
  • Delinquency status transition from t=0 to t=1 is feasible
Over 10,800
6
  • Default = 30 days delinquent
Over 22,100
4 7
  • Default = 90 days delinquent
  • Had a Chase deposit account for the 6 months before baseline (t=-12)
Over 3,900
5
  • Had a Chase deposit account in January 2013
  • Had one Chase mortgage for all of 2013-2014 and did not miss more than 2 payments (i.e. default) in 2013
8
  • Had observed verified income at origination
Over 580,800
9
  • Had observed total DTI at origination
Over 820,000
10

Across all of our samples, more than 95 percent of the mortgages were originated between 2000 and 2013. For the sample of defaulted mortgages that forms the basis for Findings 1 through 4, about 60 percent were originated prior to 2008, and 40 percent were originated between 2008 and 2013, and we include both homeowners who received a modification (about 40 percent of the sample) and those who did not. For the larger sample of mortgages that forms the basis for Finding 5, about half were originated prior to 2008 and the other half were originated between 2008 and 2013.

Suggested Citation

Farrell, Diana, Kanav Bhagat, and Chen Zhao. 2018. “Falling Behind: Bank Data on the Role of Income and Savings in Mortgage Default” JPMorgan Chase Institute.


 

Acknowledgements

We thank our research team, specifically Yuan Chen, Annie Gao, and Melissa O’Brien for their hard work and contribution to this report.

We would also like to acknowledge the invaluable input of academic experts Michael Barr, Peter Ganong, and Pascal Noel, and industry expert Michael Fratantoni from the Mortgage Bankers Association, all of whom provided thoughtful commentary, as well as the contribution of other Institute researchers, including Chris Wheat, Amar Hamoudi, Chex Yu, Max Liebeskind, and Kerry Zhang. In addition, we would like to thank Michael Weinbach and the Chase Mortgage Banking team for their support, especially Peter Muriungi, Tina Shell, Bill Zaboski, Murdock Martin, and Andrew Lewis, as well as other experts within JPMorgan Chase, including Chris Henry, Subra Subramanian, Matt Jozoff, and John Sim. We are deeply grateful for their generosity of time, insight, and support.

This effort would not have been possible without the critical support of our partners from the JPMorgan Chase Consumer & Community Bank and Corporate Technology teams of data experts, including Samuel Assefa, Connie Chen, Anoop Deshpande, Senthilkumar Gurusamy, Ram Mohanraj, Karen Narang, Stella Ng, Rob Rappa, Ashwin Sangtani, Michael Harasimowicz, and Anmol Karnad, and JPMorgan Chase Institute team members Sruthi Rao, Kelly Benoit, Courtney Hacker, Jolie Spiegelman, Elizabeth Ellis, Maggie Tarasovitch,Carla Ricks, Sarah Kuehl, Alyssa Flaschner, and Gena Stern.

Finally, we would like to acknowledge Jamie Dimon, CEO of JPMorgan Chase & Co., for his vision and leadership in establishing the Institute and enabling the ongoing research agenda. Along with support from across the Firm—notably from Peter Scher, Max Neukirchen, Joyce Chang, Patrik Ringstroem, Lori Beer, and Judy Miller—the Institute has had the resources and support to pioneer a new approach to contribute to global economic analysis and insight.

Authors

Chen Zhao

Housing Finance Research Lead

Diana Farrell

Founding and Former President & CEO