A FAIR HOUSING STUDY

*****

I. PURPOSE AND INTRODUCTION

The purpose of this 2005 study was to identify impediments to Fair Housing choice in the area of mortgage lending in Baltimore. Specifically, this is a study of the lending performance by race of 30 leading lending institutions. This study utilizes the 2004 federal Home Mortgage Disclosure Act (HMDA) data to examine possible mortgage lending discrimination. The HMDA requires that all financial institutions receiving a mortgage loan application from the public provide certain information to the Federal Financial Institutions Examination Council (FFIEC), which then creates aggregate and disclosure reports for each metropolitan area that are available to the public.

This study was conducted by the Greater Baltimore Community Housing Resource Board, Inc., (GBCHRB) under contract with the City of Baltimore utilizing funding from the Community Development Block Grant Program of the U. S. Department of Housing and Urban Development. This study first examines the topic of mortgage lending discrimination, then considers available 2004 HMDA data. Recommendations conclude.

*****

II. DISCRIMINATION IN MORTGAGE LENDING

A. INTRODUCTION TO THE PROBLEM

The mortgage lending process consists of a complex series of stages, including advertising and outreach by lending institutions, responses to pre-application inquiries from potential borrowers, approval or denial of loan applications and determination of loan terms and conditions, and finally, loan administration. Discrimination may occur at any of these stages and may take different forms at different stages. The detection of discriminatory practices in mortgage lending is very difficult because of the complexity of the mortgage lending process and the fact that most of the process occurs between lending and banking professionals out of the view and purview of consumers. However, studies conducted in other jurisdictions have found discrimination in mortgage lending can be identified in the following ways:

(1) A consumer can be discriminated against in plain sight - that is, via the actions, behavior, and/or attitudes of the mortgage lending professional providing direct service.

(2) A paired test can be conducted by a nonprofit advocacy group or a governmental agency of the mortgage lending process at a particular lender.

(3) Possible discriminatory practices can be identified via statistical analysis of HMDA data. While this data is not definitive indicating discrimination, it can indicate problem areas and problem lenders for further study.

*****

B. THE HOME MORTGAGE DISCLOSURE ACT: AN INTRODUCTION

1. History and Function

The Home Mortgage Disclosure Act (HMDA), enacted by Congress in 1975, is implemented by the Federal Reserve Board's Regulation C. This regulation provides the public loan data that can be used to assist in determining whether financial institutions are serving the housing needs of their communities; public officials in distributing public sector investments so as to attract private investment to areas where it is needed; and in identifying possible discriminatory lending patterns. This regulation applies to certain financial institutions, including banks, savings associations, credit unions, and other mortgage lending institutions. In 2002, there were approximately 28 million loan records for calendar year (CY) 2001 reported by 7,631 financial institutions. In 2001, 7,713 financial institutions reported approximately 19 million loan records for CY 2000.

Using the loan data submitted by financial institutions, the Federal Financial Institutions Examination Council (FFIEC) creates aggregate and disclosure reports for each metropolitan area that are available to the public at central data depositories. The FFIEC provides to the public various electronic, paper and magnetic media items through the FFIEC's web sites and data request order form. The annual LAR & TS Raw Data CD may be purchased from the FFIEC for $50. This data also can be downloaded from the CD into a spreadsheet application. In August, 2005, the FFIEC released the 2004 Aggregate and Disclosure Reports on CD-ROM. The CD-ROM contains each institution’s 2004 HMDA Disclosure Report, as well as the 2004 HMDA aggregate data. Institutions make their disclosure statements on this CD-ROM available to the public at the institutions’ home offices. Alternatively, institutions’ complete disclosure statements and aggregate reports are available at the FFIEC's HMDA web site, www.ffiec.gov/hmda.

The Appendix contains a Loan Application Register Format and Loan/Application Register Code Sheet, as well as an actual LAR Record for an applicant to First Mariner Bank. The sheets indicate the various fields of data reported on each mortgage lending applicant. Included are:

• The loan type, purpose, and amount.

• Race and sex of applicant.

• If the application was denied, the denial reasons.

• Characteristics of housing unit's neighborhood.

2. Definitions and Glossary of Terminology

The following are glossaries of important terms in the HMDA reporting process. The source of information is the FFIEC's HMDA website (www.ffiec.gov/hmda).

Applicant Race

Race or national origin of the applicant. This is reported for originated loans and for loan applications not resulting in an origination (i.e., approval). Institutions may, but are not required to, report these data for purchased loans. "Not applicable" is used for loans or applications to businesses.

Co-Applicant Race

Race or national origin of the co-applicant. This is reported for originated loans and for loan applications that do not result in an origination. Institutions may, but are not required to, report these data for loans that are purchased. If there is no co-applicant, the numerical code for not applicable is reported. Not applicable may also be used for loans/applications to businesses.

Dwelling

Dwelling means any residential structure whether or not attached to real property, including condominium and cooperative units and mobile or manufactured homes. It refers to both one-to-four-family and multifamily structures. Recreational vehicles such as boats or campers are not dwellings under HMDA.

Loan Application Register (LAR)

A reporter submits HMDA data using the loan application register. The FFIEC uses information provided on the loan/application registers to produce HMDA disclosure statements for each reporting institution, as well as aggregate tables for all covered lenders in each metropolitan area.

Transmittal Sheet

The transmittal sheet is submitted with an institution's data. The transmittal sheet includes the following information: reporter's HMDA ID, regulatory agency, reporter's tax ID, number of loans submitted, HMDA activity year, name of reporting institution, address, contact name, telephone number, and fax number.

For purposes of this study, a mortgage application was deemed approved if the following was reported by the institution:

"Originated."
"Approved but not accepted."
"Loan purchased by institution."

*****

C. DATA ON MORTGAGE LENDING DISCRIMINATION IN BALTIMORE

1. Included Institutions

There follows a listing of the 30 lending institutions included in this study. The institutions solely were chosen because of the comparative number of applications.

Advance Bank

American Home Mortgage Corp.

Bank of America, N. A.

Bradford Bank

Chase Manhattan Mortgage Corp.

CitiFinancial Mortgage Company

CitiMortgage, Inc.

Columbia National, Inc.

Countrywide Home Loans

First Home Mortgage Corp.

First Horizon Home Loan Corp.

First Mariner Bank

Fleet National Bank

GMAC Mortgage

Harbor Bank of Maryland

Household Finance Corporation

Key Bank USA, N. A.

Liberty Mortgage Corporation

M&T Bank

Municipal Employees Credit Union

Nationwide Advantage Mortgage

People's Mortgage Company

Provident Bank of Maryland

Rosedale Federal Savings & Loan

SIB Mortgage Corporation

Slavie Federal Savings Bank

Wachovia Bank

Wells Fargo Bank

Wells Fargo Home Mortgage

WMC Mortgage Corporation

2. Findings of the Analysis

More Black Applications

The first finding is that there were definitely more applications from Blacks than Whites. This is significant because the City's population is approximately 60% Black. For the 30 lending institutions, 51.9% were from Blacks and 45.9% from Whites.

This is a reverse of the 2003 period, in which 53.0% were from Whites and 43.7% from Blacks. It also represent a significant net percentage gain for Asian/Pacific Islanders, and a larger percentage loss for Hispanics.

	*Applicants*	*Percentage*
White	21068	52.9%
Black	23797	42.4%
Hispanic	231	1.1%
Asian/Pacific Islander	762	3.6%
Total	21123

Discrepancy in Applications by Institution

The second finding is that there was a great discrepancy in the number of 2004 loan applications to financial institutions between White and minority group applicants. Some 9 of the 30 institutions (30.0%) studied had a disproportionate percentage of either White or Black. The following four institutions (13.3%) had a disproportionate percentage of White applicants:

	*Black*	*White*
Bradford Bank	7	43
First Home Mortgage Corp.	72	179
First Horizon Home Loan Corp.	276	591
Slavie Federal Savings Bank	2	43

These 5 institutions - 16.7% - had a disproportionate number of Black applicants:

	*Black*	*White*
Advance Bank	25	4
CitiFinancial Mortgage Company	102	39
Household Finance Corporation	1452	380
Nationwide Advantage Mortgage	27	8
Wachovia Bank	573	250

This indicates a somewhat segregated lending market, with Whites apparently going to some lenders and Blacks to others.

The percentage of disproportionate institutions has remained fairly consistent between 2003 and 2004.

Higher Approval Rates for Whites

The third finding is that the percentage approval rate of 2004 applications also varied considerably by institution by race.

Whites' mortgage applications were approved 71.0% of the time, whereas only 54.5% of Blacks were approved.

	*Approved*	*Denied*
White	71.0%	29.0%
Black	54.5%	45.5%
Hispanic	50.2%	49.8%
Asian/Pacific Islander	66.1%	33.9%

Approval by race also varied considerably among the individual institutions. The following 11 lending institutions had much higher approval rates for Whites than Blacks:

	*Black*	*White*
Bank of America, NA	59.1%	74.7%
Bradford Bank	57.1%	81.4%
Chase Manhattan Mort. Corp.	72.8%	86.0%
CitiMortgage, Inc.	73.0%	94.1%
Countrywide Home Loans	65.3%	82.3%
Fleet National Bank	57.1%	78.6%
Nationwide Advantage Mortgage	37.0%	50.0%
Provident Bank of Maryland	36.8%	56.3%
SIB Mortgage Corporation	62.5%	80.0%
Slavie Federal Savings Bank	50.0%	90.7%
Wells Fargo Bank, NA	67.5%	83.0%

While 23 of the 30 institutions (76.7%) had higher approval rates for Whites, these five (16.7%) had slightly higher rates for Blacks:

	*Black*	*White*
First Mariner Bank	85.1%	82.3%
Key Bank USA, NA	73.4%	67.3%
Liberty Mortgage Corporation	81.0%	77.7%
People's Mortgage Corporation	76.9%	72.3%
WMC Mortgage Corporation	92.8%	85.3%

The 2004 data again is very similar to 2003. In 2003, 80.0% of the lending institutions had higher approval rates for Whites, and 20.0% had Blacks with higher approval rates.

For purposes of this study, a mortgage application was deemed approved if the following was reported by the institution: "Originated," "Approved but not accepted," or "Loan purchased by institution."

Asian/Pacific Islanders

The performance of lending to Asian/Pacific Islanders and Hispanic households varied widely among the 30 institutions. A number of institutions had approval rates higher than Whites; a number were below Whites. Hispanic households generally had lower rates. The primary reason for this wide variance is probably the comparatively fewer number of applications from Asian/Pacific Islanders and Hispanics during the year in study.

	*Applicants*	*Applicant Percentage*	*Approved*	*Denied*
White	21068	99.7%	71.0%	29.0%
Black	23797	112.7%	54.5%	45.5%
Hispanic	231	1.1%	50.2%	49.8%
Asian/Pacific Islander	762	3.6%	66.1%	33.9%
Total	45858

*****

D. IS THERE MORTGAGE LENDING DISCRIMINATION IN BALTIMORE?

Qualifications

The application of any analysis of HMDA data must be carefully qualified. First, while HMDA data provides a "snapshot" of an institution's performance, it is limited because of the complexities of the mortgage application process. Second, HMDA data does not include some other information regarding the applicant which could possibly influence a decision, such as household assets and credit history.

Findings

Given those qualifications, the study's data has indicated some trouble spots in mortgage lending in Baltimore:

(1) Several financial institutions are disproportionately serving either Black or White applicants. The majority of minority City residents - especially Blacks and Hispanics - are applying to a limited number of institutions. While this could be because of marketing by the institutions receiving minority applications, it also could be because some institutions are perceived as discriminatory or are actually being discriminatory in discouraging applications from minorities.

(2) Approval rates of White applications are significantly higher than of Blacks and Hispanics. This may be because of other financial factors (e.g., credit history).

*****

III. RECOMMENDATIONS

A. GBCHRB Mortgage Lending Study - Second Part

The second part of this study will examine possible lending discrimination by analyzing the 2004 HMDA data in more detail. Advanced statistical analysis will be utilized. A report will be prepared and released in December, 2006.

B. Need for More Education About Mortgage Lending Discrimination

Consumer education is necessary to alert prospective homeowners of the possibility of discriminatory practices. It is recommended that the GBCHRB receive increased funding in order to provide education to alert prospective homeowners of the type and possibility of discriminatory practices in the mortgage lending area.

C. Testing of Discrimination in Mortgage Lending Should be Conducted

Testing is a proven method of detecting discrimination in housing. In light of the findings of this study concerning possible problems in mortgage lending, testing should be conducted to determine if there is discrimination. The cost of conducting testing in mortgage lending is much more costly than housing rental or purchase testing because of the longer time periods required for application completion and interview time. Also, the amount and confidentiality of personal information required to be submitted as part of the application process makes this type of testing more difficult and costly.

*****

APPENDIX

(Please email the GBCHRB for a hard copy of this material.)

Loan/Application Register Code Sheet

Loan Application Register Format

Sample LAR Record

Approval Rates Of Mortgage Applications In Baltimore, 2004.

Differences in Approval Percentages by Race, 2004.

Number of Mortgage Applications by Race in Baltimore, 2004.