Subprime Originations and Foreclosures in New York State: A Case Study of Nassau, Suffolk, and Westchester Counties Cambridge, MA Lexington, MA Hadley, MA Bethesda, MD Washington, DC Chicago, IL Cairo, Egypt Johannesburg, South Africa Executive Summary August 2002 Prepared for American Association of Retired Persons (AARP) Public Policy Institute Abt Associates Inc. 55 Wheeler Street Cambridge, MA 02138 Prepared by Kimberly Burnett Christopher E. Herbert Bulbul Kaul
Executive Summary Introduction One of the most striking features of residential mortgage finance in the 1990s was the rapid growth of subprime lending. The term subprime generally refers to loans made to borrowers who have low credit ratings, but can also include loans to creditworthy borrowers that exceed the property value. Between 1993 and 2000 the number of subprime loans to purchase homes or refinance existing mortgages increased ten-fold from about 100,000 to 1,000,000. There are both positive and negative aspects to the development of the subprime mortgage market. By providing greater flexibility in the mortgage market, subprime loans provide borrowing options for households who might otherwise not be able to access the mortgage markets to purchase or refinance a home. But at the same time subprime lending has proven to offer opportunities for unscrupulous or predatory lenders to take advantage of borrowers by charging excessive interest rates and fees and using mortgage proceeds to pay inflated costs for home repairs or insurance products. The most common victims of these predatory lending practices have been found to include the elderly, minorities, and lowincome households. While not all subprime lending is predatory, subprime and predatory loans share the distinction of charging borrowers interest rates that are higher than conventional rates. There is also evidence that some borrowers who would otherwise qualify for prime loans may be steered to subprime loans, incurring higher than necessary mortgage costs. The costs of predatory and unwarranted subprime lending for the individuals and communities affected are potentially enormous. The Coalition for Responsible Lending in North Carolina has estimated that the cost of excessive fees and interest rates on subprime loans across the country may total as much as $9 billion annually. In addition to home equity lost to high fees and other costs and income wasted on inflated interest rates, borrowers also face the loss of their home through foreclosure. Since subprime lending has been concentrated in minority and low-income neighborhoods, these communities also bear the costs of predatory subprime lending. Rising foreclosures in these communities lead to more abandoned and dilapidated properties, depressing property values and fostering neighborhood instability. Evidence on the rise in foreclosures associated with the increase in subprime lending has been found in previous studies of the Chicago, Boston, and Atlanta areas. Over the last few years the problems associated with subprime and predatory lending have attracted a great deal of attention from policy makers at all levels of government and have led to the development of legislation aimed at providing greater protection for borrowers from unfair lending practices. It remains to be seen whether these efforts will be sufficient to Abt Associates Inc. Executive Summary ii
protect borrowers. But even if these efforts are successful, there will still be a large number of households left with costly subprime mortgages that they may not be able to afford. In addition to their efforts to curtail future predatory loans, policy makers will also need to consider remedies for past victims of these practices. Purpose One purpose of this study is to examine whether subprime lending in New York State has exhibited the same rapid increase and heavy concentration in low-income and minority areas as have been evident elsewhere in the country. This analysis supports the primary contribution of this study, which is an examination of recent foreclosure trends in three New York counties (Nassau, Suffolk, and Westchester) to determine whether a rise in subprime lending has also been associated with rising foreclosures, particularly in low-income and minority neighborhoods. Previous studies on subprime foreclosures have focused on innercity areas of Chicago, Atlanta, and Boston. The purpose of this study is to evaluate whether the trends in foreclosures found in these areas are also evident in smaller metro areas or more suburban portions of metro areas. Two factors were considered in selecting Nassau, Suffolk, and Westchester Counties for study. As noted, an important goal of this study is to examine the impact of subprime lending on areas outside of the core areas of large metropolitan areas. However, our ability to select areas for study was limited by the availability of foreclosure data. Since data on foreclosures were readily available for the New York metropolitan area, Nassau, Suffolk, and Westchester counties were selected as the more suburban portions of the New York metro area. While these areas are not necessarily representative of New York State generally, they may nonetheless shed light on what may be happening elsewhere as virtually all areas in the state have experienced a sharp rise in subprime lending over the past decade. Methodology To examine trends in subprime lending, this study relies on data on purchase and refinance mortgage originations reported under the Home Mortgage Disclosure Act (HMDA) from 1994 to 2000, the latest year for which data have been reported. HMDA mandates that lenders disclose certain information on all home mortgage applications they receive, including the race, census tract, and income of borrowers and the action taken on the application. Loans are classified as either prime or subprime based on the U.S. Department of Housing and Urban Development (HUD) Manufactured Home and Subprime Lender List for 1994 through 2000. Loans originated by lenders not listed as being subprime lenders are considered to be prime loans. 1 It is important to note that while this method appears to 1 Loans originated by lenders classified by HUD as manufactured housing lenders are excluded from this analysis. Abt Associates Inc. Executive Summary iii
capture a significant portion of subprime loans reported in HMDA, it certainly does not identify all subprime loans. The recent consolidation of companies in the mortgage industry has also blurred the lines between subprime and prime lenders. More primarily conventional lenders have entered the subprime market. The method used to identify subprime loans for this study will not encompass these subprime loans made by primarily conventional lenders. In contrast to the ready availability of information on mortgage originations, information on mortgage foreclosures is much more difficult to obtain as there is no systematic reporting process similar to that required for mortgage originations. Rather, information on mortgage foreclosures can only be obtained by reviewing public notices or filings at county courthouses. Data on foreclosures for this study were obtained from Foreclosure Trac, Inc., a firm that gathers and publishes data on foreclosure notices in the New York metropolitan area. Foreclosures are measured based on published notices announcing the sale of foreclosed properties at public auction. Foreclosure information for the three selected counties was available for the period from 1998 through 2001. The process of identifying subprime lenders in the foreclosure data is similar to that used for HMDA. The creditors listed as the foreclosing entity in the public notices and records are compared to HUD s list of subprime lenders to identify subprime loans. However, several factors make this process somewhat more imprecise than is the case with HMDA data. One complication is that the foreclosing lender is not necessarily the firm that originated the loan since loans and loan servicing rights are frequently sold after origination. In addition, in recent years prime lenders have purchased many of the large subprime lenders. While in many cases the subprime entity will continue to operate under its original name, in some cases these operations may be absorbed into the purchasing lender s operations. For all these reasons, subprime loans may not have a subprime lender listed as the foreclosing creditor and so there may be a bias toward undercounting subprime foreclosures. Much of the analysis of the data on mortgage originations and foreclosures focuses on neighborhood racial and income characteristics. The neighborhoods used in this study are census tracts, which are areas defined by the Census Bureau to capture neighborhoods of about 4,000 in population. By linking the HMDA and Foreclosure Trac data to the characteristics of census tracts, we are able to characterize the distribution of mortgage lending and foreclosure activity with regard to neighborhood racial composition and income level. Subprime Lending in New York State As has been true nationally, subprime lending has grown rapidly in New York State since the early 1990s. The overall volume of subprime mortgages nearly tripled between 1994 and 1998, growing from about 15,000 loans to more than 56,000 loans. Over the entire period from 1994 to 2000 nearly a quarter of a million subprime loans were originated in the state. Abt Associates Inc. Executive Summary iv
By 2000 subprime lenders had come to account for 14 percent of all new loans. Subprime lending has been even more concentrated in refinance loans, where subprime lenders accounted for one out of every three loans in 2000. The statewide trend in rapid growth in subprime lending was also evident in Nassau, Suffolk, and Westchester Counties. Growth in lending volumes was highest in Suffolk County, where subprime loans nearly quadrupled between 1994 and 1998. Meanwhile, subprime loan volumes more than tripled in Westchester County. Nassau County experienced the slowest growth in subprime lending (168 percent), but in part this simply reflects the fact that subprime lenders had made greater inroads into this market by 1994. By 2000, subprime lenders accounted for 35 percent of all refinance loans in Nassau County, 34 percent in Suffolk County, and 27 percent in Westchester County. To some extent these counties reflect the variation in subprime market shares across the 14 metropolitan areas around the state. Nassau-Suffolk ranked fifth among the state s metro areas in subprime lenders share of refinance loans, with higher shares in Newburgh (35 percent), New York (36 percent), Buffalo-Niagara Falls (39 percent), and Jamestown (40 percent). In contrast, Westchester County was among the areas with the lowest share of subprime lending. Only Dutchess County (23 percent), Elmira (26 percent), and Rochester (26 percent) had lower shares. In most market areas around the country, subprime lenders account for much higher shares of loan originations in low-income and, to a greater extent, minority neighborhoods. This pattern is clearly evident in New York State as well. Over the period from 1997 to 2000, when growth in subprime lending had leveled off, subprime lenders accounted for 26 percent of all loans in very low-income neighborhoods compared to 20 percent in low-income neighborhoods, 13 percent in moderate-income neighborhoods, and 9 percent in high-income neighborhoods. The concentration is even more dramatic when refinance loans are considered. In very low-income areas subprime lenders accounted for one out of every two loans across the state, compared to a third in low-income areas, a fifth of loans in moderateincome areas, and 13 percent of loans in high-income areas. Very similar patterns are evident in Nassau and Westchester Counties, with the subprime share increasing sharply as neighborhood incomes decline. However, in Suffolk County there was little variation in the subprime share with neighborhood income. There is also a clear association between the racial composition of a neighborhood and the subprime-lending share. In areas where minorities make up a majority of the population, subprime lenders had come to account for a quarter percent of all mortgage originations in New York State and 46 percent of refinance loans. Among all other neighborhoods the subprime share of loans was about 10 percent of all loans and 19 percent of refinance loans or two and a half times lower than in minority neighborhoods. The statewide racial pattern to subprime lending is also clearly evident in all three of the counties studied. Abt Associates Inc. Executive Summary v
The importance of neighborhood racial composition in determining subprime lending share can be seen when both neighborhood income and racial composition are examined simultaneously. Regardless of income, in minority neighborhoods across New York State subprime lenders have accounted for about one quarter of all loans and more than 40 percent of all refinance loans since 1997. In comparison, very low-income white neighborhoods have had subprime shares of 12 percent for all loans and 20 percent of refinance loans half the level in minority areas. This pattern is also evident in the three counties that are the focus of this study. Residential Foreclosures in Nassau, Suffolk, and Westchester Counties Residential foreclosures have risen rapidly in all three counties in recent years. The growth has been particularly significant in Suffolk and Westchester Counties where foreclosures more than doubled over the three-year period from 1998 to 2001. While foreclosure growth in Nassau County was more moderate, there was still a fairly sizable increase of 35 percent over this period. When foreclosures are measured as a share of the owner-occupied housing stock, Suffolk County stood out as having the highest foreclosure rate. In 2001 there were 2.5 foreclosures for every 1,000 homeowners in Suffolk County, while in Nassau and Westchester Counties there were 1.6 foreclosures for every 1,000 homeowners. In each county, foreclosures by subprime lenders grew more quickly than foreclosures by other lenders. Despite having a lower subprime market share than the other counties, the growth in subprime foreclosures was most pronounced in Westchester County where subprime foreclosures grew two and a half times faster than the increase in foreclosures by other lenders. In Nassau and Suffolk Counties subprime foreclosures increased by about one and a half times the rate among other lenders. Even if subprime loans were no more likely to foreclose than other loans, given the rapid increase in subprime originations, it would be expected that foreclosures of these loans would also increase. In order to determine whether these loans were foreclosing at a higher rate than other loans we compare the share of foreclosures accounted for by subprime lenders with their share of originations. If these lenders account for a greater share of foreclosures it can be interpreted as the greater risk of foreclosure associated with these loans. When subprime lenders share of foreclosures is compared to their share of originations, the three counties exhibit somewhat different patterns. By 2001 subprime lenders in Nassau County were found to have about equal shares of originations and foreclosures, while in Suffolk County their share of foreclosures exceeded their share of originations by about 20 percent. The most pronounced difference was found in Westchester County, where the share of foreclosures by subprime lenders was more than twice their recent share of originations in 2000 and 60 percent higher in 2001. Abt Associates Inc. Executive Summary vi
While foreclosures by subprime lenders have been growing rapidly and have accounted for an increasing share of foreclosures (at least in Suffolk and Westchester), in all three counties about 80 percent of the growth in foreclosures has been accounted for by primarily prime lenders. It is possible that the methodology used to identify subprime foreclosures has not accurately identified subprime lenders. Loans are identified as subprime based on the foreclosing lender s name. However, since prime lenders have become increasingly involved in subprime lending, both by originating loans directly and by acquiring servicing rights to loans originated by subprime lenders, the distinction between prime and subprime lenders may be blurred. One indication that subprime loans may not have been accurately identified is given by the concentration of foreclosures in areas where subprime lenders have been most active. When neighborhoods were characterized by the share of loans accounted for by subprime lenders since 1994, we found that 59 percent of all foreclosures occur in neighborhoods where subprime lenders have been most active. In comparison, only 23 percent of all originations occurred in these areas. Thus, foreclosures were disproportionately located in areas where subprime lenders have been active. In addition, there was little relationship between the subprime lenders share of originations and their share of foreclosures. The share of foreclosures accounted for by subprime lenders was only slightly lower in areas where these lenders had small market shares than in areas where they had high market shares (12 percent versus 17 percent). Areas where subprime lenders had the largest market share are the only areas where subprime lenders did not account for more foreclosures than originations. It seems likely that the rise in subprime lending in these areas must be associated with the rise in foreclosures. Given that subprime lending is most prevalent in minority neighborhoods, not surprisingly a similar pattern is evident when trends in foreclosures are compared across neighborhoods categorized by their income and racial composition. In all three counties, foreclosures have grown much more rapidly in neighborhoods where minorities comprise a majority of the population. Foreclosures increased by 73 percent, 214 percent, and 256 percent, respectively, in Nassau, Suffolk, and Westchester Counties. These high growth rates were also associated with high shares of foreclosures in these areas. Thirty-eight percent of foreclosures in the three counties were in minority neighborhoods, compared to only 10 percent of mortgage originations. Yet, despite the fact that subprime lenders accounted for a large share of originations in minority areas, these lenders did not account for a disproportionate share of foreclosures in these areas. In fact, subprime lenders shares of foreclosures varied little with neighborhood racial composition and income. Interestingly, many of the most active prime lenders, such as Homeside Lending, Washington Mutual, Wells Fargo, LaSalle Bank, and Countrywide, were Abt Associates Inc. Executive Summary vii
found to have very high shares of their foreclosures in minority areas higher even than known subprime lenders. Conclusion This study has found that subprime lending has grown rapidly in New York State since the early 1990s, with particularly high shares of mortgages accounted for by subprime lenders in low-income and minority neighborhoods. These patterns are also evident in the three counties that are the focus of this study. Among the three counties the main differences observed were that subprime lenders had a somewhat lower market share in Westchester County and that there was less variation in subprime market share by income in Suffolk County. However, in all three counties the highest market share of subprime lending has been in minority neighborhoods. Our analysis of trends in residential foreclosures revealed that all three counties experienced significant increases in foreclosures since 1998, with particularly large increases in Suffolk and Westchester Counties. Foreclosures have been particularly concentrated in minority neighborhoods in each county, where the growth in foreclosures has been several times the growth in the remainder of these counties. While it is true that the increase in foreclosures by subprime lenders has been more rapid than by other lenders, most of the growth in foreclosures has been by prime lenders. However, the heavy concentration of foreclosures in neighborhoods where subprime lenders have had high market shares raises questions about the accuracy of the methodology used to identify subprime loans. It seems likely that the high rate of foreclosures in areas where subprime lenders have been most active, particularly minority neighborhoods, was related to the prevalence of subprime lending, but the increased activity by prime lenders as originators and servicers of subprime loans may be blurring this distinction. Further research is needed to determine whether the foreclosures by prime lenders in areas where subprime lenders have been active do, in fact, have their roots in loans originated by subprime lenders. Abt Associates Inc. Executive Summary viii