Federal Deposit Insurance Corporation
FDIG
Acting Chi
Andrew C. Hove, jr.
Division of Research
and Statistics,
Director
Win, Roger Watson
Editor
George E. French
Editorial Committee
Frederick Sr Cams
Gary S. Fissel
Arthur J. Munon
Administrative
Manager
Dctta Voesar
Editorial Secretary
Cathy Wrighr
Graphics and Composition
Geri Bonebrake
The views expressed are Lhose
of the authors and do not neces
sarily reflect official positions of
[he Federal Deposit Insurance
Corporadon. Articles may he
leprinred or abstracted if the
FDIC Banking Review and au-
Lhor(s) arc credited. Please pro
vide the FDlC's Division of
Research and Sratisrics with a
copy of any publications con-
raining reprinted material.
Single-copy subscripriens are
available to I he public free of
charge. Requests for subscrip
tions, back issues or address
changes should be mailed to:
FDIC Banting Review, Office of
Corporate Communications,
Federal Deposit Insurance Cor
poration, 550 17rh Street, N.W.,
Washington, D.C. Z0429.
Spring/Summer 1994
Vol. 7, No. 1
Evaluating the Role of Race in Mortgage Lending
by David K. Home page 1
A statistical analysis by the Federal Reserve Bank of Boston of mortgage-lending patterns in the
Boston area suggests that mortgage lenders discriminated against minority applicants. Examination
of the loan files that comprised the statistical data uncovered a number of factors that may
systematically bias the model's estimates of the race effect. As a consequence, it cannot be
determined whether the higher rejection rates observed for minority applicants reflect these factors
or result from lending discrimination.
Risk Measurement, Actuarially-Fair Deposit Insurance Premiums and the
FDIG's Risk-Related Premium System
by Gary S. Fissel page 16
The author compares the FDIC's risk-related premium system with independent risk classifica
tions derived from a proportional hazards model (PHM). The PHM estimates actuarially-fair
insurance premiums based on econometric estimates of expected time-to-failure. The paper
concludes that the FDIG's relative risk rankings are generally consistent with those of the PHM as
well as with historical failure rates. The premium rate spread between high- and low-risk institu
tions, however, is considerably narrower than what is suggested by the PHM.
Bank Powers and the Separation of Banking from Commerce:
An Historical Perspective
by Christine E. Blair page 28
The author provides an historical perspective on bank powers and the lines of separation between
banking and other activities. This separation has taken the form of both prohibitions on bank
activities and affiliations, and "firewalls" intended to prevent inappropriate flows of capital, funds
and information between affiliates. The scope of permissible bank activities has evolved over the
course of American banking history, changing in response to economic incentives, technological
innovations and political interests. As such change is likely to continue, the debate on appropriate
lines of separation between banking and other activities should remain lively.
Recent Developments Affecting Depository Institutions
by Benjamin B. Christopher page 39
This regular feature of the FDIC Banting Review contains information on regulatory agency actions,
state legislation and regulation, and articles and studies pertinent to banking and deposit insurance
issues.
FDIG
Acting Chi
Andrew C. Hove, jr.
Division of Research
and Statistics,
Director
Win, Roger Watson
Editor
George E. French
Editorial Committee
Frederick Sr Cams
Gary S. Fissel
Arthur J. Munon
Administrative
Manager
Dctta Voesar
Editorial Secretary
Cathy Wrighr
Graphics and Composition
Geri Bonebrake
The views expressed are Lhose
of the authors and do not neces
sarily reflect official positions of
[he Federal Deposit Insurance
Corporadon. Articles may he
leprinred or abstracted if the
FDIC Banking Review and au-
Lhor(s) arc credited. Please pro
vide the FDlC's Division of
Research and Sratisrics with a
copy of any publications con-
raining reprinted material.
Single-copy subscripriens are
available to I he public free of
charge. Requests for subscrip
tions, back issues or address
changes should be mailed to:
FDIC Banting Review, Office of
Corporate Communications,
Federal Deposit Insurance Cor
poration, 550 17rh Street, N.W.,
Washington, D.C. Z0429.
Spring/Summer 1994
Vol. 7, No. 1
Evaluating the Role of Race in Mortgage Lending
by David K. Home page 1
A statistical analysis by the Federal Reserve Bank of Boston of mortgage-lending patterns in the
Boston area suggests that mortgage lenders discriminated against minority applicants. Examination
of the loan files that comprised the statistical data uncovered a number of factors that may
systematically bias the model's estimates of the race effect. As a consequence, it cannot be
determined whether the higher rejection rates observed for minority applicants reflect these factors
or result from lending discrimination.
Risk Measurement, Actuarially-Fair Deposit Insurance Premiums and the
FDIG's Risk-Related Premium System
by Gary S. Fissel page 16
The author compares the FDIC's risk-related premium system with independent risk classifica
tions derived from a proportional hazards model (PHM). The PHM estimates actuarially-fair
insurance premiums based on econometric estimates of expected time-to-failure. The paper
concludes that the FDIG's relative risk rankings are generally consistent with those of the PHM as
well as with historical failure rates. The premium rate spread between high- and low-risk institu
tions, however, is considerably narrower than what is suggested by the PHM.
Bank Powers and the Separation of Banking from Commerce:
An Historical Perspective
by Christine E. Blair page 28
The author provides an historical perspective on bank powers and the lines of separation between
banking and other activities. This separation has taken the form of both prohibitions on bank
activities and affiliations, and "firewalls" intended to prevent inappropriate flows of capital, funds
and information between affiliates. The scope of permissible bank activities has evolved over the
course of American banking history, changing in response to economic incentives, technological
innovations and political interests. As such change is likely to continue, the debate on appropriate
lines of separation between banking and other activities should remain lively.
Recent Developments Affecting Depository Institutions
by Benjamin B. Christopher page 39
This regular feature of the FDIC Banting Review contains information on regulatory agency actions,
state legislation and regulation, and articles and studies pertinent to banking and deposit insurance
issues.
Mortgage-Lending Bias
Evaluating the Role of Race
in Mortgage Lending
by David K. Home*
The controversy concerning
racial discrimination in mort
gage lending has intensified
since the release of the 1990 Home
Mortgage Disclosure Act (HMDA)
data. Although data on mortgage lend
ing have been collected since HMDA
was enacted originally in 1977, lenders
were not required until 1990 to report
applicant characteristics such as race,
ethnicity, sex, age and income. The
aggregate data reveal large discrepan
cies in loan outcomes by race. Ap
proximately 35 percent of black
applicants, 28 percent of Hispanic ap
plicants, and 24 percent of American
Indian/Alaskan native applicants
weredenied mortgage loans in 1991.
versus the 16 percent rejection rate
for white applicants.1 These differ
entials persisted when rejection rates
were compared for applicants within
income categories. Similar patterns
were observed in the 1990 data as
well.2 To many observers, the large
and persistent difference in rejection
rates provides compelling evidence
that mortgage- lending institutions dis
criminate against minority applicants.
Numerous federal regulatory
agencies have responsibility to ensure
that banking institutions comply with
laws that expressly prohibit lending
discrimination, including the Fair
Housing Act and the Equal Credit
Opportunity Act. Compliance exam
inations, which are conducted by
bank regulators on a regular basis, are
one means of evaluating whether
lenders are extending credit in an un
biased manner. The HMDA data are
one source of information that is eval
uated in the compliance examination
process. HMDA reports are exam
ined to determine whether the num
ber of minority applications appears
reasonable, given the demographics
of the lender's market area. Compli
ance examiners also inspect the mi
nority rejection rates to determine
whether these may indicate disparate
treatment on the basis of race or
ethnicity. If evidence suggests that
further investigation is warranted, ex
aminers may analyze different aspects
of the lending program, such as
whether the types of products offered
are appropriate to meet the needs of
the community and whether these
products are marketed to reach all
segments of the community. Compli
ance examiners also scrutinize in
dividual loan application files to
determine whether applicants are
treated fairly without regard to race,
ethnicity, or other protected factors.
Approximately 90 percentof bank
ing institutions receive compliance
ratings of satisfactory or better, appar
ently indicating that the substantive
provisions of the antidiscrimination
regulations are satisfied.3 Rarely have
compliance examinations provided
sufficient data to support a conclusion
chat an institution has engaged in dis
criminatory lending practices. Thus,
the compliance ratings given by bank
regulatory agencies are not consistent
with widespread racial discrimination
suggested by the HMDA data.
In order to explore this inconsis
tency, the Federal Reserve Bank of
Boston (hereafter referred to as the
Boston Fed) surveyed lending insti
tutions in the Boston area to supple
ment the 1990 HMDA data. Many
valid underwriting criteria that lend
ers consider when evaluating poten
tial default risk, such as assets, debt,
'Economist, Division of Research and Sta
tistics, FDIC. The author is indebted to George
Benston, George French, Arthur Murton, Lynn
Nejezchleb, and Win Roger Watson for iheir
helpful comments and suggestions. The author
also thanks Charles Haddad foi assistance re
garding on-site visitations, the many examiners
who devoted considerable time and effort to
examining and documenting information in the
loan files, Jimmy Loyless and Paul Weichman
who supported the participation of staff in the
Boston regional office, and Steven Guggenmos
who provided excellent research assistance.
Asian/Pacific Islander applicants, in con
trast, experienced denial rates (18 percent]
nearly equal to those of white applicants. These
statistics are based on the final 1991 HMDA
data for owner-occupied housing, excluding
multifamily units for the purpose of this com
parison.
For an extended analysis of the preliminary
HMDA data in 1990 and 1991, see Glenn B.
Canner and Dolotes S. Smith (1991), and Can-
nerandSmith(1992).
3In 1992, 90 percent of the 5,602 banking
institutions examined for compliance with CRA
received outstanding or satisfactory ratings (ac-
cording to the July 20, 1993 American Banker,
referencing statistic? repotted in the
CRA/HMDA Updare), while 58 received the low
est racing of "substantial noneompliance." In
the first half of 1993, 93 petcent of the institu
tions received outstanding or satisfactory ratings
and 0.4 percent (ten out of 2,606 lenders) wete
judged to be in substantial noneompliance.
Evaluating the Role of Race
in Mortgage Lending
by David K. Home*
The controversy concerning
racial discrimination in mort
gage lending has intensified
since the release of the 1990 Home
Mortgage Disclosure Act (HMDA)
data. Although data on mortgage lend
ing have been collected since HMDA
was enacted originally in 1977, lenders
were not required until 1990 to report
applicant characteristics such as race,
ethnicity, sex, age and income. The
aggregate data reveal large discrepan
cies in loan outcomes by race. Ap
proximately 35 percent of black
applicants, 28 percent of Hispanic ap
plicants, and 24 percent of American
Indian/Alaskan native applicants
weredenied mortgage loans in 1991.
versus the 16 percent rejection rate
for white applicants.1 These differ
entials persisted when rejection rates
were compared for applicants within
income categories. Similar patterns
were observed in the 1990 data as
well.2 To many observers, the large
and persistent difference in rejection
rates provides compelling evidence
that mortgage- lending institutions dis
criminate against minority applicants.
Numerous federal regulatory
agencies have responsibility to ensure
that banking institutions comply with
laws that expressly prohibit lending
discrimination, including the Fair
Housing Act and the Equal Credit
Opportunity Act. Compliance exam
inations, which are conducted by
bank regulators on a regular basis, are
one means of evaluating whether
lenders are extending credit in an un
biased manner. The HMDA data are
one source of information that is eval
uated in the compliance examination
process. HMDA reports are exam
ined to determine whether the num
ber of minority applications appears
reasonable, given the demographics
of the lender's market area. Compli
ance examiners also inspect the mi
nority rejection rates to determine
whether these may indicate disparate
treatment on the basis of race or
ethnicity. If evidence suggests that
further investigation is warranted, ex
aminers may analyze different aspects
of the lending program, such as
whether the types of products offered
are appropriate to meet the needs of
the community and whether these
products are marketed to reach all
segments of the community. Compli
ance examiners also scrutinize in
dividual loan application files to
determine whether applicants are
treated fairly without regard to race,
ethnicity, or other protected factors.
Approximately 90 percentof bank
ing institutions receive compliance
ratings of satisfactory or better, appar
ently indicating that the substantive
provisions of the antidiscrimination
regulations are satisfied.3 Rarely have
compliance examinations provided
sufficient data to support a conclusion
chat an institution has engaged in dis
criminatory lending practices. Thus,
the compliance ratings given by bank
regulatory agencies are not consistent
with widespread racial discrimination
suggested by the HMDA data.
In order to explore this inconsis
tency, the Federal Reserve Bank of
Boston (hereafter referred to as the
Boston Fed) surveyed lending insti
tutions in the Boston area to supple
ment the 1990 HMDA data. Many
valid underwriting criteria that lend
ers consider when evaluating poten
tial default risk, such as assets, debt,
'Economist, Division of Research and Sta
tistics, FDIC. The author is indebted to George
Benston, George French, Arthur Murton, Lynn
Nejezchleb, and Win Roger Watson for iheir
helpful comments and suggestions. The author
also thanks Charles Haddad foi assistance re
garding on-site visitations, the many examiners
who devoted considerable time and effort to
examining and documenting information in the
loan files, Jimmy Loyless and Paul Weichman
who supported the participation of staff in the
Boston regional office, and Steven Guggenmos
who provided excellent research assistance.
Asian/Pacific Islander applicants, in con
trast, experienced denial rates (18 percent]
nearly equal to those of white applicants. These
statistics are based on the final 1991 HMDA
data for owner-occupied housing, excluding
multifamily units for the purpose of this com
parison.
For an extended analysis of the preliminary
HMDA data in 1990 and 1991, see Glenn B.
Canner and Dolotes S. Smith (1991), and Can-
nerandSmith(1992).
3In 1992, 90 percent of the 5,602 banking
institutions examined for compliance with CRA
received outstanding or satisfactory ratings (ac-
cording to the July 20, 1993 American Banker,
referencing statistic? repotted in the
CRA/HMDA Updare), while 58 received the low
est racing of "substantial noneompliance." In
the first half of 1993, 93 petcent of the institu
tions received outstanding or satisfactory ratings
and 0.4 percent (ten out of 2,606 lenders) wete
judged to be in substantial noneompliance.