Follow Insights Blog

CoreLogic

CoreLogic Econ

LATEST CORELOGIC ECON TWEETS

Locating Home Price Risk with the CoreLogic Home Price Index Forecasts Stress-Testing Scenarios

David Stiff    |    Mortgage Performance, Property Valuation

Bank and bank holding company stress tests rigorously assess financial institutions’ capital adequacy, including the development and maintenance of effective loss-estimation methodologies. This is mandated by the Federal Reserve’s Comprehensive Capital Analysis and Review (CCAR) and the Dodd-Frank Act. Home Price Index (HPI) supervisory scenarios are specified for the entire nation through the Federal Reserve, which makes them less useful for stress-testing since a financial institution’s residential mortgage holdings are usually unevenly distributed across the country. Moreover, even if an institution’s holdings are evenly distributed, its home price and mortgage risk will still be heavily concentrated in a small subset of geographic areas.

During the 2006-2011 U.S. housing market crash, the worst home price declines and mortgage losses were confined to a small fraction of metro areas, and within those metro areas, inside a subset of neighborhoods and towns. The CoreLogic HPI Forecasts Stress-Testing Scenarios help banks, investors, government entities and others effectively satisfy critical regulatory requirements and—available for all states and Core Based Statistical Areas (CBSAs), plus counties and ZIP Codes with historical CoreLogic HPI coverage —can be used to identify geographic pockets of home price risk since they disaggregate the Federal Reserve’s national-level stress scenarios.1

In Table 1, CBSAs where single-family home prices dropped by more than 40 percent during the 2006-2011 housing market crash are separated from all other CBSAs. Only 71 out of 937 CBSAs fall into the worst-performing group, accounting for just 14 percent of owner-occupied housing stock in the first quarter of 2006. However, despite making up a small fraction of the national housing market, these CBSAs generated disproportionate housing market losses. The aggregate value of housing fell by $1.5 trillion in these markets, which is one-third of all CBSA losses. Foreclosure rates were also disproportionately higher, so that the worst-performing markets produced one-third of completed foreclosures between 2006 and 2011. The second to last row in this table contains estimates of the aggregate losses of residential mortgage lenders from 2006 to 2011,2 when the worst-performing CBSAs were responsible for more than one-half of aggregate mortgage losses. Finally, at the end of the home price crash, unemployment rates in these CBSAs were substantially higher due to the direct and indirect effects of devastating housing market losses.

Why were mortgages losses so much greater in the worst-performing CBSAs? Demand for residential properties fell in nearly every market during the housing market crash. The collapse in private mortgage lending along with the largest job losses since the Great Depression directly undercut housing demand. But even homeowners who kept their jobs and access to mortgage credit were not interested in buying an asset that was falling in value. In markets with significant home price declines, these direct and indirect effects spawned demand-side negative feedback loops, while falling prices caused housing demand to drop which led to further price declines. In the worst-performing CBSAs, this effect was compounded by supply-side feedback loops. On the supply-side, rapidly falling home prices pushed large numbers of mortgage borrowers underwater, leading to record numbers of foreclosures. As foreclosures mounted, the worst-performing markets became overwhelmed with for-sale inventories of bank-owned properties, putting additional downward pressure on prices.

Even worse, in the hardest-hit markets, the demand-side and the supply-side negative feedback loops combined to generate even larger employment and housing market losses. This is because real estate-related industries (residential construction, real estate brokerage, retail sales at home improvement stores, etc.) were a huge source of job and income growth in these markets during the housing market boom. When this process reversed itself, jobs losses in the real estate sector accelerated the negative demand-side feedback loops. At the same time, increasing supplies of foreclosed properties amplified real estate-related job losses by bringing new housing construction to a grinding halt.

In Table 2, the cross-sectional correlations between CBSA-level home price changes, unemployment rates and foreclosure rates illustrate these feedback loops. In the better performing CBSAs, the strongest correlation is between 2006-2011 home price changes and the post-crash unemployment rate. In those markets, job losses undercut the demand for housing so that local housing market downturns were mostly limited to the demand-side and foreclosures exerted only moderate downward pressure on home prices. In the worst-performing markets, there is a similar correlation between job losses and falling home prices (demand-side), but substantially larger correlations between foreclosure rates and falling home prices (supply-side), as well as between the unemployment and foreclosure rates (demand-side and supply-side interaction). Foreclosure activity was far more problematic for these markets because their oversized real estate-related industries increased the vulnerability of both the demand-side (job losses) and supply-side (rising for-sale inventories) to falling home prices.

Feedback Affects Amplified Worst Housing Market Downturns

Feedback Affects Amplified Worst Housing Market Downturns

The CoreLogic HPI Forecasts Stress-Testing Scenarios can help identify CBSA-level markets that are the most vulnerable to large home price declines in stressed economic and housing market conditions. These scenarios are constructed by combining historical HPI information with CoreLogic’s standard HPI forecasts (which incorporate two supply-side drivers – housing starts and REO inventories). In a future blog post, I will describe why a CBSA’s past HPI performance can be a good metric for its future vulnerability to housing market stress and how the county and ZIP Code-level stress-testing scenario forecast can be used to identify localized pockets of home price risk within a CBSA.

[1] CoreLogic now produces HPI Forecast Stress-Testing Scenarios that are consistent with the Federal Reserve’s National Supervisory Scenarios for the Comprehensive Capital Analysis and Review (CCAR) and Dodd-Frank Act Stress Testing (DFAST) programs. Since the Federal Reserve’s national scenarios are developed in collaboration with the OCC and the FDIC, the new CoreLogic scenario forecasts are applicable for stress-testing by any financial institution with more than $10 billion in assets. The HPI Forecast Stress-Testing Scenarios may also be useful for smaller institutions since OCC guidelines state that: "The OCC expects every bank, regardless of size or risk profile, to have an effective internal process to (1) assess its capital adequacy in relation to its overall risks, and (2) to plan for maintaining appropriate capital levels.” Furthermore, the National Credit Union Administration (NCUA) issued rules that will require federally-insured credit unions with more than $10 billion in assets to perform annual stress tests.
[2] These calculations rely on counts of outstanding loans and completed foreclosures from the CoreLogic MarketTrends database. The estimates of aggregate mortgage losses almost certainly underestimate actual financial institution losses for two primary reasons: (1) the Market Trends database covers about 85 percent of outstanding loans and (2) in the calculations, lenders and investors are assumed to recover the market value of their foreclosed properties. Nevertheless, the estimates of each CBSA’s relative contribution to aggregate mortgage losses should be fairly accurate.

​© 2015 CoreLogic, Inc. All rights reserved.