The Subprime Lending Crisis

The Economic Impact on Wealth, Property Values and Tax Revenues, and How We Got Here
Report and Recommendations by the Majority Staff of the Joint Economic Committee 

SENATOR CHARLES SCHUMER, CHAIRMAN & REP. CAROLYN B MALONEY, VICE CHAIR 25oct2007

Executive Summary

As the losses caused by the subprime lending crisis continue to work their way through the financial markets, there is a growing awareness among policymakers and financial market regulators that we need to prevent the continuing foreclosure wave from affecting the broader economy. A significant increase in lax (and often predatory) subprime lending during a period of rapid housing price appreciation put risky adjustable rate mortgages in the hands of vulnerable borrowers who are now facing substantial payment shocks and risk foreclosure when their loans reset this year and next.

Part I of this report shows that unless action is taken, subprime foreclosure rates are likely to increase as housing prices flatten or decline, and the effects of the subprime crisis are likely to extend beyond the housing market to the broader economy. The decline in housing wealth will negatively affect consumer spending, and the forced sale of large numbers of homes is likely to negatively impact the prices of other homes.

Part II of this report shows that, unless action is taken, the number and cost of subprime foreclosures will rise significantly. For the period beginning in the first quarter of 2007 and extending through the final quarter of 2009, if housing prices continue to decline, we estimate that subprime foreclosures alone will total approximately 2 million.

Part II also includes forward looking, state-level estimates of subprime foreclosures and associated property losses and property tax losses, covering the second half of 2007 through the end of 2009. For that shorter period, and assuming only moderate housing price declines, we estimate that:

• Approximately $71 billion in housing wealth will be directly destroyed through the process of foreclosures.
• More than $32 billion in housing wealth will be indirectly destroyed by the spillover effect of foreclosures, which reduce the value of neighboring properties.
• States and local governments will lose more than $917 million in property tax revenue as a result of the destruction of housing wealth caused by subprime foreclosures. Part III of the report highlights the underlying causes of the subprime crisis and explains how incentive structures in the subprime market work against the interests of borrowers and have had much to do with the dimensions of this crisis.

Finally, in Part IV, policy options aimed at reducing foreclosures and preventing the crisis from reoccurring in the future are offered.

Part I: The Housing Downturn and Its Impact on Subprime Mortgage Foreclosures

Over the past few months, as residential investment and housing prices have declined, delinquency and foreclosure rates for subprime mortgages have spiked sharply upward. The deteriorating performance of subprime loans is not surprising. As the subprime market expanded rapidly after 2001, so did the share of adjustable rate, “hybrid” loans issued to financially vulnerable borrowers. The ability of these borrowers to sustain hybrid mortgages has depended heavily on house price appreciation. As housing prices have flattened and declined, the ability of these households to refinance their mortgages has been reduced. The resulting rise in subprime foreclosures is likely to harm an already weak housing market, and the reduction in housing wealth has the capacity to reduce consumer spending and economic growth.

HOUSING PRICE DECLINES WILL WORSEN SUBPRIME LOAN DELINQUENCIES AND HOME FORECLOSURES

The root of the subprime mortgage crisis is the prevalence of troubling loans called “2/28” and “3/27” hybrid adjustable rate mortgages (ARMs) that were largely sold to financially vulnerable borrowers without consideration for their ability to afford them. A typical “2/28” hybrid ARM has a fixed interest rate during the initial two year period. After two years, the rate is reset every six months based on an interest rate benchmark (such as the London Interbank Bid Offered Rate, or “LIBOR”). In the current environment, resets have caused payments to rise by at least 30 percent, to an amount that many borrowers can no longer afford.

As a result, the delinquency and foreclosure rates for subprime adjustable rate mortgages have been sharply rising. For more information about the characteristics of subprime loans and borrowers, see Box A [below].

When housing prices were rising, subprime borrowers could sell or refinance their homes to pay off their loans before they reset to unaffordable rates. As housing prices flatten or decline, these options dwindle. This section explains how the weakening housing market is likely to impact subprime delinquencies and foreclosures in the months ahead. For a detailed examination of the subprime market and its expansion, see Box B [below] .

Subprime Lending Has Depended on Rapid House Price Appreciation

The period of rapid housing price appreciation that began in 1997 has helped fuel increased volumes of subprime lending and masked the weaknesses in underwriting quality and predatory tactics that accompanied it.

Beginning in 1997, the U.S. witnessed house price appreciation that was highly unusual in historical terms. Between 1997 and 2006, real home prices increased by nearly 85 percent.1 Sustained price increases near this magnitude have only been observed once during the twentieth century, in the period immediately after World War II2 (See Figure 1). In fact, during the period 2001 through 2005, the annual rate of house price appreciation accelerated. The S&P/Case-Shiller® Home Price Index shows annual price appreciation rising from slightly over eight and one-half percent in 2001 to more than 15 percent in 2005.

Figure 1: U.S. Housing Market in Historical Perspective Shiller U.S. Real Housing Price Index and Other Economic Indicators, 1938-2007

Source: Irrational Exuberance, 2nd Edition, 2005, by Robert J. Shiller, Figure 2.1 as updated by author.

Not every part of the housing market witnessed this rate of home price appreciation. In some states and cities there was significant price appreciation, while it was more moderate in others. For example, Figure 2 shows the difference between home price appreciation in Michigan, Ohio, California, and Florida. But price increases were sufficiently widespread to produce significant nationwide increases in housing prices.

Housing Price Appreciation Reduced Subprime Delinquencies and Foreclosures

The deterioration in underwriting standards in the subprime market as the market expanded is well documented. (For a discussion on declining underwriting standards in subprime lending, see Box B.) Although underwriting standards in the subprime lending market began to decline after 2001, the effects of this decline were, until recently, mitigated by house price appreciation. If a borrower is struggling to make mortgage payments, but the value of his house has appreciated, he can solve his financial problems at least temporarily by refinancing the mortgage. Cash can be withdrawn from the increased equity in the house, and the new, higher mortgage can be sustained for a while. The house can also be sold, and the loan principal repaid. However, when house price appreciation does not create equity, borrowers’ financial weakness cannot be disguised and default rates rise.

Figure 2: House Price Appreciation Has Varied Across States House Price Index for Homes in Michigan, Ohio, California and Florida, Q1:1995-Q2:2007

Source: Office of Federal Housing Enterprise Oversight

There is systematic evidence that when home prices appreciate, subprime mortgage defaults decline. Using a very large sample of subprime mortgages securitized between 1999 and 2002, researchers at the Center for Responsible Lending found statistically significant correlations between the odds of foreclosure and cumulative price appreciation in a Metropolitan Statistical Area (MSA).3

The option to sell or refinance also should reduce delinquencies, which are the precursors to default and foreclosure. Recent work by economists at the Federal Reserve Bank of San Francisco shows strong negative correlations between delinquency rates and cumulative house price appreciation across MSA’s during 2006.4 This research also indicates that house price appreciation significantly improved the performance of subprime loans.

SUBPRIME PROBLEMS ARE LIKELY TO ACCELERATE HOUSE PRICE DECLINES

The Housing Market Is Contracting

Unfortunately, conditions in the housing market indicate that house price appreciation will no longer be able to disguise the financial precariousness of the millions of borrowers whose subprime adjustable rate mortgages are about to reset. The decade of steady house price appreciation appears to be at an end. Nationally, house prices began to decline in 2006 and are now down approximately 3.2 percent from their peak in the second quarter of 2006.⁵

Figure 3: Home Production Has Outpaced Demand 

Source: Bureau of the Census, U.S. Department of Commerce.

In fact, the housing market has contracted significantly for more than a year. Inventories of unsold new homes have increased, and the monthly supply of new homes has risen (See Figure 3). The Federal Reserve has estimated that so far, declines in residential investment have reduced the annual rate of GDP growth by about three-fourths of a percent over the past year and a half.⁶

A Housing Asset Bubble May Be Bursting

As residential investment in construction declines and house prices fall, there is reason to be concerned about the longer term prospects for housing values. There is apprehension that the economy is experiencing the bursting of a housing price “bubble” – a situation in which housing prices are high only because market participants believe that prices will be high tomorrow. In other words, home prices deviate significantly from the equilibrium level consistent with market fundamentals. When an asset bubble bursts, large price appreciation can be followed by sudden and large price declines.

If a housing price bubble does exist, then house price levels can be affected dramatically by shifts in expectations.⁷ There is some evidence that expectations about housing prices are changing. The National Association of Home Builders/Wells Fargo Housing Market Index (HMI), based on monthly surveys of a panel of homebuilders, reached an historic low in October 2007.⁸ See Figure 4.

Subprime Foreclosures Will Put Additional Downward Pressure on the House Prices

It is widely expected that, as the large number of subprime 2/28 and 3/27 hybrid ARMs originated during and after 2004 reset to their higher payment rates, the volume of subprime delinquencies and defaults will rise substantially. Many financially vulnerable borrowers will be facing substantially higher payments, and the lack of house price appreciation will prevent sale or refinance.

The Federal Deposit Insurance Corporation (FDIC), citing First America LoanPerformance data on securitized subprime and near-prime (so-called “Alt-A”) mortgages, estimated in March 2007 that there were approximately 2.1 million hybrid nonprime ARMs outstanding. LoanPerformance data cover about 70 percent of subprime originations.¹³ This implies that as of March there were roughly 3 million nonprime mortgages, many of which will reset in the next three years.

From Mortgage Bankers Association (MBA) data we know that the average value of all sub- prime ARM loans in 2005 was about $200,000. If we use this number as the average value of for all nonprime loans then there were approximately $600 billion in outstanding non- prime mortgages as of March. Since then, the number and amount of hybrids yet to reset will be somewhat smaller. However, the numbers are significant.

Figure 4: Expectations About Housing Market Reached Historic Lows in October 2007 NAHB/Wells Fargo Housing Market Index (HMI) and Its Three Components Seasonally Adjusted, January 1985-August 2007

Source: National Association of Homebuilders

A NOTE ON THE HOUSING BUBBLE DEBATE

There is a substantial body of economic research that attempts to explain housing prices in terms of supply and demand fundamentals such as construction costs, interest rates, employment growth, and household income.⁹ On the basis of this line of research, some economists argue that the housing price appreciation we have witnessed is not a bubble. These economists focus on the characteristics of local markets, and argue that once accurate measures of local supply and demand factors are carefully examined, there is scant evidence that housing prices have deviated significantly from fundamental values.¹⁰

There is, however, substantial evidence pointing in the other, less sanguine direction. Using state- level data for 1985 through 2002, Case and Shiller provide econometric evidence that, in eight states, fundamentals do not explain home price appreciation.¹¹ Dean Baker from the Center for Economic and Policy Research argues that at the aggregate level it is difficult to point to changes in economic fundamentals that convincingly explain why housing prices began to increase in the mid-1990’s, rather than at some other time.¹² He points to data showing that GDP, income, and population growth during this period were not unusually high, and notes that any constraint on supply caused by urban density or building regulation surely existed well before prices began to climb. The data in Figure 1 are consistent with the points made by Baker.

While many outstanding subprimes are hybrids, there are many other subprime borrowers who are also at high risk of default. Several studies of subprime mortgages show that cumulative default rates are very high. Estimates range from almost 18 percent to more than 20 percent. ¹⁵ Should housing prices decline further, cumulative defaults are likely to increase.

Using data on individual subprime mortgages originated between 1998 and the first three quarters of 2006, researchers at the Center for Responsible Lending estimated cumulative foreclosures of 2.2 million, with losses to homeowners of $164 billion.¹⁶ Although this forecast tried to take account of the effect of slowing house price appreciation, it was published in December 2006. Since that time housing prices have continued to decline.

THE EFFECTS OF FORECLOSURES AND HOUSE PRICE DECLINES WILL BE SIGNIFICANT

Foreclosures Will Harm Neighboring Home Owners and Local Housing Markets

Foreclosures can have a significant impact in a community in which the foreclosed property is located. This is particularly true when the factors that led to one foreclosure drive a concentration of foreclosures in the same neighborhood, for example in a spatial concentration of sub- prime lending. A concentration of home foreclosures in a neighborhood hurts property values in several ways. A glut of foreclosed homes for sale depresses home market values for the other owners. Neighboring businesses often experience a direct monetary loss from reduced sales and neighborhood landlords experience a loss or reduction in rental income. Moreover, the homes left vacant by foreclosure lower the desirability of the neighborhood since there is often an increase in crime associated with a vacant house.¹⁷

As concentrated foreclosures persist in a community, the value of surrounding homes may decline. Dan Immergluck and Geoff Smith survey the literature on this subject and estimate the impact of foreclosures on nearby property values using data on foreclosures and neighborhood characteristics in the Chicago area.¹⁸ They found that conventional foreclosures have a statistically and economically significant effect on nearby property values. In particular, they found that each conventional foreclosure within a one-eighth mile of a single-family home produces at least a 0.9 percent lower property value, and may be closer to 1.5 percent in low to moderate income communities.

Similarly, Shlay and Whitman find significant affects of abandoned property on nearby housing values in Philadelphia.19 They find that an abandoned property will lower property values on homes located within 150 feet by $7,627 (or 10.1 percent) and will lower property values on homes located within 450 feet by $3,542 (or 4.7 percent). As did Immergluck and Smith in Chicago, Shlay and Whitman find that the effects of abandoned properties on nearby home values are cumulative. They find that, on average, home values on the block decline by 9.1 percent in the case of one abandoned home on the block, and decline on average by 15.0 percent for 5 abandoned properties on the block.

Large House Price Declines Have the Potential to Reduce Growth and Employment

Should housing prices decline dramatically, the effects could be significant. To the extent that price declines reflect a decline in demand for new housing, construction activity will decline. This contraction is already under way, and has reduced residential investment sufficiently so that GDP growth has declined markedly in the past year.

House price declines can also affect economic activity through their effect on household wealth. Econometric work has established that household wealth, along with income, helps to determine the level of aggregate consumption. Higher levels of wealth lead to higher consumption, all things being equal. Since declines in home prices reduce wealth, they reduce consumption and thus output and employment.²⁸ These effects occur with significant time lags.

Federal Reserve Board Governor Frederic Mishkin has reported on simulations of Federal Reserve macroeconomic models of the U.S. economy in which housing prices are assumed to experience an exogenous 20 percent decline. One model shows real GDP declining one-half percent relative to baseline after three years, another shows a GDP decline of one and one- half percent, with the largest decline occurring somewhat earlier.²⁹

While these outcomes are significant, they may understate the effects of large price declines. If the price of houses were to fall 20 percent in a short period of time, we might well see a shift in overall business confidence. This could produce negative effects on credit markets, as recent events have illustrated. Higher interest rates or restrictions on business credit can in turn reduce real economic activity. In addition, business decision-making and capital investment can be affected by any changes in confidence.

Part II: State-Level Estimates of the Economic Effects of Subprime Foreclosures

To better understand how subprime lending and declining housing prices may affect households and communities in the near future, we have made quantitative estimates of the potential scale of foreclosures and their costs at the state and national levels. We first discuss entirely forward looking, state level estimates, covering the second quarter of 2007 through the end of 2009. We estimate the number of foreclosures, the loss in housing value that directly results from each foreclosure, the effect that a foreclosure has on the value of neighboring houses, and the state and local government tax revenues that will be lost as housing values decline.

As is made clear below, these state level estimates rely on housing price forecasts which show moderate housing price declines. It was necessary to use these forecasts to obtain state level results. However, it is quite possible that housing price declines will be substantially larger. Therefore we also present national level foreclosure and property loss estimates, assuming larger future housing price declines. This allows us to learn about the scale of economic damage if the housing market evolves in a less favorable way.

The results of the state level estimates, although based on forecasts of moderate housing price decline, are quite sobering. We estimate there will be approximately 1.3 million foreclosures and a loss of housing wealth of more than $103 billion through the end of 2009 (including approximately $71 billion in direct costs to homeowners and $32 billion in indirect costs caused by the spillover effects of foreclosures). The estimated aggregate cumulative subprime foreclosure rate for this period is 18 percent (See Figures 5 and 6). The total loss in property tax revenue is also high, amounting to more than $917 million. The ten states with the greatest number of estimated foreclosures, in descending order, are California, Florida, Ohio, New York, Michigan, Texas, Illinois, Arizona, Pennsylvania and Indiana.³⁰ There are, unfortunately, several others that are close behind in the rankings.

The effects of larger price declines could considerably increase the magnitude of these damages. For example, Moody’s forecasts that, in the aggregate, housing prices will decline by about 6.9 percent between Q3 2007 and Q2 2009 and rise mildly thereafter. If we instead assume that the aggregate price decline is 20 percent over that period, the total number of foreclosures for the period beginning in the first quarter of 2007 and extending through the final quarter of 2009 would be nearly 2 million and the loss of property values would total about $106 billion.

Several assumptions are necessary to make the state level estimates, and we have been deliberately conservative when making them. We have assumed that all foreclosures over the 2007-2009 period will come from the stock of subprime mortgages outstanding at the end of the second quarter of 2007. This is a very conservative assumption. The growth in the outstanding stock of subprime loans through the second quarter of 2007 indicates that incremental subprime loans are still being made. However, because we cannot forecast the course of future lending, we assume that all foreclosures come from the existing stock. This biases our estimates downward. We also assume that once a mortgage enters foreclosure it is foreclosed within a year. Although there are variations across jurisdictions, the average maximum amount of time to foreclose is less than a year.³¹

Figure 5: Impact of Subprime Foreclosures on Home Equity, Property Values and Property Taxes 

Sources: Number of outstanding subprime mortgages and current subprime foreclosure rates from Mortgage Bankers Association survey data; average home value calculated using the 2006 Home Mortgage Disclosure Act (HMDA) data for subprime first-lien loans and loan-to-value ratios courtesy of the Center for Responsible Lending; historical home price indices from the Office of Federal Housing Enterprise Oversight (OFHEO); forecasts of OFHEO price indices from Moody's Economy.com; Congressional Budget Office (CBO) forecasts of personal consumption expenditure deflators; state property tax rates from U.S. Census Bureau and the Tax Foundation; state household densities, by MSA, from the U.S. Census Bureau.

 

Figure 6: Projected Economic Costs of the Subprime Mortgage Crisis State-by-State

Source: JEC Calculations.

Figure 7: State-Level Foreclosure Rate Regressions

		               Dependent Variable        
Independent	     Foreclosure Rate	Foreclosure Rate
Variable	         ARM		      FRM

House Price            -14.80 **           -9.27 **
Appreciation
(2004-2006)            (2.058)             (1.655)

Employment Growth      -19.22 **           -11.19 *
(2004-2006)            (5.930)             (5.384)

Constant               8.88 **             5.42 **
                       (0.570)             (0.523)

Observations           51                  51

R2           	       0.712               0.490

* Significant at 95% level. 
** Significant at 99% level. 

Data Sources: Foreclosure rate are Mortgage Bankers Association “foreclosure inventory”;
House Price Appreciation is calculated from Office of Federal Housing Enterprise
Oversight housing price indices; Employment Growth is calculated from Bureau of Labor 
Statistics “employees on non-farm payrolls,” seasonally adjusted. All data accessed via
Haver Analytics.

To estimate the numbers of mortgages that will be foreclosed, we begin by examining what determines the fraction of mortgages in foreclosure (foreclosure rate) during a year. It is reasonable to suppose that, holding the risk characteristics of borrowers constant, the foreclosure rate will depend heavily on house price appreciation and the economic fortunes of borrowers. ³² If house prices appreciate, refinance or sale is easier. If general economic conditions are good, it is more likely that households will be able to meet their financial commitments. As it turns out, both these factors are significant determinants of the foreclosure rate. Figure 7 shows the results of state-level cross sectional regressions of subprime foreclosure rates for 2006 on two independent variables – cumulative housing price appreciation between 2004 and 2006, and cumulative employment growth in the same period. The cumulative housing price appreciation variable is an index of changes in home equity, and the cumulative employment growth variable is an index of the ease of finding employment and the overall performance of the real economy. Both variables are statistically significant. The significance of the employment variable highlights the importance of developments in the real economy for loan outcomes. However, we do not attempt to estimate changes in employment when we use these results. If employment growth were to slow during our forecast period, foreclosure rates likely would be higher than our estimates.

To estimate future foreclosure rates, we use current foreclosure rates, the coefficients on house price appreciation reported in Figure 7, and estimates of future housing prices. That is, we calculate foreclosure rates according to FC_t = FC_t-1 + ß(HPA_t), where FC_t is the foreclosure rate in year t, FCt-1 is the foreclosure rate in the previous year, .HPAt is the change in cumulative two-year housing price appreciation between years t and t-1, and ß is the estimated coefficient of HPA (house price appreciation) as reported in Figure 7. The values for the variable HPA_t are calculated using forecasts of state-level housing price indices from the Office of Federal Housing Enterprise Oversight (OFHEO). The forecasts were produced by Moody’s Economy.com. We estimate foreclosure rates separately for fixed rate and adjustable rate mortgages. These foreclosure rates are used to calculate the absolute number of foreclosures in a given period. ³³

Using our estimates of the number of subprime foreclosures, we then estimate the associated economic costs. Research has shown that foreclosure causes a decrease in the value of the foreclosed house.³⁴ We estimate this direct loss in housing wealth by discounting the average loan value of a subprime mortgage. We apply a 22 percent discount rate to the average home value associated with subprime loans (net of the loss due to the decline in home prices) to calculate this loss. ³⁵

Foreclosures also affect the values of neighboring houses. We estimate the effect of a foreclosure on surrounding house prices as 0.9 percent of the value of all single family houses within 1/8th mile of a foreclosed house.³⁶ We use MSA-level population densities to estimate the number of houses within one-eighth mile of each foreclosed house.³⁷

The loss in property taxes caused by housing price losses is calculated by assuming that average state property tax rates remain unchanged through the end of 2009. Tax losses are calculated by applying existing property tax rates to the change in housing values caused by foreclosure (net of the loss due to the decline in home prices).

We conclude by noting that the forecast values for housing prices clearly play a pivotal role in this analysis, and that the price forecasts we have used are likely to be conservative. The Moody’s data are forecasts of future values of OFHEO housing price indices. However, in recent quarters the OFHEO indices have not reflected the same downward movement in housing prices registered in other price measures. For example, the national OFHEO index had not peaked by the second quarter of 2007, but the S&P/Case-Shiller® U.S. national home price index peaked in the second quarter of 2006 and had declined by 3.2 percent by the end of the second quarter of 2007. Therefore it is possible that the price forecasts we have used will not pick up all of the likely housing price declines over the near term.

To account for this possibility, we have applied the procedure developed for state level estimates to aggregate foreclosures, assuming a 20 percent decline in aggregate home prices. A price decline of that amount is not out of the question. When simulating the possible macroeconomic effects of housing price declines, the Federal Reserve recently assumed a 20 percent decline in aggregate housing prices.³⁸ Moreover, futures contracts based on the S&P/ Case-Shiller® indices are predicting that housing prices may decline as much as 10 percent over the coming year.³⁹ Since the S&P/Case-Shiller® indices already show a 3.2 percent decline over the past year, calculating subprime foreclosures by assuming a 20 percent decline in the OFHEO price indices over two years seems unfortunately plausible. Under these assumptions, the number of foreclosures for the period covering the third quarter of 2007 through the end of 2009 is approximately 1.66 million, and the associated property loss is about $106 billion.⁴⁰ If we add in an estimate of foreclosures in the first half of 2007,

Part III: The Origins of the Subprime Lending Crisis

The discussion above highlights the potential economic damage that could result if subprime foreclosures are allowed to proceed unchecked. In this section we investigate the underlying causes of the subprime mortgage crisis in an effort to identify policy approaches that could prevent the reoccurrence of such a threat to homeownership, household wealth, and the broader economy.

FINANCIAL INTERMEDIARIES DROVE THE EXPANSION OF THE SUBPRIME MARKET

Most Lending Organizations Make Few Subprime Loans

The expansion of subprime mortgages during the years 2001 through 2006 came, for the most part, through a well defined channel of financial intermediaries. The intermediaries in this channel – brokers, mortgage companies, and the firms that securitize these mortgages and sell them on to the capital markets – had strong incentives to increase the supply of these loans. One outcome was a significant increase in the rate of homeownership. From 1994 to 2005, the overall homeownership rate rose from 64 to 69 percent.⁴¹ However, since brokers and mortgage companies are only weakly regulated, another outcome was a marked increase in abusive and predatory lending.

Most Subprime Loans Are Originated Through Mortgage Brokers

The mortgages underwritten by subprime lenders come from many sources, but the overwhelming majority is originated through mortgage brokers. For 2006, Inside Mortgage Finance estimates that 63.3 p