by Maurice Jourdain-Earl

Transcription

1 The Forec losure Cr isis and Racial Dispar ities in Access to Mor tgage Credit by Maurice Jourdain-Earl February 9, 2011

2 Table of Contents Introduction... 1 Purpose... 2 Methodology... 3 Significance of the Study... 3 Summary of Findings... 6 Clear Differences in the Distribution of Loan Types by Race... 6 African Americans and Latinos have Increasingly Less Access to Prime Loans than Whites... 6 Whites are the Largest Users of Subprime Rate Loans... 7 FHA Benefits Whites more than African Americans and Latinos... 7 Fannie Mae and Freddie Mac Purchased Low Shares of Loans from Borrowers of Color... 8 Trend of CRA eligible Lending Activity Varies by Race... 8 Implications of the Findings... 9 The Correlation Between Defaults and Foreclosures and Loan Type Loan Type Differences by Race Impact on Cost of Credit and Loan Performance The Racial Disparities in Access to Mortgage Credit Disparities in the Dollar Volume of Prime Rate Loans The Number of Prime Loans Decreased Disproportionately The Market Share Trend of Prime Rate Loans Favors Whites Prime Loan Origination Rate Disparities Subprime Rate Loan Disparities Disproportionate Drop in Dollar Volume of Subprime Rate Loans by Race Racial Disparities in the Number of Subprime Rate Loans Market Share Changes of Subprime Rate Loans by Race Disparities in the Incidence and Magnitude of Subprime Rate Loans i

3 Disparities in Access to FHA Loans Dramatic Increase in the Dollar Volume and Number of FHA Loans Market Share of FHA Loans FHA Loan Origination Disparities Low Levels of Loan Diversity at Fannie Mae and Freddie Mac CRA eligible Lending Favors non Hispanic Whites Conclusion Recommendations About the Author About COMPLIANCETECH Table of Figures Figure 1: Six Year Average Loan Type Distribution by Race and Ethnicity Figure 2: Detailed Loan Type Distribution by Race and Ethnicity Figure 3: Loan Type Summary by Race and Ethnicity Figure 4: Subprime Rate Loans by Census Tract Percent Minority Figure 5: Distribution of Prime Loans Figure 6: Prime Loans Race Trend Figure 7: Prime Loan Origination Rates by Race Figure 8: Prime Loan Origination Disparity Index Figure 9: Race Distribution of Subprime Loans Figure 10: Subprime Race Trend Figure 11: Percent of Subprime by Race and Ethnicity Figure 12: Incidence of Subprime Disparities (SIDI) Figure 13: Race Distribution of FHA Loans Figure 14: FHA Race Trend Figure 15: FHA Loan Origination Rates Figure 16: FHA Loan Disparity Index ii

4 Figure 17: Fannie Mae and Freddie Mac Loan Purchase Activity by Race Index of Tables Table 1: Percent Default and Foreclosure (August 1, 2010) Table 2: Dollar Volume of Prime Loans 2004 to Table 3: Prime Loan /Decline by Race Table 4: Market Share /Decline of Prime Loans by Race Table 5: Prime Loan Origination Disparities Table 6: Dollar Volume of Subprime Rate Loans Table 7: Top 15 MSAs with non Hispanic white Subprime Rate Loans Table 8: Subprime Loan /Decline by Race Table 9: Market Share of Subprime Loans by Race /Decline Table 10: 2004 Subprime Disparities by Race and Ethnicity Table 11: 2005 Subprime Disparities by Race and Ethnicity Table 12: 2006 Subprime Disparities by Race and Ethnicity Table 13: 2007 Subprime Disparities by Race and Ethnicity Table 14: 2008 Subprime Disparities by Race and Ethnicity Table 15: 2009 Subprime Disparities by Race and Ethnicity Table 16: FHA Summary by Dollar Volume Table 17: FHA Loan /Decline by Race Table 18: Market Share /Decline of FHA by Race Table 19: Fannie Mae and Freddie Mac Ratio of Loans ( ) to Population Table 20: Fannie Mae and Freddie Mac Loan Purchase Activity by Race ( ) Table 21: Annual Percent Change of Fannie Mae and Freddie Mac Loan Purchase Activity by Race ( ) Table 22: CRA Eligible Loans ( ) Relative to US Population (2000 Census) Table 23: CRA eligible Loans by Race ( ) Table 24: Annual Percent Change of CRA eligible Loans by Race ( ) iii

5 INTRODUCTION This study, using Home Mortgage Disclosure Act (HMDA) 1 data from , shows evidence of how access to mortgage credit for African Americans and Latinos has plummeted compared to non Hispanic whites. The disparities exist with all loan types 2 ; prime, subprime 3, and FHA and by all measures; the number, dollar volume and share of the mortgage market. Furthermore, the racial disparities exist with lenders that originated loans sold to Fannie Mae and Freddie Mac and loans eligible for Community Reinvestment Act (CRA) 4 credit. 1 Congress enacted HMDA in 1975 to: provide the public with information to judge whether lenders are serving their communities; to enhance enforcement of laws prohibiting discrimination in lending; and to provide private investors and public agencies with information to guide investments in housing. HMDA requires mortgage lenders located in metropolitan areas to collect data on their mortgage application transactions, report the data annually to the government, and make the data publicly available 2 Loans are limited to Conventional 1 st lien, owneroccupied, home purchase and refinance transactions on 1 4 family properties. 3 Subprime or higher cost loans are purportedly made to borrowers with less than pristine credit. HMDA data defines such loans as 1 st lien loans with an APR greater than 3 percentage points above a comparable maturity treasury security; 2 nd liens 5 percentage points. 4 The Community Reinvestment Act is intended to encourage depository institutions to help meet the credit needs of the communities in which they operate, including low and moderate income neighborhoods, consistent with safe and sound banking operations. It was enacted by the 1 Focused only on loans, this study analyzes origination disparities by race and loan type. The study does not attempt to explore the myriad of reasons why loan outcomes are different. Outcomes are different due to reasons as varied as a lower number of applications submitted, or higher denial and fallout rates based on credit, collateral, income and/or employment. Denial and fallout rate differences across racial and ethnic groups persist, although the HMDA data does not include sufficient information to determine the extent to which these differences stem from illegal discrimination or other factors. 5 Ultimately, regardless of the reason for the loan outcome differences, either a loan is granted or it is not. This study seeks to measure the disparities in access to credit by race and describe how the disparities are having adverse effects on borrowers of color and the communities in which they live. Today, African Americans and Latinos have more difficulty financing a home, whether they are first time home buyers or homeowners looking to refinance an existing mortgage. Moreover, access to credit is further restricted if the borrowers Congress in 1977 (12 U.S.C. 2901) and is implemented by Regulations 12 CFR parts 25, 228, 345, and 563e. 5 The 2009 HMDA Data: The Mortgage Market in a Time of Low Interest Rates and Economic Distress by Robert B. Avery, Neil Bhutta, Kenneth P. Brevoort, and Glenn B. Canner of the Federal Reserve Board s Division of Research and Statistics, pg. 3.

6 are looking to buy or refinance in census tracts with high foreclosure activity 6. Many assume the foreclosure crisis was caused by government polices to expand homeownership opportunities to borrowers of color and low and moderate income (LMI) 7 borrowers in LMI communities. The policies in question are the fair lending laws, CRA and Fannie Mae s and Freddie Mac s affordable housing goals. 8 Those offering this viewpoint cannot offer any hard data to support this position. Moreover, it is unclear whether (or to what extent) the above assumption influences housing 6 Ibid, Lending activity in census tracts with high foreclosure activity has declined more than in other neighborhoods and the decline has been particularly severe for refinancing activity pg Low Income: individuals and geographies having a median family income less than 50 percent of the area median income. Moderate Income: individuals and geographies having a median family income of at least 50 percent and less than 80 percent of the area median income. 8 The Federal Housing Finance Agency (FHFA) established new housing goals for Fannie Mae and Freddie Mac (the Enterprises) for The Housing and Economic Recovery Act of 2008 (HERA) required FHFA to establish housing goals for the Enterprises for targeted segments of the mortgage market. The final rule establishes three single family, owner occupied home purchase mortgage goals for low income families, very lowincome families, and families living in geographical areas with lower income populations, areas with high concentrations of minority residents, and federally declared disaster areas. The latter goal also includes a specialized sub goal to ensure that the Enterprises address housing needs in lowerincome and minority areas. The final rule also contains a goal for single family, owner occupied refinance mortgages for low income families. policy makers and lenders marketing and credit allocation practices. PURPOSE The purpose of this report is to shine a bright light on the facts regarding credit access by: Providing empirical evidence of the disparities in mortgage credit availability for African Americans and Latinos. Unbundling the erroneous perception that the root cause of the financial crisis was unsustainable lending to African Americans and Latinos by showing how whites are also adversely impacted by the subprime crisis. The report provides data that shows how whites received the highest number and dollar volume of subprime rate loans and are likely to have more loans in foreclosure, thereby having a greater impact on the collapse of the market. Showing how FHA, in its efforts to rescue the market and manage risk, grew tremendously in 2008 and 2009, but the growth was more with whites (and Asians) than with African Americans and Latinos. By increasing loan limits and raising credit requirements, it appears FHA may be moving away from its congressionally mandated mission of supporting affordable housing. Illustrating how lenders, Fannie Mae, and Freddie Mac underserved borrowers of color contrary to the 2

7 popular notion that they somehow over served these segments. METHODOLOGY HMDA data from 2004 through 2009 were examined for evidence of the disparities in credit availability for African Americans and Latinos. The data were stratified by various dimensions and analyzed to show the year over year percent change in the following: 1. The growth or decline in the dollar volume, number and market share of prime, subprime and FHA loans by race; 2. Racial disparities in the origination rates of prime and FHA loans, 3. The racial distribution of loans purchased by Fannie Mae and Freddie Mac, and 4. The racial distribution of CRAeligible LMI loans. Disparity ratios are used to measure and compare the outcomes of African Americans and Latinos to non Hispanic whites. Such disparities are comparisons of origination rates of non Hispanic whites to these two minority groups. SIGNIFICANCE OF THE STUDY Analyzing the issue of access to mortgage credit by race is significant because of the central role homeownership plays in building personal wealth. Homeownership provides many benefits to the family, children and the community, such as increased access to quality 3 education, lower crime rates, health benefits and increased family and neighborhood stability. 9 Most Americans accumulate assets through homeownership. Homeownership provides multiple opportunities for households to save and accumulate wealth. Before buying a house, households increase savings to accumulate a down payment. Once they become home owners they are forced to save through amortization of their mortgages. As homeowners repay their mortgages, their debt burden declines while housing equity and overall wealth rises. Some home owners further boost their savings as they plan for maintenance and upgrading projects. Finally, housing appreciation offers another means for wealth accumulation by delivering tax exempt capital gains when house prices increase. Of course, as with any other assets, home prices are not guaranteed to go up all the time, but historically national home price measures did and on average rose faster than inflation. 10 Home equity accounts for roughly 44 percent of total measured net worth 11. Consequently, the impact of diminished homeownership due to limited access to 9 National Association of Homebuilders, Homeownership, Financial Flexibility, and Wealth, Housing Policy, July 8, 2005by Natalia Siniavskaia, Ph.D. &genericcontentid=42728&print=true 10 Ibid. 11 See The Hidden Cost of Being African American, by Thomas M. Shapiro, pg. 107

8 credit and increased foreclosures for African Americans and Latinos is huge. According to the testimony of Federal Reserve Board (FRB) Governor Elizabeth Duke, the number of foreclosures initiated on residential properties has soared from about 1 million in 2006, the year that house prices peaked, to 2.8 million in The FRB saw a further 1.2 million foreclosure filings and an additional 2.4 million homes in the foreclosure pipeline at the end of June All told, the FRB expects about 2.25 million foreclosure filings in 2010 and again in 2011, and about 2 million more in In November 2010 the FRB indicated that 5 million mortgage loans were 90 days or more past due or in foreclosure 12. More recently, on January 13, 2011 RealtyTrac released its Year End 2010 U.S. Foreclosure Market Report, which shows a total of 3,825,637 foreclosure filings default notices, scheduled auctions and bank repossessions were reported on a record 2,871,891 U.S. properties in 2010, an increase of nearly 2 percent from 2009 and an increase of 23 percent from The report also shows that 2.23 percent of all U.S. housing units (one in 45) received at least one foreclosure filing during the year, up from 2.21 percent in 2009, 1.84 percent in 2008, 12 Governor Elizabeth A. Duke Foreclosure documentation issues before the Financial Services Subcommittee on Housing and Community Opportunity, U.S. House of Representatives, Washington, D.C., November 18, mony/duke a.htm 1.03 percent in 2007 and 0.58 percent in While very little research has been conducted on the racial distribution of foreclosures, the Center for Responsible Lending (CRL) released a study in June 2010 entitled Foreclosures by Race and Ethnicity: The Demographics of a Crisis. 14 The authors estimated that 2.5 million foreclosures were completed. The vast majority of these foreclosures were on owner occupied properties with mortgages that were originated between 2005 and Major findings in the report include the following: The majority (an estimated 56%) of families who lost homes were non Hispanic and white, but African American and Latino families were disproportionately affected relative to their share of mortgage originations. Among recent borrowers, CRL estimates that nearly 8% of both African Americans and Latinos have lost their homes to foreclosures, compared to 4.5% of whites. The racial and ethnic disparities in these estimated foreclosure rates hold million us properties receiveforeclosure filings in 2010 despite 30 month lowin december Foreclosures by Race and Ethnicity: The Demographics of a Crisis, CRL Research Report, Debbie Gruenstein Bocian, Wei Li, and Keith S. Ernst, June 18,

9 even after controlling for differences in income patterns between demographic groups. The CRL report also stated that even if foreclosures represented nothing more than a one time cost only to the families involved, these findings would be troubling. But the costs are extensive, multifaceted and long term, extending far beyond individual families to their neighbors, communities, cities and states. As the foreclosure crisis threatens the financial stability and mobility of families across the country, it will be particularly devastating to African American and Latino families, who already lag behind their white counterparts in terms of income, wealth and educational attainment. Furthermore, the indirect losses in wealth that result from foreclosures as a result of depreciation to nearby properties will disproportionately impact communities of color. CRL estimates that, between 2009 and 2012, $194 and $177 billion, respectively, will have been drained from African American and Latino communities in these indirect spillover losses alone. The significance of this study is that it is supported by data. It measures the access to credit issue by reporting on the six year change in the number, dollar volume and market share of loans by race and the differences in origination rates. The disparity in credit access has a direct impact on wealth and opportunities for wealth accumulation. Unless the wealth gap is addressed it will be a harbinger of even greater racial inequities to come. Thomas M. Shapiro in his book The Hidden Cost of Being African American discusses how the asset gap or wealth gap 5 continues to perpetuate inequities. He states that wealth has been a neglected dimension of the social sciences concern with the economic and social status of Americans in general and racial minorities in particular. 15 He further states that we have been much more comfortable describing and analyzing occupational, educational, and income inequality than examining the economic foundation of a capitalistic society, private property. 16 Shapiro argues that wealth motivates much of what Americans do, is the grounds for life changes, and provides enduring advantages and disadvantages across generations The Hidden Cost of Being African American, by Thomas M. Shapiro, pg Ibid 17 Ibid

10 SUMMARY OF FINDINGS The study provides clear insight into the duality of the mortgage market; one where African Americans and Latinos have lower origination rates and when they do get loans they disproportionately receive higher cost loans. These higher cost subprime rate loans are more prone to default and foreclosure. However, non Hispanic whites had the largest number and dollar volume of subprime rate loans and thus are likely to have more loans in default and foreclosure. CLEAR DIFFERENCES IN THE DISTRIBUTION OF LOAN TYPES BY RACE Between 2004 and 2009, white loans were disproportionately prime rate loans at 77.95% of total white loans; subprime, 12.47%; and FHA, 9.58%. By contrast, the distribution of African American loans was prime, 48.53%; subprime, 34.96%; and FHA, 16.51%. The distribution of Latino loans was prime, 62.12%; subprime, 26.17%; and FHA, 11.71%. Asians had the highest use of prime mortgage credit at 87.12%, and the lowest use of subprime and FHA at 8.84% and 4.04%, respectively. AFRICAN AMERICANS AND LATINOS HAVE INCREASINGLY LESS ACCESS TO PRIME LOANS THAN WHITES Between 2004 and 2009 the number of white prime loans decreased 30.99% while prime loans made to African Americans and Latinos decreased 75.57% and 75.55%, respectively. Prime loans to Asians decreased 28.29%, slightly less than whites. 6 All minority groups combined received substantially fewer prime rate loans compared to whites. Whites obtained 25.4 million prime rate loans between 2004 and 2009; Latinos, 3.1 million; Asians, 1.7 million; and African Americans, 1.6 million. Only whites and Asians had positive prime loan market share growth between 2004 and 2009 at 12.54% and 19.60%, respectively. By contrast, the market share of prime rate loans for African Americans and Latinos decreased 62.67% and 61.62%, respectively. The prime loan ODI 18 shows evidence of prime mortgage credit disparities with African Americans and Latinos. Between 2004 and 2009 the average African American prime loan ODI was In other words, whites were two times or 201% more likely to receive 18 The Origination Disparity Index (ODI ) is a function of the origination rate. It is a ratio of the White origination rate divided by a minority groups origination rate. It is a measure of the likelihood of a minority application being originated relative to the likelihood of a White application being originated. For example, if 85 % of White applications result in a loan while 72% of Hispanic applications result in a loan, then the Hispanic ODI is 85% / 72% = Taken at face value the ratio in this example suggests that White applicants are 1.18 times more likely to be originated than Hispanic applicants. An ODI of 1.00 indicates parity. An ODI greater than 1.00 indicates a disparity. An ODI less than 1.00 are referred to as negative disparity and suggest (on its face) that a minority group faired better than Whites.

11 prime rate loans than African Americans. The average Latino ODI was The Asian ODI was indicating virtually no disparity. WHITES ARE THE LARGEST USERS OF SUBPRIME RATE LOANS Between 2004 and 2009, whites received 4.1 million subprime rate loans, followed by Latinos, 1.3 million; African Americans, 1.2 million; and Asians, 179 thousand. In each year after 2005, subprime rate lending had negative growth rates overall for each racial group. The drop in the number of subprime rate loans was greatest for African Americans and Latinos. For example, in 2004, African Americans received 218,665 subprime rate loans. By 2009, the number was down to 10,416; a decrease of 95.24%. A similar pattern occurred with Latinos. In 2004, Latinos received 190,335 subprime rate loans and was down to 14,774 by 2009, a decrease of 92.24%. The number of white subprime rate loans decreased from 727,790 in 2004 to 141,545 in 2009, a decrease of 80.55%. Finally, the number of Asian subprime rate loans decreased 86.62% from 25,033 in 2004 to 3,350 in share of subprime loans decreased significantly between 2004 and 2009 at 69.03% and 49.54%, respectively. The Asian share of the subprime market was negligible, reaching its peak of 2.79% in 2006 and its low point of 1.79% in Between 2004 and 2009 the Asian market share of the subprime market declined 13.15%. Whites received the largest dollar volume of subprime rate loans in each year under review ( ). Latinos received the second largest dollar volume and number of subprime rate loans followed by African Americans and Asians. African Americans had the highest Subprime Incidence Disparity Index (SIDI ) at 3.68 in This indicates that African Americans received subprime rate loans 3.68 times the rate of whites. Over the six year period of this study African Americans on average received subprime rate loans 2.38 times the rate of whites or 238% more often. The average Latino SIDI was 1.95 indicating they received subprime rate loans 195% more often than non Hispanic whites. FHA BENEFITS WHITES MORE THAN AFRICAN AMERICANS AND LATINOS The white market share of subprime rate loans stood at 61.80% in 2004; by 2009 it reached 81.71%, a 32.22% increase. Whites were the only racial group to have positive growth in the market share of subprime rate loans from 2004 to For African Americans and Latinos the market 7 Whites had the largest dollar volume and number of FHA loans in each period with a low of 227,445 FHA loans in 2006 to a high of 1,225,605 in Between 2004 and 2009 the white number of FHA loans grew %, compared to African Americans at 80.25% and Latinos at %,

12 respectively. Between 2004 and 2009, Asians had the highest FHA growth rate at %. Asians, who have traditionally been limited users of FHA, had their highest number of FHA loans in 2009 at 42,380. Whites increased their market share of FHA loans from a low of 64.43% in 2004 to a high of 74.05% in 2009, a 14.94% growth rate. During the same period the African American percent change with FHA loans decreased 38.71%. The percent change of Latino FHA loans decreased 25.48%. Asians had a significant market share growth with FHA loans at 85.56% but with a relatively low number of loans. The data reveals that African Americans had the highest FHA origination rate disparities with an average ODI of 1.19 or a 19% higher origination rate. The average Latino FHA ODI was The average Asian FHA ODI was FANNIE MAE AND FREDDIE MAC PURCHASED LOW SHARES OF LOANS FROM BORROWERS OF COLOR Fannie Mae and Freddie Mac (Government Sponsored/Owned Enterprises [GSEs]) do not make loans directly. Instead, they buy and securitize loans originated to their underwriting specifications from approved lenders. According to the HMDA data, together these GSEs purchased 14,271,173 loans from lenders between 2004 and Non Hispanic whites were the primary beneficiaries of Fannie Mae and 8 Freddie Mac s secondary market activities. Despite the presence of government mandated affordable housing goals, on average between , more than 81% of loans purchased by the two mortgage giants went to non Hispanic whites. African American s accounted for only 4.77% of loans purchased; Latinos, 8.14%; and Asians, 5.40%. With respect to the number of loans purchased by the GSEs between 2004 and 2009, African American loans decreased from 152,207 in 2004 to 66,791 in 2009 or 56.12%. Latino loans purchased decreased from 268,884 in 2004 to 128,984 in 2009 or 52.03%. White loan purchases decreased slightly from 2,554,128 in 2004 to 2,440,653 in 2009 or 4.44%. Only Asian loans purchased by the GSE s had positive growth from 158,937 in 2004 to 197,624 in 2009, an increase of 24.34%. TREND OF CRA ELIGIBLE LENDING ACTIVITY VARIES BY RACE CRA eligible loans made either to LMI borrowers or on homes located in LMI census tracts totaled 15,177,002 between 2004 and These loans were distributed by race as follows: white, 68.24%; African American, 12.49%; Latinos, 15.03%; and Asian, 3.47%. CRA eligible lending activity decreased for all racial groups between 2004 and 2009, but more with African Americans and Latinos. White and Asian CRAeligible loan originations decreased

13 29.32% and 23.99%, respectively. By contrast, CRA eligible loans to African Americans and Latinos decreased 64.65% and 63.48%, respectively. IMPLICATIONS OF THE FINDINGS On one hand, the count and volume of subprime loans tells us something about which racial groups benefitted most from subprime financing. On the other hand, the racial concentration of subprime rate loans tells us something about the likely effect on racial groups. Because of the disproportionate number of subprime rate loans made to African Americans and Latinos, a larger proportion of their loans are likely to result in defaults and foreclosures. Further, because of the high concentration of subprime rate loans in communities of color, those communities are likely to be hit the hardest. However, non Hispanic whites had the largest number of subprime rate loans and consequently they are likely to experience the greatest number of defaults and foreclosures. Currently there is no public data available to ascertain the actual racial distribution of defaults and foreclosures. One can only estimate this based on the distribution of subprime rate loans by race and other contributing factors, such as unemployment statistics. The debate points to a drastic need for the public disclosure of foreclosure data by race and other demographic criteria such as borrower income, income of census tracts and the distribution by communities of color. Foreclosure management and fair lending analysis would be greatly improved with this type of granular information. The effect of defaults and foreclosures and diminished access to credit is already evidenced by lower homeownership rates. According to the U.S. Census Bureau News 19 the African American and Latino homeownership rates decreased more than whites. From the 1 st quarter 2007 to the 1 st quarter 2010 the African American homeownership rate dropped from 48.0% to 45.6%, Latinos dropped from 50.1% to 48.5%, while whites dropped from 75.3% to 74.5% qtr110/files/q110press.pdf, April 26, 2010, pg. 7 9

14 THE CORRELATION BETWEEN DEFAULTS AND FORECLOSURES AND LOAN TYPE Theoretically, there is an inverse relationship between the cost of credit and the risk of loss due to default. Prime rate loans cost less and have a lower expected risk of loss due to default and foreclosure. On the other hand, subprime rate loans cost more and have a higher expected risk of loss due to default and foreclosure. FHA loans are priced somewhere in between the two. A review of data published by the Federal Reserve Bank of New York, First American CoreLogic and Lender Processing Service (LPS) reveals wide disparities in the default and foreclosure rates by loan type (Table 1). TABLE 1: PERCENT DEFAULT AND FORECLOSURE (AUGUST 1, 2010) Table 1 reveals that subprime rate loans are 5.75 and 4.60 times more likely to be 90 days past due than prime and government loans and 4.41 and 5.33 times more likely to be in foreclosure than prime or government loans, respectively. For this reason, particular emphasis is placed on the number and proportion of subprime rate loans. The fact that subprime rate loans are more likely to go into default and foreclosure supports one of the main theses of this report: Those racial groups with the largest absolute number of subprime rate loans are likely to have a larger number of mortgage defaults and foreclosures and a larger impact on the economy. On the other hand, the racial groups that disproportionately used subprime were most affected by this concentration, but due to the relatively low level of loans, had a lesser impact on the economy. Loan Type Percent with a payment 90+ days past due Percent in foreclosure Prime* Subprime** 18.40% 12.80% Government*** *Loans which have first lien rights with high credit score, less risky and with full required supported documents (A Paper). Those who qualify these requirements come under prime loan. ** A type of loan that is offered at a rate above prime to individuals who do not qualify for prime rate loans. Quite often, subprime borrowers are often turned away from traditional lenders because of their low credit ratings or other factors that suggest that they have a reasonable chance of defaulting on the debt repayment. ***Loans insured or guaranteed by the Federal Housing Administration or the Veterans Administration. Also included are loans insured by the Farmers Home Administration/Rural Housing Service. 10

15 LOAN TYPE DIFFERENCES BY RACE IMPACT ON COST OF CREDIT AND LOAN PERFORMANCE In effect, differences in loan type distributions by race create a dual mortgage market. A market where on one hand, non Hispanic whites and Asians have significantly higher proportions of low cost prime rate mortgage loans, lower proportions of subprime and even lower proportions of FHA. On the other hand, a market where loans to African Americans and Latinos are disproportionately less represented in low cost prime rate loans and over represented in high cost subprime rate loans and FHA. Figure 1 shows the average percent of loan type distributions from 2004 to 2009 for whites, African Americans, Latinos and Asians. Figure 1 reveals that Asians had higher average levels of prime rate loans (87.12%) and lower average levels of subprime (8.84%) and FHA loans (4.04%) than whites. The average levels for whites were prime 77.95%; subprime 12.47%; and FHA 9.58%. By contrast, the average levels of Latino loans were prime 62.12%; subprime 26.17%; and FHA 11.71%. African Americans had the lowest average proportion of prime rate loans at 48.53%; the highest average proportion of subprime rate loans at 34.96%; and the highest percent of FHA loans at 16.51%. FIGURE 1: SIX YEAR AVERAGE LOAN TYPE DISTRIBUTION BY RACE AND ETHNICITY 11

16 Figure 2 includes the number and percent of prime, subprime and FHA loans for each racial group between 2004 and Two patterns stand out. First, is the low and decreasing percentage of prime rate loans for African Americans. Second, is the increasing number of subprime rate loans made to Latinos. Both are indicative of how African Americans and Latinos have different experiences with mortgage finance than whites and Asians. The data confirms that African Americans and Latinos pay more for mortgage credit and have a disproportionate share of loan types that are more likely to result in defaults and foreclosures. 12

17 FIGURE 2: DETAILED LOAN TYPE DISTRIBUTION BY RACE AND ETHNICITY Loan Type Distribution by Race and Ethnicity White 2004 % 2005 % 2006 % 2007 % 2008 % 2009 % Total % Prime 6,018, % 5,167, % 3,982, % 3,580, % 2,623, % 4,011, % 25,384, % Subprime 727, % 1,227, % 1,108, % 609, % 244, % 141, % 4,059, % FHA 362, % 252, % 227, % 267, % 784, % 1,225, % 3,120, % 7,110, % 6,650, % 5,320, % 4,459, % 3,654, % 5,380, % 32,564, % African Am % 2005 % 2006 % 2007 % 2008 % 2009 % Total % Prime 456, % 371, % 338, % 287, % 122, % 101, % 1,677, % Subprime 218, % 385, % 378, % 176, % 38, % 10, % 1,208, % FHA 91, % 58, % 49, % 65, % 142, % 164, % 570, % 766, % 815, % 766, % 530, % 302, % 275, % 3,456, % Latino 2004 % 2005 % 2006 % 2007 % 2008 % 2009 % Total % Prime 858, % 742, % 639, % 490, % 222, % 195, % 3,149, % Subprime 190, % 451, % 445, % 182, % 41, % 14, % 1,326, % FHA 95, % 54, % 42, % 48, % 142, % 209, % 593, % 1,146, % 1,250, % 1,129, % 724, % 408, % 421, % 5,069, % Asian 2004 % 2005 % 2006 % 2007 % 2008 % 2009 % Total % Prime 426, % 347, % 266, % 235, % 188, % 302, % 1,767, % Subprime 25, % 64, % 56, % 24, % 6, % 3, % 179, % FHA 7, % 4, % 3, % 4, % 19, % 42, % 81, % 461, % 418, % 327, % 266, % 215, % 350, % 2,029, % 13

18 Figure 3 is a graph that illustrates the different distribution patterns by race and loan types. The graphs reveal that African Americans were the only group where levels of subprime rate loans actually superseded levels of prime rate loans. This occurred in 2005 and 2006 when the percent of subprime rate loans was greater than the percentage of prime rate loans. Latinos have a similar pattern, but only African Americans had a higher level of subprime rate loans compared to prime, during the six year period of this study. FIGURE 3: LOAN TYPE SUMMARY BY RACE AND ETHNICITY White Loan Type Summary '04 '09 Black Loan Type Summary '04 '09 7,000, ,000 6,000,000 5,000, , , ,000 4,000,000 3,000,000 2,000,000 Prime Subprime FHA 300, , , ,000 Prime Subprime FHA 1,000, ,000 50, Latino Loan Type Summary '04 '09 1,000, , , , ,000 Prime 500,000 Subprime 400,000 FHA 300, , , , , , , ,000 Asian Loan Type Summary '04 '09 Prime Subprime FHA 200, , ,000 50, Source: HMDA Data The data provides clear evidence of a dual lending market. Based upon the type of loans received, African Americans and Latinos pay more for mortgage credit. Because of their collective vulnerability, everyone in the loan origination food chain profited from the higher prices paid for loans by African Americans and Latinos; from loan originators, lenders, Wall Street investment bankers and investors in 14

19 mortgage backed securities. The higher cost for mortgage credit translates into less money for basic necessities. The higher cost for mortgage credit also translates into African Americans and Latinos having lower homeownership rates and lower opportunities to build wealth, lower educational achievement and higher unemployment. Consequently, it is not surprising that they are disproportionately experiencing higher levels of defaults and foreclosures and greater disparities in access to credit. The foreclosure activity is especially high in high minority neighborhoods as the distribution of subprime rate loans were disproportionately originated in communities of color. Figure 4 below shows the distribution of subprime rate loans for each racial group by census tracts and communities of color. Each graph is very revealing. For whites, it shows that subprime rate loans were concentrated in predominately white census tracts, i.e., less than 30% minority. The pattern with African Americans was just the opposite. The majority of subprime rate loans to African Americans were in census tracts with high proportions of persons of color, i.e % minority. The Latino pattern was similar to African Americans, but not as dramatically skewed towards census tracts with high concentrations of persons of color. Finally, the Asian pattern was similar to whites, i.e. higher distributions in predominately white census tracts. Consequently, the communities of color with high concentrations of subprime rate loans are disproportionately impacted. 15

20 Figure 4: Subprime Rate Loans by Census Tract Percent Minority 16

21 THE RACIAL DISPARITIES IN ACCESS TO MORTGAGE CREDIT DISPARITIES IN THE DOLLAR VOLUME OF PRIME RATE LOANS Every racial group experienced significant declines in the volume of prime loans; however, the greatest declines occurred with African Americans and Latinos. Between 2004 and 2009 the dollar volume of prime loans made to African Americans and Latinos declined 76.24% and 76.45% respectively. African American prime loan volume dropped from $82 billion in 2004 to $24.7 billion in 2008 and dropped further to $19.5 billion in Latino prime loan volume decreased from $171 billion in 2004 to $47.3 billion in 2008, and decreased to $40.2 in African Americans and Latinos are the only groups whose prime loan volume dropped from 2008 to 2009 loans. However, Latinos experienced the greatest overall percentage decrease in the dollar volume of prime mortgage loans. Between 2004 and 2009, the dollar volume of Latino prime loans decreased 76.45%. The decreases in prime credit availability is indicative of how lenders focused on lending opportunities to African Americans and Latinos in the earlier time period of this study. But now, lending to the African American and Latino market segments is occurring less frequently, more evidence of the diminished access to prime mortgage credit. During the same period ( ) white and Asian prime loan volume decreased 26.71% and 25.11%, respectively. White prime loan volume decreased from $1.2 trillion in 2004 to $595 billion in 2008, but increased to $876 billion in In 2004, the prime loan volume for Asians was $121 billion; decreased to $60 billion by 2008 and increased to $90.6 billion in 2009 (Table 2). With regards to disparities in the dollar volume of prime credit availability, African Americans appear to be hit the hardest. Compared to whites, Latinos and Asians, African Americans received the smallest dollar volume of prime mortgage 17

22 TABLE 2: DOLLAR VOLUME OF PRIME LOANS 2004 TO 2009 Race % +/ % +/ % +/ % +/ % +/ White $1,195,047,404 $1,155,199, % $933,254, % $843,684, % $594,804, % $875,802, % African Am $81,970,528 $76,615, % $74,584, % $61,398, % $24,699, % $19,479, % Latino $170,667,596 $172,997, % $163,524, % $120,192, % $47,322, % $40,188, % Asian $121,080,536 $113,667, % $92,516, % $82,620, % $60,119, % $90,672, % Native Am $5,512,028 $4,328, % $3,456, % $2,940, % $1,654, % $2,091, % Hawaiian $9,172,531 $9,728, % $8,131, % $6,118, % $2,837, % $3,135, % Multi Race $3,937,582 $4,172, % $3,681, % $3,357, % $2,262, % $2,982, % Total $1,814,108,095 $1,735,559, % $1,279,150, % $1,274,035, % $835,009, % $1,034,352, % 18

23 THE NUMBER OF PRIME LOANS DECREASED DISPROPORTIONATELY The number of prime loans originated decreased disproportionately by race. White prime loans dropped from 6 million in 2004 to 2.6 million in 2008, but increased to 4.0 million units in 2009 or 52.92%. Between 2004 and 2009, the number of white prime rate loans decreased 33.34%. By comparison, African American prime rate loans dropped from 456,713 in 2004 to 122,032 in 2008 and dropped further to 101,168 units in 2009 or by 17.10%. This adds up to a six year African American decrease of 77.85%. Latino prime rate loans decreased from 858,339 in 2004 to 222,662 in 2008 and decreased further to 195,062 in This was the second largest overall percentage decrease in the dollar volume of prime rate loans or a drop of 77.27%. Asians had the lowest overall percentage decrease in the number of prime rate loans between 2004 and 2009 at 29.07%. The Asian prime loan volume decreased from 426,962 units in 2004 to 188,448 in 2008 and (like whites) increased to 302,834 units or 60.70% in 2009 (Figure 5 and Table 3). Figure 5: Distribution of Prime Loans

24 TABLE 3: PRIME LOAN GROWTH/DECLINE BY RACE Year % % % % % % 04/09 White 6,018,339 5,167, % 3,982, % 3,580, % 2,623, % 4,011, % 33.34% African Am 456, , % 338, % 287, % 122, % 101, % 77.85% Latino 858, , % 639, % 490, % 222, % 195, % 77.27% Asian 426, , % 266, % 235, % 188, % 302, % 29.07% Native Am. 28,838 19, % 15, % 13, % 8, % 10, % 62.54% Hawaiian 36,664 32, % 25, % 19, % 10, % 11, % 67.50% Total 7,827,859 6,684, % 5,269, % 4,629, % 3,176, % 4,635, % 40.78% Source: 20

25 THE MARKET SHARE TREND OF PRIME RATE LOANS FAVORS WHITES Whites maintained an increasing share of prime rate loans throughout the time period and experienced positive percentage gains in every year except 2005 to 2006 with a small decrease of 2.22%. Whites reached their highest market share of prime loans in 2009 at 86.54%; the lowest in 2006 at 75.61%. Asians had the lowest market share of prime rate loans among all racial groups, but like whites had positive market share growth of 19.60% between 2004 and 2009 (Figure 6 and Table 4). By comparison, the prime loan market share for African Americans never exceeded 6.43% (2006) and the lowest share in 2009 of 2.18%. Latinos had double digit market shares in every year until Latinos had their highest prime market share of 12.14% in 2006, their lowest in 2009 at 4.21%. By 2009 the bottom fell out of the prime market for African Americans and Latinos. Between 2008 and 2009 their share of the market decreased 43.23% and 39.94%, respectively. The Asian share of the prime loan market ranged from a low of 5.1% in 2007 to a high of 6.53% in 2009, exceeding the market share of African Americans and Latinos by 4.35% and 2.32%, respectively. The number of Asian prime rate loans exceeded the number made to African American in both 2008 and 2009 and exceeded the number made to Latinos in Figure 6: Prime Loans Race Trend

26 Table 4: Market Share /Decline of Prime Loans by Race Share /Year % % % % % % 04/09 White % % % % % 12.54% African Am % % % % % 62.67% Latino % % % % % 61.62% Asian % % % % % 19.60% Native Am % % % % % 37.84% Hawaiian % % % % % 44.68% Source: 22

27 PRIME LOAN ORIGINATION RATE DISPARITIES Further evidence of the disparities in prime mortgage credit availability is revealed through an analysis of prime mortgage origination rates and the corresponding origination rate disparity index (ODI ) by race. Table 5 shows that the white and Asian origination rates had a consistently similar pattern between 2004 and Both averaged prime mortgage origination rates of 54+ percent. Between 2004 and 2008, the average white prime origination rate was in a tight and narrow range; a high of 56.85% in 2004 to a low of 49.69% in 2006, but increased significantly in 2009 to 64.50%. slightly to 50.34% in 2008 and increased significantly in 2009 to 60.38%. Only white and Asian origination rates exceeded 60% in By contrast, the African American and Latino prime origination rates both had significant year over year decreases until In 2009, the African American and Latino prime origination rates tipped upwards to 35.30% and 43.34%, respectively. While both African Americans and Latinos had lower prime origination rates, African Americans fared the worst. The highest Latino prime origination rate was 47.23% in 2004 or 14.97% higher than African Americans. In every year the African American origination rates were lower than the Latino prime origination rates. (Figure 7) The Asian prime origination rate pattern was similar to whites. In 2004 Asians had the highest prime origination rate of 60.11% dropping to 49.39% in 2007, rose Figure 7: Prime Loan Origination Rates by Race 23

28 TABLE 5: PRIME LOAN ORIGINATION DISPARITIES Prime Loans 2004 Originations 2005 Originations 2006 Originations 2007 Originations 2008 Originations 2009 Originations Race # % # % # % # % # % # % White 6,018, ,167, ,982, ,580, ,623, ,011, African Am 456, , , , , , Latino 858, , , , , , Asian 426, , , , , , Native Am 28, , , , , , Hawaiian 36, , , , , , Multi Race 19, , , , , , Unk/NA 989, , , , , , Totals 8,834, ,452, ,875, ,182, ,541, ,159,

29 As the mortgage market meltdown unfolded there was a pervasive pattern of prime mortgage credit disparities with African Americans and Latinos. In 2004 the African American prime loan ODI was By 2008 the African American prime loan ODI was In other words, whites were 76% to 224% more likely to receive prime rate loans than African Americans. The Latino ODI had a steady year over year increase, starting in 2004 at 1.20 and ending in 2008 at In 2009, for both African Americans and Latinos, the disparity in the prime rate loan originations actually improved as the ODI decreased to 1.83 and 1.49, respectively. Despite the 2009 improvement, the ODI is further evidence that African Americans and Latinos have broad disparities in access to lower cost prime mortgage credit. The net effect is that African Americans and Latinos pay more than whites and Asians to finance home mortgages in America. All things being equal this leaves African Americans and Latinos with less disposable income to manage other financial obligations and responsibilities, such as, food, clothing, transportation, and education, etc.. Further, because African Americans and Latinos have lower dollar volumes, numbers and market shares of prime loans, it follows they have higher shares of higher cost subprime and FHA loans. Figure 8: Prime Loan Origination Disparity Index 25

30 SUBPRIME RATE LOAN DISPARITIES It may come as a surprise to many people, the degree to which whites dominated the subprime lending market. Between 2004 and 2009, whites had the highest dollar volume, the largest number and the largest market share of subprime rate loans. Moreover, whites were the only racial group to have positive growth with subprime rate loans between 2004 and African Americans and Latinos had a higher proportion of subprime rate loans relative to total loans, but whites had the largest number. Thus, based on the actual number of subprime rate loans, whites are likely to have more loans in default and foreclosure. This section explores the racial composition and the pattern of growth or decline in the dollar volume, number and market share of subprime rate loans. To measure disparities, a Subprime Incidence Disparity Index (SIDI ) is used. The SIDI is calculated by dividing the percent of total loans that are subprime for borrowers of color by the percent of total loans that are subprime for whites. The result is a ratio that compares the incidence or frequency of subprime rate loans by race. The higher the ratio, the higher the incidence of subprime rate loans with borrowers of color. If the SIDI is less than 1.00, it indicates that whites had a higher incidence of subprime than the target group. This section also includes tables for each year ( ) that details total loans, total subprime, percent subprime, the SIDI, average spread and a ratio called the Average HMDA Spread Disparity Index (AHSDI ). The AHSDI measures the disparity in the magnitude, i.e. average cost of subprime rate loans. Unfortunately, the AHSDI for 2009 is distorted because the Federal Reserve Board (FRB) implemented a change in the way subprime rate loans are identified. The effective date of the change was October 1, 2009, as opposed to January 1, In effect, the change impairs any 2009 full year analyses of pricing disparities since two different calculation methods were used that year. DISPROPORTIONATE DROP IN DOLLAR VOLUME OF SUBPRIME RATE LOANS BY RACE The racial distribution and the percent change in the dollar volume of subprime rate loans is shown in Table 6. The table shows a triple digit increase in the dollar volume of subprime rate loans between 2004 and In this period, Asians had the largest percentage increase in the dollar volume of subprime rate loans at %, but in comparison to other groups the actual dollar volume was very low. During the same period, Latinos had a % increase in the dollar volume of subprime rate loans. Their volume grew from $31 billion in 2004 to almost $100 billion by The dollar volume of African American subprime rate loans grew from $29 billion in 2004 to $65 billion in 2005 or %. White subprime loan volume was $100 billion in 2004 and grew to $210 billion in 2005 or %. 26