2005 lending study key findings.

I. What's New?

In an ongoing commitment to cast light on lending patterns and lending disparities in Memphis and Shelby County, the Community Development Council of Greater Memphis (CD Council) and the Center for Community Building and Neighborhood Action (CBANA) at the University of Memphis continue to strengthen their annual analysis of Home Mortgage Disclosure Act (HMDA) data. The 2005 Lending Study includes more in-depth analysis of data reported to HMDA, reconfiguration of the data to include more zip code level analysis, and the integration of HMDA data with other housing market data for Memphis and Shelby County. More specifically, the 2005 Lending Study has been expanded to include:

* More comprehensive analysis of housing market indicators and trend data at the zip code level. Data reported by lenders required to report under the Home Mortgage Disclosure Act (HMDA) is included for both 2004 and 2005. (1) Other trend data is extracted from records filed with the Shelby County Register of Deeds, and is reported for 2000-2006. Indicator data provides greater context within which to interpret lending patterns in particular zip codes.

Zip code level data includes:

* Mortgage originations reported to HMDA

* Subprime mortgage originations reported to HMDA

* Mortgage originations recorded with the County Register of Deeds (2)

* Arms-length residential property sales Quit-claim residential property transfers

* Foreclosure notifications

* Substitute trustee deeds

* Expanded analysis of the "dual mortgage market" to include more comparison between lending by local retail depository institutions (banks and credit unions) and "non-local" lenders. Non-local lenders include depository institutions that do not have a local branch bank but may or may not have a local storefront lending operation, and independent or affiliated mortgage companies that may have a local storefront, work through a local broker, or are accessible through the internet. The expanded analysis includes a discussion of the prime lending market share that has been forfeited to non-local lenders by a mortgage lending retrenchment on the part of the "Big Five" local banks. It is also clear that race-related lending disparities are very much associated with the function of the dual market, and are as important to understand in these terms as in terms of approval and denial rates.

* More in-depth analysis of HMDA-supplied "reasons for loan denial" including more information on the characteristics of applicants who have been denied loans. It is important to note that credit scores are still excluded from HMDA reporting, making it difficult to estimate the independent effect of racial bias in the continuing pattern of racially disparate lending in Shelby County. This deficiency could be countered somewhat by geographic-based credit data available for purchase from the three national bureaus; CBANA is negotiating a data partnership that may make this analysis possible for next year. In the meantime, we compare characteristics of applicants denied by both local and non-local institutions.

* As with the 2004 lending study, "center city" and "non-center city" comparisons have been replaced by comparisons among Zones 1 (center city), 2 (mid-city) and 3 (suburban). (See planning districts overlaid by zones in Map 1.)

* Applications and loans to self-identified Hispanics remain too few for meaningful analysis, so that lending disparities continue to be couched in terms of comparisons between white and African-American applicant-borrowers.

II. Key Findings for the 2005 Lending Study

The Executive Summary reports key findings that 1) build on themes from the 2004 study; 2) develop new themes that are resulting for more comprehensive and contextual analysis; or 3) document significant change from 2004. As always, the study is informed by the desire to understand (and counter) racial and ethnic disparities in lending and grapple with changes in mortgage markets that undermine neighborhoods or might be leveraged to produce positive change.

Key Finding 1: High-cost (3) lending continued to expand in Shelby County from 2004 to 2005, prior to the national realignment in the subprime mortgage market in 2006. The number of subprime loans by 59% from 9,511 in 2004 to 15,080 in 2005.

While we may anticipate some changes in 2006 HMDA data-based on the fallout from an overpriced housing market where "risky" lending was encouraged because of anticipated double-digit annual appreciation--it is not at all clear that the Shelby County housing market was or is being driven by the same considerations. While Shelby County may be influenced by the exit of some subprime lenders from the market, the fact that Shelby County housing continues to be "undervalued" (according to national real estate investment analysts) might mean new money diverted to the relatively safe longer term local market by secondary market investors. It is too soon to tell.

[ILLUSTRATION OMITTED]

What is clear, however, is that we will continue to experience the consequences of high subprime lending--and its related high foreclosure rate in Memphis and Shelby County--probably into at least 2008. The median time to foreclosure for failed subprime loans nationally is 1.8 years; (4) this means that 2006 foreclosures are already being driven by the lending patterns we see in 2005, and that those patterns will continue to have repercussions into 2007 and 2008. A mitigating factor may be Tennessee's Home Ownership Preservation Act, anti-predatory lending legislation that took effect in January 2007. Like reform efforts that are gaining momentum nationally in the wake of spiking foreclosures in overheated markets, the state legislation will depend on successful monitoring and litigation to give it teeth. In the meantime, neighborhood markets that have been targeted by high-cost (and likely predatory) marketers will continue to be in need of anti-foreclosure intervention and "rescue funds"; local interventions are not as well developed as they could be and merit serious examination of what other cities are doing to counter the fallout from high-cost lending. (5)

1.1 High-cost lending increased from 25% to 38% of all loans originating in Shelby County in 2005--a 59% increase.

1.2 High-cost lending is associated with non-local lenders, where 49% of all loans are high cost compared to only 10% for local retail banks and credit unions.

1.3 Non-local lenders account for 93% of all high-cost loans.

Key Finding 2: High-cost lending is heavily skewed toward Zone 2, which accounted for 60% of the total. Zone 2 consists mostly of neighborhoods "in transition" Zone 2 neighborhoods are suburban-style neighborhoods that were typically developed beginning in the 1970s; development may be continuing on the city-county borders of neighborhoods like Hickory Hill, Whitehaven, and Raleigh. These neighborhoods experienced a decline in income and education levels between 1990 and 2000, but remain well above citywide medians for income and college education, and are typically thought of as black middle class neighborhoods. The geographic concentration in Zone 2 suggests a marketing strategy that targets black homebuyers.

2.1 Zone 2 includes the four planning districts with the highest rate of high-cost lending in 2005: Frayser at 35%, Nonconnah with 19%, Raleigh-Bartlett at 13.2%, and McKellar-Whitehaven at 11.4%. Zip codes within the planning districts reveal more concentrated evidence of high-cost lending.

2.2 Twenty of 32 Shelby County zip codes (60% of all zip codes) have high-cost lending rates of at least one-third.

2.3 Eleven zip codes have at least 50% subprime--one of every three zip codes countywide.

Key Finding 3: High-cost lending is shifting to Zone 3, where the number of high-cost loans in Cordova zip codes 38016 and 38018, for example, more than doubled from 2004 to 2005. Zone 3 includes zip codes in suburban, recently annexed, and still unincorporated areas of Shelby County. African-American borrowers are more likely than white borrowers to be working with a high-cost, non-local loan in Zone 3 as elsewhere. Among African-American borrowers with high-cost loans, 97% of loans were from non-local lenders. Even among prime borrowers, 80% of loans to African-Americans for 2005 in 38016 and 38018 involved non-local lenders.

Since research by Fannie Mae has documented that some 30% of high-cost borrowers, based on borrower qualifications, should have been offered prime-rate loans, it is important to question the association between African-American borrowers and high-cost loans. That the pattern is being replicated in higher income and higher priced zip codes suggests the efficacy of particular marketing strategies that target African-American buyers and borrowers in Memphis. The strong involvement of non-local lenders suggests that community developers need to be looking at alternatives and reexamining the role of local lenders.

3.1 Five zip codes saw at least a doubling of their high-cost loans from 2004 to 2005, including three in Zone 3 (Cordova and the Poplar Corridor).

3.2 In annexed Cordova and unincorporated Cordova, zip code 38016 experienced a 139% increase from 272 loans to 649 and 38018 experienced a 100% increase from 403 to 807 high-cost loans.

3.3 The Poplar Corridor zip code 38119 increased from 147%, from 73 to 180.

3.3 Suburban Collierville 38107 recorded a 96% increase from 318 to 622 high-cost loans.

3.4 Suburban Lakeland-Arlington zip code 38002 experienced a 74% increase from 202 to 352 high-cost loans.

3.5 Bartlett zip codes experienced double-digit percentage increases in high-cost loans, from 90 to 155 in 38133 (72%); from 386 to 717 in 38134 (86%); and from 221 to 393 in 38135 (77%).

3.6 Germantown zip code 38138 increased from 133 to 233 high-cost loans (75%).

3.7 Paralleling Zone 3, the two most affluent zip codes in Zone 1 (Downtown 38103 and Midtown 38104) also doubled; Downtown remained low in real numbers, increasing from 9 to 24 (167%), but Midtown recorded an increase from 119 to 239 (100%).

Key Finding 4: Foreclosure notifications in Shelby County increased by 128% from 2004 to 2005, for a total of 10,515 in 2005. From 2000 through 2006, nearly 50,000 notifications were published countywide, with real numbers and increases from year to year paralleling zip code analysis of high-cost lending. Zone 2 remains hardest hit and Zone 3 is experiencing the most dramatic increases.

It is likely that more detailed analysis of property transaction data collected by CBANA for Hickory Hill, Frayser, Orange Mound, and the LeMoyne-Owen neighborhood in 2006 (the Property Transaction Database records every property transaction including sales and other transfers and all mortgage transactions) will reveal foreclosures falling into several patterns vis avis high-cost lending. Preliminary analysis suggests that one group of higher income buyer-borrowers is stretching to buy ever higher-priced properties using "exotic" interest-only and balloon mortgages, while other low to moderate income borrowers are buying only modest prices homes but putting themselves at special risk when high-cost loans result in unrealistic expenses. Table 1 documents foreclosure notifications by zip code for 2000 through 2006. Map 2 shows the percentage of high-cost loans by zip code. Map 3 show the concentration of foreclosures by zip code for 2006, while Map 4 shows the cumulative concentration of foreclosures by zip code between 2000 and 2006.

4.1 Zip codes with at least 2,000 foreclosure notifications for the seven year period 2000-2006 include Frayser, Raleigh, Whitehaven, Fox Meadows, Hickory Hill, and

southeast Shelby County zip codes 38125 and 38141; most are in Zone 2.

4.2 In Zone 3, Cordova once again stands out, with a 487% increase in foreclosure notifications from 52 in 2000 to 305 in 2006 for zip code 38016 and a 486% increase in 38018 from 59 52 in 2000 to 346 in 2006. For the seven year period, both zip codes combined featured 1,267 foreclosure notifications in largely annexed neighborhoods.

Key Finding 5: Neighborhoods with large concentration s of high-cost loans and high numbers and incidence of foreclosure can be characterized as "foreclosure-tolerant" and "foreclosure-driven" Foreclosure tolerance means that housing sales remain robust amidst continued high volume lending; indeed, there appears to be something about high-risk lending that has attracted brokers, lenders, and real estate agents to support a high volume of sales activity. Foreclosure-driven means that foreclosed properties (as well as short sales and other desperation measures) account for a high proportion of property transfers (6)--1ikely attracting investors rather than homeowners in the process.

Each neighborhood housing market can be better understood by comparing HMDA data with the full range of property transactions--public records of which are typically available by zip code through the Register of Deeds. We know too little about local mortgage lending that is not reported through HMDA, but it is clear that there is a significant volume of non-reported mortgage activity in Shelby County.

5.1 Comparing number of foreclosure notifications to number of sales, Zone 2 neighborhoods appear to have as much or more than half of property transfers related to foreclosure; Frayser and Hickory Hill stand out.

5.2 A high volume of mortgages to sales suggests a great deal of refinancing, much of which may be counterproductive. We may be able to establish some optimal level, beyond which a higher volume would be a red flag for intervention.

5.3 Substitute trustee deeds mean a completed formal foreclosure. Based on preliminary analysis of chain of title in neighborhoods being tracked through CBANA's Property Transaction Database; the gap between foreclosure notifications and substitute trustee deeds appears to be driven more by "rescue" sales more than workouts or other forms of mitigation that benefit the homeowner.

5.4 Quit claim activity is highly variable by neighborhood and merits further investigation of how investors use quit claims to avoid code enforcement, flip property for tax advantages, and use other tactics that can undermine neighborhoods.

Key Finding 6: Local lenders' relatively small 17% share of mortgage applications translates to untapped mortgage lending opportunities. Local lenders collected 15,924 applications for mortgage loans compared to 78,555 for non-local lenders in 2005; they originated 10,904 loans (69% of applications) compared to 28,444 (36% of applications) for non-local lenders.

It is important to recognize that even though almost half of non-local lending is high-cost lending, the remaining half of the loans are "prime" market loans with borrower characteristics that rival local lenders' own customers and loans. Table 2 compares non-local loans and borrowers fairly favorably with local loans and borrowers.

6.1 Local retail banks and credit unions market share of loan applications decreased from 19% to 17% from 2004 to 2005.

6.2 Local lenders originated 28% of all mortgage-related loans in 2005, accounting for 31% of dollar volume. Basically unchanged from 2004, higher approval rates for local lenders (79% for local lenders compared to 59% for non-local) results in a higher market share of originated loans than loan applications. Higher approval rates are associated with the local lender customer base, which is somewhat more affluent and more likely to be white compared to the non-local customer base. Local lenders are clearly less connected to the African-American market.

[ILLUSTRATION OMITTED]

6.3 The dollar volume of non-local lending is over twice that of local lending: $3,359,544,000 compared to $1,536,010,000 (see Charts 1 and 2 and Table 3).

6.4 Non-local lenders have greater attrition from the application pool in the form of incomplete and withdrawn applications and "approved but not accepted by borrower" outcomes than local lenders (29% for non-local vs. 11% for local). This pattern suggests that brokers for non-local lenders are submitting multiple applications, which can reflect on either the attractiveness of the applicant or venue shopping to benefit brokers; we know too little about why there is so much attrition.

6.5 Non-local lenders have clearly cultivated the high-cost loan market in Shelby County; 49% of non-local loans are high cost compared to 10% of local lender loans. Based on national research, we know enough about the role of independent mortgage companies and broker practices to infer that much of this activity is predatory and involves persuading borrowers to accept less favorable terms than they qualify for in about 30% of cases.

6.6 It is important to emphasize, however, that non-local lenders originate more prime market loans than all local lenders combined. The value of prime loans originated by non-local lenders is 23% greater than all local loans combined: $1,892,061,000 for prime non-local loans compared to $1,536,010,000 for all local loans. (See Chart 2.)

6.7 Wells Fargo and Countrywide, both of which have local storefronts, led the pack among non-local lender prime originations with 1,580 and 1,415 loans originated, respectively, for 2005. This total of 2,995 loans is equivalent to 27% of all loans originated by local lenders combined. When we add Community Mortgage (836 prime originations) and Washington Mutual (434 prime originations), total prime loan originations (4,265) on the part of the "big four" non-local lenders equals 39% of originations by local lenders.

[ILLUSTRATION OMITTED]

6.8 Either local lenders--especially the "Big Five"--have de-emphasized mortgage lending in their business plans and are not too concerned with these missed opportunities, or have failed to develop effective outreach and marketing strategies to better connect with the African-American market in Memphis and Shelby County. (7) Table 4 compares loan originations for local lenders, non-local originations as a whole, and non-local prime loan originations.

Key Finding 7: The "Big Five" local retail banks (in terms of overall deposits) are forfeiting market share of mortgage lending to the smaller local banks, and all local institutions continue to forfeit market share to non-local lenders, who are in turn associated with high-cost lending and foreclosures.

7.1 The number of applications collected by the Big Five declined 31% from 2003 to 2004 and declined another 11% from 2004 to 2005.

7.2 Smaller local banks are growing, however, and have become more aggressive with mortgage lending. Their applications collected increased by 30%. (The 2005 Lending Study detailed tables include information for all local banks and credit unions.)

Key Finding 8: Racial disparities continue to be suggested by a remarkably consistent 20% approval gap: 67% of white borrowers were approved in 2005 compared to 48% for black borrowers.

Local lenders continue to accept a higher percentage of applications than non-local lenders, but their pool of black applicants is significantly small. Since black borrowers appear more likely to submit multiple applications compared to white borrowers (reflected in high rates of attrition from the applicant pool), it is hard to estimate to what extent potential borrowers remain without access to loans of any kind. Changes in mortgage lending suggest that the type of loan, rather than the historical absence of credit, is the contemporary version of disparity--an argument developed under the rubric "predatory greenlining" in the 2004 study and underscored once again in 2005.

[ILLUSTRATION OMITTED]

Reasons for denial are similar for black and white borrowers and for local and nonlocal lenders, with credit history being the primary reason for denial. (It remains problematic that HMDA reporting does not require credit scores or other criteria on which underwriting decisions are made.) Table

3 compares profiles for borrowers and unsuccessful applicants. More detailed information and discussion of lending disparities is included in Section V of the Analytical Report.

8.1 Denied applications are clearly skewed toward lower income applicants, with 48% of denied applications in 2005 from would-be borrowers with incomes at or below 80% of the area median income.

8.2 Not only are African-Americans over-represented among denials, but so are borrowers interested in property in neighborhoods with minority populations of at least 30%, or where the median income is at or less than 80% of the area median income. This pattern belies the high volume lending outcomes identified for Zone 2. Ultimately, there is such high volume throughput for broker-assisted non-local lending than seemingly contradictory patterns can both be true. The complexity of contemporary lending disparities defies explanation in terms of old-fashioned one-on-one discrimination, but remains perhaps the most fundamental driving force for neighborhood housing markets in Memphis.

(1) It is important to note that not all mortgage lending is reported through HMDA. Investor purchases of residential property using commercial loans, for example, are not reportable under HMDA. Other situations that are unlikely to meet HMDA reporting requirements include brokered "correspondent" loans and liens posted in conjunction with home repair contracts. Broker correspondent loans involve a broker who makes the lending (underwriting) decision while the loan is simultaneously "purchased" by assignment to another lender; these loans need not be reported if the broker-correspondent does not originate 100 or more purchase and/or refinance loans, or does not have company-related assets of over $10,000,000 in the preceding calendar year. Liens for home repair are typically negotiated by lender-contractor partnerships that do not include HMDA-reporting institutions.

(2) Includes all mortgage liens and will typically exceed HMDA-reported mortgages. The significance of this complementary lending stream is discussed under key findings.

(3) We use the term "high cost" instead of "subprime" to distinguish loans having at least a 3.00 point rate spread over prime (for first liens) or at least a 5.00 rate spread for second liens. Since subprime lending is not necessarily but the terms are often used as inappropriately interchangeable, we are using the high-cost terminology simply to avoid the confusion.

(4) Center for Responsible Lending, "Losing Ground: Foreclosures in the Subprime Market and Their Cost to Homeowners" December 2006. http:// www.responsiblelending.orglpdfs/foreclosure-paper-report-2-17.pdf

(5) The Neighborworks Center for Foreclosure Solutions is helping to organized local intervention teams that take a multi-dimensional approach; efforts to date are most likely to be found in older industrial cities in the north. See www.nw.org/network

(6) Substitute trustee deeds typically indicate that the property has been officially foreclosed rather than subject to short sale or other form of foreclosure avoidance; occasionally--and we need to learn more about optimizing this outcome--a homeowner may be "rescued" through foreclosure mitigation that preserves equity, but we expect that the gap between notifications and substitute trustee deeds by zip code is owing more to a range of not-so-desirable outcomes than to effective rescue mitigation.

(7) National research demonstrates that effective marketing to historically underserved African-American borrowers is based on the proactive and personalized outreach associated with affinity marketing; brokers for non-local lenders have cultivated this strategy while local retail banks have not.

This article is taken from the "2005 Lending Study" prepared by the Center for Community-Building and Neighborhood Action and published by the Community Development Council. This portion of the report taken from the Executive Summary is reprinted by permission.

The Center for Community-Building and Neighborhood Action (CBANA) specializes in research and community-based problem-solving on housing, neighborhoods, and community development, where low-income women and their families experience the brunt of housing hardship and diminished support services in urban neighborhoods. Research on the relationship between predatory and other mortgage lending products and high rates of mortgage foreclosure in African-American neighborhoods and among female borrowers is supported by The Brookings Institution's Urban Markets Initiative. CBANA is also affiliated with The Urban Institutes National Neighborhood Indicators Partnership, and is working to develop a set of neighborhood-level indicators to track social inequality and social change in Memphis and Shelby County.

The Community Development Council of Greater Memphis (CD Council) is a network of organizations working to revitalize Memphis neighborhoods through a broad array of community development strategies, including affordable housing, economic development, commercial development, workforce programs, and many others. The Council supports these organizations with training and technical assistance programs; research, publications, and information sharing; and public policy activities focused on reducing barriers to neighborhood redevelopment.

Phyllis G. Betts, Ph.D., Director, Center for Community Building and Neighborhood Action, Fellow, The Urban Child Institute, Research Associate, Center for Research on Women, and Associate Professor, School of Urban Affairs and Public Policy: T.K. Buchanan, Senior Research Associate; Carol Gothe, Data Systems Programmer I; Adam Foster, GIS Mapper; and John D. Smith, Research Associate with special thanks to Emily Trenholm, Executive Director, Community Development Council of Greater Memphis

Table 1. Shelby County Published Foreclosure Notices, 2000-2006

Location
Memphis City        2000     2001     2002     2003

         38103       14        9       19       21
         38104       80       85      106      124
         38105       24       26       33       43
         38106      235      272      357      366
         38107      152      151      191      246
         38108      191      176      231      265
         38109      385      392      501      603
         38111      283      319      386      397
         38112      125      125      150      158
         38114      206      238      297      347
         38315      292      260      310      322
         38116      256      226      360      371
         38117       65       88      127      116
         38118      337      329      462      488
         38119       39       45       53       67
         38120       26       26       37       33
         38122      156      186      230      245
         38125      138      199      258      318
         38126       30       39       32       33
         38127      452      449      563      691
         38128      310      342      459      463
         38141      271      310      435      403
Partial Memphis/Cordova--Shelby County
         38016       52       82      104      149
         38018       59       93      140      147
Bartlett
         38133       69       82      128      129
         38134      105      129      171      191
         38135       73       75      132      118
 Collierville
         38017       59       88      115      123
Eads/Fisherville
         38028        2        5        7       11
Germantown
         38138       27       48       61       69
         38139       13       17       29       18
Lakeland/Arlington
         38002       20       48       74       69
Millington
         38053       63       82      114      134
Totals            4,609    5,041    6,672    7,278

Location
Memphis City        2004     2005      2006

         38103       23       22        30
         38104      108      119       125
         38105       30       41        39
         38106      373      453       509
         38107      221      218       292
         38108      235      259       317
         38109      625      654       829
         38111      449      391       469
         38112      164      154       192
         38114      272      352       433
         38315      364      385       548
         38116      396      430       541
         38117      111      136       147
         38118      499      529       710
         38119       84      102       134
         38120       26       29        32
         38122      245      260       303
         38125      381      477       576
         38126       24       33        39
         38127      714      731       969
         38128      494      529       669
         38141      428      457       592
Partial Memphis/Cordova--Shelby County
         38016      153      201       305
         38018      202      234       346
Bartlett
         38133      113      151       144
         38134      233      216       307
         38135      148      183       240
 Collierville
         38017      139      136       212
Eads/Fisherville
          38028      10       12        11
Germantown
          38138      84       63        67
          38139      36       28        39
Lakeland/Arlington
          38002      98      102       196
Millington
          38053     133      150       153
Totals            7,615    8,237    10,515

Location                    Percent
Memphis City     Totals    Increase

         38103      138        114%
         38104      747         56%
         38105      236         63%
         38106    2,565        117%
         38107    1,471         92%
         38108    1,674         66%
         38109    3,989        115%
         38111    2,694         66%
         38112    1,068         54%
         38114    2,145        110%
         38315    2,481         88%
         38116    2,580        111%
         38117      790        126%
         38118    3,354        111%
         38119      524        244%
         38120      209         23%
         38122     1,625        94%
         38125     2,347       317%
         38126       230        30%
         38127     4,569       114%
         38128     3,266       116%
         38141     2,896       118%
Partial Memphis/Cordova--Shelby County
         38016     1,046       487%
         38018     1,221       486%
Bartlett
         38133       816       109%
         38134     1,352       192%
         38135       969       229%
 Collierville
         38017       872       259%
Eads/Fisherville
         38028        58       450%
Germantown
         38138       419       148%
         38139       180       200%
Lakeland/Arlington
         38002       607       880%
Millington
         38053       829       143%
Totals            49,967       128%

Source: Memphis Daily News.

Table 2. Local Banks and Non-Local Lenders/Independent Mortgage
Companies, Loan Characteristics, 2005

                                            Local Banks

                                             Percent/
                                              Summary     Percent
Loan Characteristics               Count     Statistic     Share

Total Applications *                 14           9          19
Home Purchase *                      80          85         106
Refinancing/Home Equity *            24          26          33
Home Improvement *                  235         272         357
Approved and Originated *           152         151         191
Approved but Not-Accepted           191         176         231
  by Consumer *
Denied *                            385         392         501
Withdrawn *                         283         319         386
Incomplete *                        125         125         150
Approval Rate (1) **                206         238         297
Attrition Rate (2)*                 292         260         310
High Loan-to-Income Ratio *         256         226         360
High Cost Loans (3)                 271         310         435
Totals                            4,609       5,041       6,672

                                   Non-Local Lenders
                                   and Independent
                                   Mortgage Companies

                                             Percent/
                                              Summary
Loan Characteristics               Count     Statistic

Total Applications *                 21       114%
Home Purchase *                     124        56%
Refinancing/Home Equity *            43        63%
Home Improvement *                  366       117%
Approved and Originated *           246        92%
Approved but Not-Accepted           265        66%
  by Consumer *
Denied *                            603       115%
Withdrawn *                         397        66%
Incomplete *                        158        54%
Approval Rate (1) **                347       110%
Attrition Rate (2)*                 322        88%
High Loan-to-Income Ratio *         371       111%
High Cost Loans (3)                 403       118%
Totals                            7,278       128%

* Reported as a percentage of initiated applications
collected within each lender type.

** Reported as a percentage of completed applications
within lender type.

(1) Includes both originated and approved-not excepted applications.

(2) Includes approved-not accepted by consumer, incomplete
applications, and applications withdrawn by consumer.

(3) 3.00+ for first liens and 5.00 for subordinate liens, reported
as a percentage of all originated loans.

Source: Memphis Daily News.

Table 3. Characteristics of Local and Non-Local Lender Oriainations

                           All Local
                             Retail             All Non-
                           Bank/Credit            Local
Loan                          Union              Lender
Characteristics               Loans              Loans *

Total Amount              $1,536,010,000      $3,359,411,000
  Originated
Average Loan              $      140,870      $      118,110
Average Borrower          $      113,400      $       75,500
  Income                            1.75                1.80

                              Non-
                             Local
                              Prime
                             Choice
Loan                        Loans **
Characteristics

Total Amount              $1,892,061,000
  Originated
Average Loan              $      174,500
Average Borrower          $       94,400
  Income                            2.12

* Includes non-local depositories and non-local mortgage
companies, some of  which have local storefronts.

** First-lien loans for purchase or refinancing.

Source: 2005 HMDA Data.

Table 4. Comparative Profiles for Originated and Denied
Loan Applications, 2005

                                       Originated        Denied
                                         Loans        Applications

Percent Purchase Loans                       59.0%            26.0%
Percent for Owner-Occupants                  86.0%            90.0%
Female Applicant Outcomes                    57.0%            43.0%
Male Applicant Outcomes                      59.0%            41.0%
Percent Black                                55.0%            66.0%
Percent Not Reporting Race                   10.0%            25.0%
Percent High Loan to Income                  15.0%            22.0%
Percent Low Income: <80% AMI                 29.0%            48.0%
Percent Low Income: <50% AMI                  9.0%            21.0%
Percent High Income: >150% AMI               45.0%            27.0%
Percent in Minority Census                   46.0%            69.0%
  Tracts: >=30% Non-White
Percent in Minority Census                   38.0%            62.0%
  Tracts: >=50% Non-White
Percent in Low Income Census                 21.0%            37.0%
  Tracts: <=80% AMI
Average Loan                          $124,420          $97,220
Median Loan Value                     $ 91,000          $76,000
Average Applicant Income              $ 85,760          $65,000
Median Applicant income               $ 60,000          $44,000
Average Loan to Income Ratio                 1.80             2.18

Source: 2005 HMDA Data.

Chart 1. Loan Dollars Originated by Local and
Non-Local Lenders, 2005

Local         31.0%   $1,536,010,000
Non-Local     69.0%   $5,359,544,000

Note: Table made from pie chart.

Chart 2. Comparative Dollar Value for Local,
Non-Local, and Non-Local
Prime Choice Lenders

Total Amount Originated

Non-Local Prime Choice     $1,892,061,000
Non-Local                  $3,359,544,000
Local                      $1,536,010,000

Note: Table made from pie chart.