FDIC Federal Register Citations

From: lois_dean@hud.gov [mailto:lois_dean@hud.gov]
Sent: Tuesday, April 06, 2004 3:26 PM
To: regs.comments@occ.treas.gov; regs.comments@federalreserve.gov; Comments; regs.comments@ots.treas.gov
Subject: Community Reinvestment Act comments

April 6, 2004
Docket No. 04-06

Communications Division
Public Information Room, Mailstop 1-5
Office of the Comptroller of the Currency
250 E St. SW,
Washington 20219

Docket No. R-1181
Jennifer J. Johnson

Secretary
Board of Governors of the Federal Reserve System
20th Street and Constitution Avenue, NW
Washington DC 20551

Robert E. Feldman
Executive Secretary
Attention: Comments
Federal Deposit Insurance Corporation
550 17th St NW
Washington DC 20429

Regulation Comments, Attention: No. 2004-04
Chief Counsel's Office
Office of Thrift Supervision
1700 G Street NW
Washington DC 20552

Dear Officials of Federal Bank and Thrift Agencies:

CRA is a law that makes capitalism work for all Americans. CRA is too vital to be gutted by
harmful regulatory changes and neglect

The proposed CRA changes would facilitate predatory lending and reduce
the ability of the general public to hold financial institutions accountable for compliance with
consumer protection laws.

The proposed changes to CRA could harm low-income communities and
lower access to homeownership opportunities for the people living there

The Community Reinvestment Act (CRA), as passed in 1977, prohibits redlining or discrimination against
neighborhoods. Banks and thrifts are obligated to serve all communities in which they are chartered
and from which they take deposits. Under CRA, federal agencies examine and rate most banks
based on how many loans, investments, and services they offer to low- and moderate-income
consumers and neighborhoods.

The proposed CRA changes would undermine the Statute and will thwart the Bush Administration's goals
of creating 5.5 million new minority homeowners by the end of the decade.

The proposed changes would reduce the rigor of CRA exams for a
large number of banks and include an anti-predatory lending test that would permit abusive lending,
instead of severely penalizing banks for predatory lending. I urge you to amend very seriously or withdraw the proposed changes to
the Community Reinvestment Act (CRA) regulations.

Your proposed changes, if not seriously amended or withdrawn,
will slow down, if not halt, the progress made in community reinvestment under every one of the previous administrations over the past 17 years.

Sincerely yours,

Lois Courtney

ldcourtne@aol.com

Note:
Further details of the damage the proposed changes would inflict are spelled out in the document below.

Report provided by Community Action Partnership and the National Community Reinvestment
Coalition.

The proposed CRA changes will thwart the Bush Administration's goals of improving the
economic status of immigrants and creating 5.5 million new minority homeowners by the end of
the decade. Instead, the proposed CRA changes would facilitate predatory lending and reduce
the ability of the general public to hold financial institutions accountable for compliance with
consumer protection laws.

The proposed changes include three major elements: 1) provide streamlined and cursory exams
for banks with assets between $250 million and $500 million; 2) establish a weak anti-predatory
lending compliance standard under CRA; and 3) expand data collection and reporting for small
business and home lending.

The beneficial impacts of the third proposal are overwhelmed by the
damage imposed by the first two proposals. In addition, the federal banking agencies did not
update procedures regarding affiliates and assessment areas in their proposal, and thus missed a
vital opportunity to continue CRA's effectiveness.

Streamlined and Cursory Exams for Smaller Banks – Devastating and Completely
Unjustified

Retain the Large Bank Exams for Institutions with Assets between $250 and $500 Million

Under the current CRA regulations, large banks with assets of at least $250 million are rated by
performance evaluations that scrutinize their level of lending, investing, and services to low- and
moderate-income communities. The proposed changes will eliminate the investment and service
parts of the CRA exam for banks and thrifts with assets between $250 and $500 million.

The notice of proposed rulemaking (NPR) brushes aside the crippling impact that streamlined
exams would have on the continued progress of community reinvestment. The NPR attempts to
minimize the impact of the proposed change by stating that the portion of industry assets subject
to the large bank exam would decline from slightly a little more than 90% to a little less than
90%.1 This presentation obscures the fact that the proposed changes would reduce the rigor of
CRA exams for 1,111 banks that account for more than $387 billion in assets. Wells Fargo and
Company, the fifth largest holding company in the United States, has assets equal to $387
billion.

While the federal banking agencies would be unlikely to propose eliminating the
investment and service test for a lender the size of Wells Fargo, the effect of streamlining the
exams for the so-called smaller banks has virtually the same impact.

1
The elimination of the investment and service tests for more than 1,100 banks translates into
considerably less access to banking services and capital for underserved communities. For
example, these banks would no longer be held accountable under CRA exams for investing in
Low Income Housing Tax Credits, which have been a major source of affordable rental housing
needed by large numbers of immigrants and lower income segments of the minority population.
Likewise, the banks would no longer be held accountable for the provision of bank branches,
checking accounts, Individual Development Accounts (IDAs), or debit card services.

Thus, the effectiveness of the Bush Administration's housing and community development programs
would be diminished.

Moreover, the federal bank agencies will fail to enforce CRA's statutory
requirement that banks have a continuing and affirmative obligation to serve credit and deposit
needs if they eliminate the community development lending, investment, and service test for a
large subset of depository institutions.

NCRC urges the federal agencies to abandon their proposal to eliminate the investment and
service tests for banks and thrifts with assets between $250 and $500 million in order to avoid
significant declines in investments and services for low- and moderate-income communities. At
the very least, the federal agencies must retain the large bank lending test for institutions with
assets between $250 and $500 million. The lending test of the large bank exam is more rigorous
and also contains a community development lending test. The community development lending
test assesses whether banks are making critically needed loans for affordable housing and
economic development activities.

Data reporting requirements regarding small business and farm loans must also remain intact for
banks and thrifts with assets between $250 and $500 million. It would be terribly ironic if the
federal agencies remove the small business and farm data reporting and thus remove the ability
of the general public to gauge if smaller banks with assets between $250 and $500 million are
continuing to respond to credit needs by making critical small business and farm loans. Small
businesses and farms, particularly those in non-metropolitan areas, rely on the smaller banks for
access to loans.

Public data disclosure is a vital means for holding the smaller banks accountable
to serving the small businesses and farms that face a restricted choice of banks.

In addition to retaining the large bank lending test and small business loan reporting, the federal
agencies must retain the investment and service tests for smaller banks that have a sizable
presence in their communities in terms of assets, branches, and loans. NCRC recommends that
the complete large bank test must be applied to banks with assets between $250 and $500 million
if these banks control ten percent or more of the total assets and branches in their assessment
areas or make more than 5 percent of the loans in their assessment areas.

These institutions have
a considerable presence in their communities, meaning that any diminution of their existing CRA
obligations would have a devastating impact on the amount of credit, capital, and banking
services available in their communities. While NCRC does not believe that the proposed
streamlining is justified for any institution, the federal agencies must consider that their proposal
will cause the most damage if they apply the streamlined exams to all banks with assets between
$250 and $500 million, regardless of their presence and weight in the community.

The presence of a holding company must remain a factor in deciding which institutions receive
the streamlined small bank exam or the comprehensive large bank exam. NCRC's analysis
below reveals that a great majority of banks with assets between $250 and $500 million are part
of holding companies with much larger asset sizes. The assets of the holding company and the
option to include affiliates on CRA exams assists small banks in making investments and
community development loans. Since smaller institutions currently utilize assets of their holding
companies and activities of their affiliates, providing streamlined CRA exams to these banks
deprives low- and moderate-income communities of valuable resources for community
development investment and lending.

The current procedure of applying the large bank exam to
small banks that have holding companies of $1 billion or more in assets must remain.

National Analysis Obscures Local and Statewide Impacts

The federal agencies' cursory reference to the relatively small amount of industry assets eligible
for the streamlined exam proposal suggests that the agencies have not scrutinized the profound
impacts on a state and local level.2 The national and aggregate perspective of the regulators is
puzzling, considering that CRA ensures that local, not national, credit and deposit needs are
addressed by banks and thrifts. The very essence of the streamlined exam proposal suggests that
the agencies are violating their statutory responsibilities to require banks to meet local needs of
all the communities in which they are chartered.

On a national level, the federal agencies can perhaps dismiss the impact of the streamlined exam
proposal by asserting that only 4.3 percent of the industry's assets would be covered by the
cursory exams. Using the FDIC database on depository institutions, NCRC has found that the
impacts in terms of assets is much larger on a state, urban, and rural level.3 In the state of Idaho,
2 During the FDIC Board meeting on January 20, 2004 considering the proposed changes, new Board member
Thomas Curry asked FDIC staff if staff had conducted an analysis of the impacts of the changes. Staff replied that
they had not.

3 The FDIC database is the Statistics on Depository Institutions (SDI), which is updated on a monthly basis
according to the FDIC web page (http://www.fdic.gov). NCRC downloaded the database in late February 2004.
The database assigns all of the bank assets to the state in which a bank is located. The publicly available FDIC
for instance, smaller banks with assets between $250 and $500 million possess $4.6 billion in
assets and control more than 55 percent of the total assets of depository institutions
headquartered in the state.

In Vermont, the smaller banks and thrifts likewise control 24 percent
of the assets or $1.8 billion in assets. Twenty-seven smaller banks and thrifts in Maryland have a
sizable $9.6 billion in assets or 21.4 percent of the assets of all the depository institutions located
in Maryland (see Table 1 in the Appendix behind the comment letter).
So-called small banks and thrifts with assets between $250 and $500 million control more than
10 percent of total depository institution assets in 20 states. In other words, the so-called small
bank and thrifts control more than twice their national share of 4 percent of assets in almost half
of the states.

Ten percent of total assets on a statewide level is quite significant. If these banks
were to shut down, the financial resources of banks available to state residents for investment
purposes would suddenly decline by 10 percent. That means much less investment available for
commercial and residential development. Yet, the elimination of 10 percent of bank assets for
investment and community development lending is precisely what the regulators propose for
low- and moderate-income communities in about half of the states.

The proposed streamlined exam would have the most devastating impact for rural America since
the so-called small banks have their largest presence in non-metropolitan areas. According to the
FDIC database, small banks and thrifts with assets between $250 and $500 million hold $126
billion of total assets of banks located in rural areas. This amount is 18.8 percent of all bank
assets in rural areas, or more than 4 times the portion of assets that the smaller banks control in
the nation as a whole. In other words, the impact of streamlining CRA exams is about 4 times
worse (in terms of assets available for bank investments and services) in rural areas than in the
nation as a whole.

In eight states, institutions with $250 to $500 million in assets control more than one third of the
bank assets in rural areas. In Vermont, just five smaller banks possess $1.7 billion in assets or
more than 53 percent of the assets in rural counties in that state. Similarly, in Utah and Idaho
banks and thrifts with assets between $250 and $500 million control more than 50 percent of all
assets in rural areas. The "smaller" banks and thrifts in Massachusetts, Washington, Virginia,
Alaska, Maryland, and Maine possess between 30 to 44 percent of the assets in non-metropolitan
counties (see Table 2).
Databases do not provide sufficient detail to determine if banks distribute their assets among their interstate
branches. For the purposes of this comment letter, NCRC assumes all of the bank assets are located in the states in
which the banks are headquartered.

Banks and thrifts with assets between $250 and $500 million control more than 20 percent of the
total assets held by depository institutions in rural areas in 18 states. These banks and thrifts
control 10 percent or more of the assets in rural areas in 41 states. While the regulatory agencies
may refer to these institutions as small banks, it is clear that they are a major source of
investments and services to rural areas. Streamlining their CRA exams would result in
disinvestment from rural parts of the country, which are least able to deal with the loss of bank
investment and community development lending.

The impact of the proposed streamlining is greater for urban areas than would be expected. In
fourteen states, banks and thrifts with assets between $250 and $500 million control 10 percent
or more of all assets of depository institutions located in metropolitan areas. In Colorado, small
banks possess a large $8.7 billion in assets or 22 percent of all the assets of lenders located in
metropolitan areas. Similarly, in Maryland, small banks and thrifts control $7.5 billion in assets
or 19.4 percent of all the bank assets in urban areas (see Table 3).

The impact when considering the number of lenders as opposed to assets is also dramatic in a
number of states. Overall, the proposal would eliminate the large bank exam for 20 percent or
more of the lenders located in 12 states. In other words, at least 20 percent of the lenders in these
twelve states have assets between $250 and $500 million. Likewise, the proposal would
eliminate the large bank exam for 10 percent or more of the lenders in 35 states.

The proposal would wipe out the large bank exam for 20 percent or more of the rural-based banks in 15 states.
In seven states, more than 30 percent of the lenders based in rural counties would be exempted
from the large bank exam. For example, 33 percent, or 20 of the 60 banks and thrifts located in
rural parts of Virginia, have assets between $250 and $500 million, and thus would no longer
undergo the large bank exam. Finally, the impact is also significant in metropolitan areas as 20
percent or more of the lenders in urban areas in 16 states would be exempt from the large bank
exam.

Reductions in Community Development Investments

NCRC analyzed the CRA exams of 40 banks and thrifts with assets between $250 and $500
million to assess the impacts on the level of investments and community development lending if
the small bank exam applied to these institutions (see Table 4 for a list of lenders in the sample).
The analysis scrutinized exams in four states (Vermont, Maryland, Colorado, and Arkansas) in
which smaller banks controlled the largest percentage of assets.4

The analysis reinforces the devastating impact of the proposed streamlining. The 40 banks and
thrifts in the sample made a total of $69,450,000 in qualified investments, according to their
CRA exams. They also issued $92,642,000 in community development loans. The community
development lending and investment combined equals more than $162 million.

For the four states of Vermont, Maryland, Colorado and Arkansas, $162 million in community development
lending and investment represents a substantial source of revitalization financing. The loss of
this financing would be felt many times over since community development investing and
lending of this magnitude creates hundreds, if not thousands of jobs, and increases the
purchasing power of local workers and communities.

Assuming that these banks and thrifts are representative of all depository institutions with assets
between $250 and $500 million, the total amount of community development lending and
investing by the "smaller" lenders equals more than $4.5 billion. This is the amount of lending
and investment that occurs roughly every two to three years or the approximate time period
between CRA exams.

Eliminating the large bank lending and investment test for these lenders
translates into dramatically fewer dollars in community development loans and investments for
low- and moderate-income communities. Even if NCRC's sample is not statistically
representative, the order of magnitude in lost investments and loans is likely to be in the
hundreds of millions, if not billions, of dollars.

As detailed below, eliminating the investment
and community development lending tests reduces the level of investment and community
development loan dollars by at least half in the NCRC sample of CRA exams.

Scrutinizing the Investment Tests of the 40 banks and thrifts in the sample, NCRC found that the
average investment amount of the 11 depository institutions receiving Outstanding ratings on the
Investment Test was $3.7 million or 1.36 percent of their assets. The average investment of the
10 depository institutions with High Satisfactory ratings on the Investment Test was $1.6 million
or .65 percent of their assets. In sharp contrast, investment dollars and percent of assets was less
than half that level for banks with lower ratings. The 16 banks and thrifts with Low Satisfactory
ratings made an average investment amount of just $734,000 or a mere .21 percent of their
assets. The 3 banks and thrifts with Needs-to-Improve ratings made a measly $171,000 in
qualified investments or .06 percent of their assets.

Idaho is that state in which smaller banks and thrifts control the largest percentage of assets. We were unable to
find large bank CRA exams for these institutions; the institutions had assets under $250 million at the time of their
most recent CRA evaluations and thus were examined under the small bank exam.

The upshot of this analysis is that it is very likely that eliminating the investment test for banks
and thrifts with assets between $250 and $500 million would reduce their investments in lowand
moderate-income communities by at least half. Banks with High Satisfactory ratings made
twice as many qualified investments (measured in terms of dollars) than banks with Low
Satisfactory ratings. The differences are even more extreme if comparisons are made among
banks with Outstanding, High Satisfactory, Low Satisfactory, and Needs to Improve ratings.

Therefore, a conservative estimate is using the difference between banks with High and Low
Satisfactory ratings. In the absence of Investment Tests, it is reasonable to assume that banks
with High Satisfactory ratings would invest at the level of banks with Low Satisfactory ratings.
This suggests that the banks with High Satisfactory ratings would reduce their level of
investments by half. Since the comparison between banks with High and Low Satisfactory
ratings is a conservative estimate of impacts, it is likely that all banks (regardless of their ratings)
would cut the dollar amount of their qualified investments by half in the absence of an
investment test.

Reductions in Community Development Lending

The decrease in community development lending is even greater for NCRC's sample of 40 banks
with assets between $250 and $500 million. The five depository institutions with Outstanding
ratings on the lending test had an average community development lending level of $4.7 million.
Their ratio of community development lending to assets was 1.46 percent. The sixteen banks
with High Satisfactory ratings on their lending test had an average of $3.2 million in community
development loans and a community development lending to asset ratio of 1.03 percent. In sharp
contrast, the nineteen banks with Low Satisfactory ratings on the lending test made an average of
only $950,000 in community development loans and had a dismal .3 percent ratio of community
development loans to assets.

The banks and thrifts with Outstanding and High Satisfactory ratings on their lending tests made
between 3 and 4 times the level of community development lending as institutions with Low
Satisfactory ratings. Again, a conservative estimate of the impact of eliminating the community
development lending test would be the difference between High and Low Satisfactory
institutions. Assuming that this difference would apply to all institutions regardless of their
ratings, the level of community development lending would be two thirds less if the federal
agencies eliminate the community development lending test of the large bank exam for
institutions with assets between $250 and $500 million.

Concrete Examples of Community Development Loans and Investments Likely to Disappear

Quantifying the proposal's likely decreases in reinvestment is compelling, but concrete examples
clearly and powerfully illustrate the looming harm of the proposals.

Simply put, the streamlining would result in much less affordable rental housing, fewer homeless shelters, less economic
development projects, and fewer community health centers and other facilities.

On most of these projects, banks realize a profit.

Projects that do not generate economic returns, such as homeless
shelters, still benefit banks and their local communities by reducing poverty and deprivation.5

If the federal agencies believe that it is desirable to substantially decrease affordable housing and
economic development activities, then they should proceed with their proposed streamlining.

If, on the other hand, the regulators come to believe that the societal and human costs of
streamlining are too high, they should immediately abandon their proposal.

In Maryland, banks with assets between $250 and $500 million have been motivated by CRA
exams to undertake a variety of critical community development loans and investments. For
instance, Arundel Federal Savings Bank invested $625,000 in Maryland Community
Development Administration bonds and purchased $20,000 of tax credits from the Anne Arundel
County Chapter of Habitat for Humanity. Bradford Bank originated a $2.5 million loan to
refinance and renovate shopping centers in eastern Baltimore County. Carrollton Bank purchased
two Fannie Mae Mortgage Back Securities totaling $3 million, which provided funds to finance
mortgages for multi-family housing dedicated to those with limited incomes. Carrollton also
made available two lines-of-credit totaling $800,000 to a nonprofit organization that operates a
Baltimore County residential treatment center for low-income adolescent females.

In Colorado, Pueblo Bank & Trust Company's overall level of community development lending
has been extraordinary, according to the most recent CRA exam. In 2001 and 2002, Pueblo
B&T originated 57 community development loans totaling approximately $24,422,000. Many of
these loans went to providing affordable housing to low- and moderate- income individuals.
Community development loans equaled an incredible 7 percent of Pueblo's assets, about 5 times
the portion of assets that banks with Outstanding ratings on the lending test in NCRC's sample
devote to community development lending. As civic minded as Pueblo Bank & Trust may be, it
is unlikely that they would continue their impressive performance should the community
development lending and investment tests be abolished.

In January 1997, First Bank of South Jeffco, Colorado purchased $800,000 in a Sheridan School
District, Arapahoe County, Refunding and Improvement Bond. Proceeds of the bonds paid the
5 In terms of economic theory, CRA has encouraged banks to "internalize" the positive externalities of some social
projects that otherwise would not be undertaken since no party realizes private profit from them.

cost of capital improvements at elementary, middle, and high schools, and an early education
center that houses a head start program. In 1999, First Bank purchased a portion of a 99 percent
limited partnership interest in the Littleton Creative Housing Limited Partnership for $2,800,000.
The partnership owns and operates the Libby Bortz Low-Income Housing Assisted Living
Center.
Also, in Colorado, First Bank of Boulder purchased a total of $3,700,000 in Colorado Housing
and Finance Authority (CHFA) Single-Family Revenue Bonds since its last evaluation. The bond
programs are specifically targeted for low- and moderate-income individuals/families in
Colorado.

In Arkansas, Citizens Bank originated $3,100,000 in loans for White River Medical Center,
according to the most recent CRA exam. The two loans provided financing for working capital
and construction of nursing home and retirement facilities, all of which primarily served lowand
moderate-income individuals and Medicaid patients. Finally, First National Bank of
Springdale originated 54 community development loans totaling $4.3 million. FNB Springdale's
community development loan portfolio consists of short-term affordable housing construction
loans
.
As these examples illustrate, elimination of the community development lending and investment
test entails the elimination of critical affordable housing, economic development, and community
facility projects.

In many small and medium-sized metropolitan areas and rural counties, it is
unlikely that banks still subject to the large bank exam would step in and fill the gap in
community development lending and investing. The banks with assets between $250 and $500
million are most likely to have assessment areas that are confined to the smaller metropolitan
areas and rural communities. In contrast, the larger banks are likely to have assessment areas
that include more geographical areas, meaning that they are less focused on the credit and
development needs of the areas served by banks with assets of $250 to $500 million. The loss of
community development lending and investing is likely to be permanent in parts of the country
least able to withstand a withdrawal of capital and credit.

Elimination of Service Test Will Reduce Access to Branches

The FDIC database also reveals the dramatic impacts that eliminating the service test will have
on access to branches. If the federal agencies eliminate the service test, it is quite likely that
small banks will de-emphasize their branching network and/or reduce the number of services and
products that the branches offer to low- and moderate-income communities.

In the United States, as a whole, small banks and thrifts with assets between $250 and $500
million own almost 10 percent of the branches. They own 7,985 of the 87,357 branches serving
the general public.6 As stated above, NCRC believes that any subset of institutions that control
either 10 percent of the assets or 10 percent of the branches in a geographical area have a
significant impact in terms of access to credit, investments and banking services. Therefore,
when just confining the analysis to a national level, the large bank exam and the service test must
not be eliminated for banks and thrifts with assets between $250 and $500 million since these
institutions have a significant branching presence across the country.

When the analysis is conducted on a state level, the branch presence is even larger for the socalled
smaller banks and thrifts. In 25 states, the smaller banks have more than 10 percent of the
branches. In 10 states, they own 15 percent or more of the branches. The branch presence of the
smaller banks is dominant in the more rural states. In Maine, the "smaller" banks own 29
percent or 146 of the 504 branches in the state. Likewise, they own 19.8 percent and 17.6
percent of the branches in South Dakota and Idaho, respectively (see Table 5).

The impact of the proposed abolition of the service test is the most severe in rural areas because
of the large presence of branches owned by the smaller banks and thrifts. Banks and thrifts with
assets between $250 and $500 million control more than 10 percent of the non-metropolitan
branches in 32 states. They possess 20 percent or more of the rural branches in 7 states. In
Virginia, for example, the "smaller" banks and thrifts own 169 of the 697 branches or 24.2
percent of the rural branches. Likewise, in New Hampshire, they control 51 of the 216 branches
or 23.6 percent of the rural branches (see Table 6).

The effect of the streamlining on urban areas is also significant. In nineteen states, small banks
and thrifts with assets between $250 and $500 million own 10 percent or more of the branches in
metropolitan areas. Not surprisingly, the more rural states such as Wyoming and Montana have
significant percentages of metropolitan area branches owned by the smaller banks. Even more
urban states including Massachusetts and Missouri have a significant portion of metropolitan
branches owned by the smaller banks (see Table 7).

The impact by deposits is also striking. Across the United States, the so-called smaller banks
and thrifts with assets between $250 and $500 control more than $302 billion in deposits. In
6 NCRC used the FDIC's Summary of Deposits Database for the analysis. The most recent data available for
downloading was June 30, 2003. NCRC eliminated branches from our sample that did not accept deposits and serve
the public. These included administrative offices, trust offices, messenger offices, loan production offices, and
consumer credit offices.

seventeen states, they control more than 10 percent of the deposits. Again, the impacts of the
streamlining would be most crippling in rural areas. In 36 states (more than two thirds of all
states), the "smaller" banks and thrifts have more than 10 percent of the deposits in rural areas.
In 18 states, they control more than 15 percent of the deposits. For instance, in Maryland, they
control more than $1.2 billion of the $5.6 billion or 21 percent of deposits collected by rural
branches.

The smaller banks and thrifts control more than 20 percent of the rural deposits in
Maine, South Dakota, Virginia, Vermont, Maryland, Idaho, and New Mexico. These states can
ill afford the smaller banks and thrifts neglecting the deposit and service needs of rural residents.
The payday and subprime lenders will sense even more of a market opportunity and replace
mainstream bank products with higher rate consumer and home loans. The resulting reductions
of community and consumer wealth will further retard economic development efforts.

Bank Holding Company Must Remain a Consideration

As stated above, removing the bank holding company as a factor in differentiating between small
and large banks will allow many institutions with sufficient resources to unfairly enjoy the
streamlined test and abdicate their responsibilities for providing branches and community
development investments and loans in low- and moderate-income communities. Using the FDIC
database, NCRC calculates that 815 of the 1,111 small banks and thrifts with assets between
$250 and $500 million are owned by holding companies. More than 73 percent of the so-called
smaller banks and thrifts are owned by holding companies. This is a higher percentage than all
banks and thrifts; about 70 percent of all banks and thrifts are owned by holding companies.
Not only are a greater percentage of smaller institutions owned by holding companies, NCRC's
sample of 40 CRA exams reveals a substantial amount of holding company assets available to
the smaller institutions. In the sample, 37 of 40 banks in the states of Arkansas, Colorado,
Maryland, and Vermont had holding companies. This is the great majority or 92 percent of the
banks in the sample. While about three quarters of the smaller banks and thrifts nationwide have
holding companies, the portion is even greater in a number of states including those in the NCRC
sample of CRA exams.

Some holding companies in NCRC's sample of CRA exams had considerable assets well above
$1 billion. These holding companies include UMB Financial with $8 billion, Mercantile
Bankshares with $9.9 billion, Fulton Financial with $6.9 billion, First Bank Holding Company of
Colorado with $5.7 billion, First Tennessee National Corporation with $23 billion, and First
Nations of Nebraska with $9.7 billion. In a couple of cases, one holding company owned a
sizable number of banks in the NCRC sample. For example, in Colorado, First Bank Holding
Company owned 11 of the 15 banks in that state.

Similarly, in Maryland, Mercantile Bankshares owned 6 of 17 banks. Moreover, in the Colorado exams of banks owned by First Bank Holding
Company, the banks often claimed credit for community development loans and investments
undertaken by affiliates.

In other words, the holding company made its resources available to their banks for CRA exam
purposes. Eliminating the holding company as a factor in differentiating between small and
large banks therefore results in major financial institutions abdicating their community
reinvestment obligations and greatly diminishes the amount of holding company assets available
to businesses and consumers in low- and moderate-income communities.

Burden Argument

The benefits of large bank CRA exams are substantial and are likely still underestimated by the
conservative approach of the NCRC analysis. The application of the large bank CRA exam to
banks and thrifts with assets between $250 and $500 million has made thousands of branches
and billions of dollars in community development loans and investments available to low- and
moderate-income communities. Consequently, the proposed elimination of the large bank exam
for the so-called smaller banks poses the threat of withdrawing access to a substantial number of
branches and financial resources for reinvestment.

The burden of large bank exams for the so-called smaller banks appears to be minimal while the
benefits of the exams are profound for low- and moderate-income communities. During a
session held by the FDIC on regulatory streamlining, Charlotte Bahin, Senior Vice President of
Regulatory Affairs of America's Community Bankers, stated publicly that most smaller
institutions no longer complain about the burden of CRA exams.7 According to Ms. Bahin,
smaller banks worry that they are compared to larger banks on CRA exams, but they are not
concerned about the CRA exam process, in and of itself. With almost a decade of experience
with CRA exams, the smaller institutions are now accustomed to the exams.

The comments of Ms. Bahin regarding perceptions of unfair comparisons to larger banks on
CRA exams can be readily put to ease by appropriate CRA examination procedures. The CRA
exams scrutinized by NCRC compared small banks against other smaller banks. This is wellestablished
CRA exam procedure. Moreover, the examiners also remark that they take into
account, when appropriate, how the presence of large banks can impact smaller bank
performance on any part of the exam. This procedure is referred to in CRA jargon as the CRA
performance context.

7 FDIC Session on the Economic Growth and Regulatory Paperwork Reduction Act, February 20, 2004.

The time spent by CRA examiners suggests that the CRA examination process for banks with
assets between $250 and $500 million is considerably less time consuming than for banks with a
few billion dollars in assets. According to a CRA examiner NCRC interviewed, a CRA exam for
a bank with half a billion dollars in assets consumes 10 to 15 days of examiner staff time. In
contrast, a CRA exam of a bank with $5 to $10 billion in assets consumes about 20 to 50 days of
staff time. Finally, a CRA exam of a bank with more than $40 billion in assets consumes about
100 days of staff time. It is reasonable to assume that CRA examiner time serves as a proxy for
bank staff time in compiling data and preparing for a CRA exam. Therefore, a CRA exam for a
bank with more than $5 billion in assets probably entails between 2.5 to 5 times the staff time as
a CRA exam of a bank with half a billion in assets. CRA exams are already streamlined for
institutions with assets between $250 and $500 million in assets.

Of course, regulations impose some costs on banks. NCRC believes, however, that an objective
cost-benefit analysis would reveal that the benefits massively outweigh the costs of large bank
CRA exams for both banks and the public at large. NCRC believes that the regulatory agencies,
themselves, must conduct a comprehensive cost-benefit analysis in considering their streamlined
proposal. NCRC contacted senior officials of the federal banking agencies, who told NCRC that
the agencies have not conducted cost-benefit analyses.

Smaller banks themselves complain much less frequently about CRA exams than they did a
number of years ago. Their lingering concern about unfair comparisons does not appear to be a
reality in most CRA exams. In the final analysis, burdens associated with large bank CRA
exams have more to do with perception than reality. In contrast, the benefits of large bank exams
are real, easily documented, and profound. Low- and moderate-income communities have
access to billions of dollars in capital and credit, which would likely disappear as the NCRC
analysis above suggests. Banks themselves have realized substantial amounts of profits as CRA
exams have motivated them to find safe and sound lending, investing, and branching
opportunities in low- and moderate-income communities.

Finally, it is strange that the federal agencies are proposing to considerably streamline CRA
exams for a large segment of banks when the banks themselves do not place CRA at the top of
their list of "burdens." According to the federal agency web site regarding the Economic Growth
and Regulatory Paperwork Reduction Act, banks regard the Bank Secrecy Act (BSA) and
Currency Transaction Reports as the "most burdensome regulations for the banking community."
Banker "outreach" meetings suggest that the "cost of compliance is high…(the BSA regulations)
are ineffective…and overly complex."

Also, high on the list for burden was the "Know Your
Customer" requirements of the USA Patriot Act.8 In contrast to the BSA regulations, the CRA
regulations are quite effective and not overly complex. The CRA regulations are the wrong
regulations to savage by a proposed streamlining.

Inadequate Predatory Lending Standard Must Be Replaced

The proposed CRA changes contain an anti-predatory screen that will actually perpetuate
abusive lending. The proposed standard states that loans based on the foreclosure value of the
collateral, instead of the ability of the borrower to repay, can result in downgrades in CRA
ratings. The asset-based standard falls short because it will not cover many instances of
predatory lending.

For example, abusive lending would not result in lower CRA ratings when it
strips equity without leading to delinquency or foreclosure. In other words, borrowers can have
the necessary income to afford monthly payments, but they are still losing wealth as a result of a
lender's excessive fees or unnecessary products. By shielding banks from the consequences of
abusive lending, the proposed standard will frustrate CRA's statutory requirement that banks
serve low- and moderate-income communities consistent with safety and soundness.

CRA exams will allow abusive lending if they contain the proposed anti-predatory standard that
does not address the problems of packing fees into mortgage loans, high prepayment penalties,
loan flipping, mandatory arbitration, and other numerous abuses.

Rigorous fair lending audits and severe penalties on CRA exams for abusive lending are necessary in order to ensure that the
new minority homeowners served by the Administration are protected, but the proposed
predatory lending standard will not provide the necessary protections. In addition, an antipredatory
standard must apply to all loans made by the bank and all of its affiliates, not just realestate
secured loans issued by the bank in its "assessment area" as proposed by the agencies.

A comprehensive anti-predatory standard must also apply to payday loans and other non-secured
consumer loans. In addition, the anti-predatory standard must apply to secondary market activity
including purchasing loans or securitizing loans for others. Many large banks are no longer
originating subprime loans, but they are purchasing large quantities of high cost loans. If the
banks do not have sufficient due diligence procedures, they will facilitate abusive lending, and
thus fail on their CRA obligation to respond to credit needs in a safe and sound manner.
Data analysis reinforces the need for a strong anti-predatory lending standard that applies to
secondary market activity as well as primary market activity.

Using CRA Wiz produced by PCI Services, NCRC calculates that depository institutions and their affiliates made 597,861
8 See http://www.EGRPRA.Gov and go to Banker Outreach Meetings. subprime and
manufactured home loans during 2002, the latest year for which HMDA data is
available. Depository institutions and their affiliates issued 38 percent of all the subprime and
manufactured home loans in 2002. They purchased a much larger percentage of total subprime
and manufactured home loans. They purchased 51.9 percent of all high cost loans or 143,288
high cost loans. In total, depository institutions and their affiliates originated and purchased
more than 740,000 high cost loans during 2002. The federal banking agencies have a solemn
responsibility to ensure that the massive amount of subprime and manufactured home loans
made by the lenders they regulate are free of abuses.

Multiple Screens Need to Penalize Predatory Loans

NCRC recommends a series of anti-predatory screens that will capture a significantly greater
amount of abusive loans than the standard imported from the OCC preemption ruling. The three
screens recommended by NCRC are: 1) a revised OCC standard, 2) a HOEPA screen, and 3) a
screen adapted from HUD's affordable housing rules for the Government Sponsored Enterprises
(GSEs).

The OCC standard can be retained, but only as one of a number of screens. The OCC standard
should also be clarified to indicate that default and delinquency are not the only circumstances
indicating lending beyond repayment ability. As the standard is constructed currently, it could
appear that only lending based "predominantly" on the foreclosure value of the collateral would
be considered abusive since the reference to foreclosure is part of the same sentence as the term
beyond repayment ability. The standard could be revised as follows:

"Lending beyond repayment ability is abusive.

The impacts of lending beyond repayment ability include, but are
not limited to, borrower financial distress, delinquency, and foreclosure. Banks and thrifts
engaging in this type of lending shall be penalized on their CRA exams.

Because the OCC standard does address the numerous abuses associated with predatory lending,
NCRC recommends a second screen of applying the prohibitions and limitations contained in the
Home Ownership and Equity Protection Act (HOEPA) to a wider subset of loans. The federal
agencies should use the Federal Reserve Board's new price reporting trigger in the Home
Mortgage Disclosure Act (HMDA) regulation. Under the HMDA regulation, lenders must report
price information for first liens with Annual Percentage Rates (APRs) that are three percentage
points above Treasury rates and for second liens with APRs that are five percentage points above
Treasury rates.

The Federal Reserve Board estimates that these triggers cover virtually all
subprime loans.9 If the regulators applied a screen to these loans, they would be scrutinizing the
great majority of subprime loans. If loans captured by the HMDA price trigger violated
HOEPA, the lender would be penalized on CRA exams.
HOEPA contains important, but not sufficient protections. Under HOEPA, lenders cannot make
high cost loans that exceed repayment ability, have short term balloon payments under five
years, contain negative amortization, or have prepayment penalties beyond five years. The
HOEPA limits and prohibitions must be supplemented by another screen since HOEPA does not
afford comprehensive protections.

A third screen to catch predatory loans is the prohibitions that the Department of Housing and
Urban Development (HUD) imposed on the Government Sponsored Enterprises (GSEs) in the
affordable housing goal rule. The GSEs cannot count loans towards their affordable housing
goals if the loans have points and fees of more than 5 percent of the loan amount, if the loans
contain single premium credit insurance, or if the loans are subprime and are made to borrowers
creditworthy for prime loans.10 Since HUD believed it had the authority to impose these
prohibitions on the GSEs in the affordable housing goal ruling, then the federal banking agencies
can surely penalize lenders for making loans with these features.

NCRC believes that the federal agencies should apply HUD's prohibitions to all loans (prime or subprime). At the very least,
the HMDA price trigger should be used to apply the HUD limits to the great majority of
subprime loans.

Since HUD implemented the current affordable housing goals in 2000, the GSEs have made
additional promises to prohibit abusive features in the loans they purchase. Recently, both GSEs
have pledged not to purchase loans with mandatory arbitration clauses and loans that apply
prepayment penalties for a period of three years after loan origination. In the 2000 affordable
housing goal rule, HUD's list of prohibitions included those that the GSE's had been voluntarily
adhering to previously. Likewise, the GSE subsequent voluntary pledges regarding mandatory
arbitration and prepayment penalties should inspire the federal banking agencies to make their
proposed anti-predatory standard more specific and comprehensive.

9 Memo from Division of Consumer and Community Affairs to Board of Governors of the Federal Reserve Board,
January 16, 2002; see page 60 saying that a large sample of subprime loans indicates that a "cutoff of 3 percentage
points or more above the comparable treasury security would cover about 98 percent of the subprime loans that are
first liens." A cutoff of 5 percentage points would cover 95 percent of subprime second lien loans.
10 Federal Register, Vol. 65, No. 211, October 31, 2000, pp.65070-65071.

Threshold Levels for Penalties on CRA Exams

The proposed anti-predatory standard conceivably allows a significant amount of predatory
lending before the bank suffers a reduced CRA rating and other penalties. The standard requires
that the lender must engage in a pattern and practice of abusive lending. Penalties are not
automatic, but depend on the strength of the evidence and whether the bank promises to take any
corrective action.

The federal agencies also do not regularly conduct a fair lending review
probing for discriminatory or abusive lending while they conduct a CRA exam. In sum, the antipredatory
standard makes it quite possible that a bank can engage in large scale abusive lending
that is scrutinized on an infrequent basis. Furthermore, if the bank is caught making predatory
loans, the bank can escape penalties if it suddenly commits to a plan for corrective action (future
commitments to corrective plans can avoid penalties per the CRA regulation).

NCRC believes that penalties must be swift and certain when the evidence indicates that a bank
has made a significant amount of predatory lending. A pattern and practice standard must not
exclude circumstances in which the abusive lending was not due to the bank's policies but due to
an operational issue that the bank did not address. For example, if a bank did not stop either a
rogue broker or bank employee from engaging in abusive lending over a significant time period,
the bank must be penalized on a CRA exam. In this case, the bank was negligent and did not halt
predatory lending, even in cases in which it may have had a policy prohibiting abusive lending.

While the predatory lending may not have been caused by an institutional practice or policy, it
nevertheless was unresponsive to legitimate credit needs by stripping wealth instead of enabling
borrowers to build wealth. The predatory lending, unresponsive to credit needs, must be
therefore penalized on CRA exams through a lower rating.

If a significant amount of abusive loans occurs in one assessment area, it must cause the bank to
receive a failing rating of at least needs-to-improve in that assessment area. A significant
amount could be related to an institutional policy and/or to an operational issue such as a rogue
broker or loan officer. Assuming a rogue broker or bank employee makes one abusive loan each
day, the number of predatory loans can easily equal 30 or more per month. If a bank does not
take action after two months (a time period in which between 50 to 60 abusive loans are made),
the bank has enabled too much wealth stripping and abuses.

The threshold for failing in one assessment area should certainly be triggered by 50 loans.
The threshold could even be less if it can be demonstrated that the abuses were not only caused by an operational issue but also an
institutional policy. For example, if a bank has a policy requiring loan officers to charge more
than 5 percent of fees on every loan, then even a few loans with usurious fees should result in a
failed CRA rating in the one assessment area.

The bank must fail its entire CRA exam if the bank exceeds the threshold for making abusive
loans in two or more assessment areas. When predatory lending occurs in two or more
assessment areas, it is most likely that the predatory lending is due to an institutional policy or
practice. Failure of the CRA exam must therefore be swift and certain.

Banks and thrifts must understand that CRA exams and fair lending reviews probing for
predatory lending will not be infrequent, but regular and stringent. The Office of the
Comptroller of the Currency (OCC) has adopted a risk-based approach to examination; the
OCC's counterparts are in the process of adopting this method as well. The difficulty with a
risk-based approach is that it can lead to infrequent fair lending exams and apply arbitrary and
inconsistent criteria for deciding which institutions are to receive rigorous fair lending reviews.

A more reliable approach is applying a rigorous anti-predatory standard for CRA exams in
deciding the frequency of fair lending reviews. If a regulator believes (or has evidence via CRA
comments from community groups or other parties) that an institution is making high cost loans
that are close to the limits established by the anti-predatory standard, then the bank must be
subject to regular fair lending reviews that occur no less frequently than the CRA exams. In
addition, federal agencies should conduct fair lending reviews coinciding with CRA exams for
those lenders with a majority of their loans exceeding the HMDA price trigger or appear on

HUD's list of subprime or manufactured home lenders. HUD currently categorizes lenders as
subprime or manufactured home loan specialists if more than half of their loans are subprime or
manufactured home loans.11 Under a risk-based approach to examination, it would seem prudent
that lenders should qualify for regular fair lending reviews on the same cycle as CRA exams
when the majority of their loans are high cost and non-prime.

Illegal Practices Standard Must be Inclusive

The proposed changes also stipulate that lenders can be penalized if they make loans violating a
number of federal fair lending and consumer protection statutes. These include the Home
Ownership and Equity Protection Act (HOEPA), the Equal Credit Opportunity Act (ECOA), and
the Truth in Lending Act (TILA). The proposal states that specific violations of these laws, such
as the right of rescission under TILA will result in penalties. NCRC believes that violations of
all substantive protections, not just violations of a few provisions, in the fair lending and
11 HUD updates its list of subprime and manufactured home lenders on an annual basis. It is available on HUD's
web page (http://www.huduser.org/datasets/manu.html). If more than 50% of an institution's loans are subprime or
manufactured home loans, the lender is classified as a subprime or manufactured home lender. The HUD list is
widely used by researchers in government agencies and elsewhere.

consumer protection laws must result in penalties. For example, if the bank consistently
understates the Annual Percentage Rate (APR), this violation of TILA must also be penalized.
Understating price can end up costing consumers several thousands of dollars over the course of
a loan. This type of deception fails to meet credit needs as mandated by CRA since equity
stripping is occurring.

The federal agencies should clarify that all substantive protections under the fair lending laws
and consumer protection laws can result in penalties. Moreover, the preamble to the proposed
rule states that violations of state law, where applicable, entail penalties. This reference to state
law must be included in the proposed regulation in order to elevate its importance. Finally, the
same thresholds as discussed above for the anti-predatory standard would apply to the cases of
violating federal and state laws

"No Documentation" Loans Must Not be Tolerated

The proposal contains a glaring inconsistency between insisting that loans cannot be
unaffordable but then allows institutions to make loans without documenting borrower income.
How can a lender ensure that it is not making a loan that is unaffordable if it does not document
borrower income? In December of 2001, the Federal Reserve Board amended Regulation Z
(implementing HOEPA) to stipulate that lenders are presumed not to consider repayment ability
if they do not verify and document borrowers' incomes.

This HOEPA standard must apply to all loans, not just high cost loans. It is quite conceivable
that lenders making prime loans will issue unaffordable loans if they do not verify borrower
incomes. If the federal agencies allow a double standard to apply to high cost loans covered by
HOEPA and loans not covered by HOEPA (which includes many subprime loans as well as
prime loans), the proposed CRA regulation can end up facilitating a significant amount of
abusive lending.

Affiliate and Assessment Area Procedures Must be Changed to Protect Against Abusive Lending
The agencies' proposals fail to close gaping loopholes in the CRA regulation regarding affiliates
and assessment areas. Banks can still elect to include affiliates on CRA exams at their option.
They can thus manipulate their CRA exams by excluding affiliates not serving low- and
moderate-income borrowers and excluding affiliates engaged in predatory lending. The current
regulations allow banks to hide their true performance in serving community credit needs by
shifting undesirable loans to their affiliates and subsidiaries. The game playing with affiliates
and subsidiaries will end only if the federal agencies require that all affiliates and subsidiaries be
included on exams.12

The proposed changes do not address the need to update assessment areas to include
geographical areas beyond bank branches. Many banks make considerable portions of their
loans beyond their branches; this non-branch lending activity is not scrutinized by CRA exams.
NCRC has previously proposed that a metropolitan or rural area must be a distinct assessment
area if a bank has a market share of loans in the area of more than one half of one percent. In
any case, assessment areas must cover the great majority of a bank's loans. The Joint Housing
Center at Harvard University estimates that depository institutions among the top 25 lenders
originate only 25 percent and 33 percent of their home purchase and refinance loans,
respectively, in their assessment areas.13

With such low percentages of bank loans in assessment
areas, the federal agencies are simply unable to assess if banks are meeting credit needs in the
communities in which they are chartered to do business.

At the very least, the federal agencies must update their procedures for considering lending
beyond assessment areas. According to the current Interagency Question and Answer document
on CRA, federal agencies will provide positive consideration for lending beyond assessment
areas if a bank has satisfied credit needs in its assessment areas. It would be woefully
inconsistent if the agencies do not penalize banks for abusive loans beyond assessment areas
made by either the banks or affiliates. If a bank is to receive positive consideration for safe and
sound lending beyond the assessment area, it must also receive negative consideration for
abusive lending beyond the assessment area. Only such an evenhanded approach can guarantee
that banks are meeting legitimate credit needs wherever they are doing business.

Existing Examinations of Subprime Lenders are Inadequate

The following case studies illustrate that the existing CRA exam scrutiny of subprime lenders is
cursory and grossly inadequate most of the time. The current CRA exam practices illustrate the
desperate need for a strong anti-predatory standard and rigorous fair lending reviews of
depository institutions that are either subprime lenders or have affiliates that are subprime
lenders. In only one instance reviewed by NCRC, does the fair lending review explicitly test for
abusive lending. In only a few cases, does the CRA exam document the extent of subprime
12 The CRA regulations define a subsidiary as an affiliate. Hence, banks can also include subsidiaries at their option
on CRA exams. NCRC comments regarding affiliates apply to subsidiaries. CRA exams must require the inclusion
of all affiliates and subsidiaries.

13 The Joint Center for Housing Studies at Harvard University, The 25th Anniversary of the Community Reinvestment
Act: Access to Capitol in an Evolving Financial Services System, March 2002, p.29.

lending by the bank or its affiliate. The CRA examiner should always carefully document the
nature and extent of subprime lending in providing a rationale of why or why not a fair lending
review was conducted.

As stated above, NCRC believes that the federal agencies must conduct fair lending reviews at
the same frequency of CRA exams for lenders on HUD's list of subprime or manufactured home
lenders and/or lenders with a majority of their loans exceeding the HMDA price trigger. Finally,
the exams of major subprime lenders often include unjustified assessment areas covering only a
tiny fraction of the institutions' loans. The current application of the CRA regulations is failing
to ensure that credit needs are met in a safe and sound manner.

Immediately below are reviews of exams of major depository institutions that are subprime.
After their CRA exams, Superior Bank, FSB and Conseco Bank failed. These are two clear
cases of regulatory failure, caused in part by inadequate scrutiny on CRA exams.

CRA Exams of Subprime Depository Institutions
Superior Bank, FSB

Before the spectacular failure of Superior Bank due to unsafe lending, the Office of Thrift
Supervision (OTS) conducted a CRA exam of Superior that clearly illustrates regulatory failure
to catch problems before they explode.
Based in Oakbrook Terrace, Illinois, Superior Bank, FSB was a major thrift and subprime
lending institution. The OTS CRA exam of September 1999 reports that this institution of $1.8
billion in assets made or purchased more than 39,000 home loans from January of 1997 through
July of 1999. The exam states that Superior has "flexible underwriting guidelines, which benefit
low- and moderate-income persons and families." It adds that "Superior's lending business
focuses on lending to "A-," "B", or "C" borrowers, who are borrowers that have varying degrees
of credit history problems."

It describes a "Universal Mortgage Product" in the following paragraph:
The Universal Mortgage Product is designed specifically to improve a borrower's day to
day cash flow by consolidating their various debts and lowering their combined payment
and overall cost of credit…The borrower is provided 26 payment reduction vouchers at
closing to use at any time during the term of the loan (when used, deferred payments are
not required to be paid at maturity). No private mortgage insurance is required and all
National Community Reinvestment Coalition * 202-628-8866 * http://www.ncrc.org 23
closing costs may be included in loan amount rather than paid at closing. If the borrower
needs additional funds after consummation, any principal amount previously paid down
on the current mortgage can be borrowed without any additional closing fees. Also, as
long as the borrower meets certain criteria, he (or she) may change the payment amount
(to as low as interest only payments) at anytime throughout the loan term. This product
does not require tax or insurance escrow payments.

The examiner accepts at face value that these terms are flexible and beneficial for low- and
moderate-income borrowers. However, since all closing costs can be financed, what percentage
of the loan amount are fees that are financed? In addition, the borrower could borrow against
principal already paid. But did Superior provide borrowers with advice about possible negative
amortization and prolonged indebtedness? Also, the Universal Mortgage Product did not require
tax or escrow payments. But did Superior explain to the borrower that tax payments are still
required to be made outside of the loan payments? Based on the eventual fate of Superior, it is
reasonable to assume that neither the OTS CRA or safety and soundness examiners answered
these basic questions about Superior's lending.

Superior's Universal Mortgage product appears to be appropriately named. In 1998, Superior
made 16 percent of its 15,338 refinance and home purchase loans to Blacks while OTS-regulated
subprime lenders made nearly10 percent of their loans to Blacks and OTS-regulated prime
lenders made about 2.5 percent of their loans to Blacks. Likewise, Superior made 37 percent of
its single-family loans to low- and moderate-income borrowers while OTS-regulated subprime
lenders made 22 percent and OTS-regulated prime lenders made 23 percent according to the
HMDAWare. software produced by Compliance Technologies. The CRA exam likewise states
that Superior had an "excellent penetration among the various types of borrowers in the
assessment area, particularly low- and moderate-income borrowers." While Superior's
mortgages may have been universally available (or at least more so than other OTS-regulated
institutions), the OTS did not probe in its CRA exam if Superior was targeting minorities and
low- and moderate-income borrowers with unsafe and abusive loans in violation of the Equal
Credit Opportunity Act and the Fair Housing Act.

The shortcomings of Superior's CRA exam extend to an unjustified restriction of considering
only one assessment area, parts of the Chicago metropolitan area, that accounts for only 1.3
percent of the thrift's loans. The explanation for not examining other areas was that Superior is a
" large, wholesale, nationwide lender." Examining a geographical area with such a small
percentage of the thrift's loans leaves open the possibility that questionable practices elsewhere
are not being examined. In addition, even the OTS examiner implicitly admitted the inadequacy
of examining the thrift in only one assessment area. The examiner gave Superior "low
satisfactory" ratings on its lending, investment, and service tests, in part because of the
" marginal" level of activity in the thrift's assessment area.

Provident Bank – Federal Reserve Board

In its July 12, 1999 Community Reinvestment Act (CRA) performance evaluation, Provident
Bank received a satisfactory rating from the Federal Reserve Bank of Cleveland. At the time, the
CRA examiners indicated that, among other things, Provident had "an adequate level of lending
within the bank's assessment area" and used "innovative and flexible lending products" to meet
the needs of its borrowers. Curiously enough, in reaching these conclusions Board examiners
failed to scrutinize or even note the fact that Provident Bank is primarily engaged in subprime
mortgage lending, having been clearly identified as a subprime lending institution by the U.S.
Department of Housing and Urban Development.14

Federal Reserve examiners made no mention of the interest rates or terms and conditions
associated with Provident's mortgage loans. In addition, Board examiners brushed aside the fact
that Provident made only 41% of its loans within its assessment area, considering Provident's
loan volume within its assessment area adequate because the bank simply made a "strategic
decision to generate loans nationally through its broker network and other forms of delivery
systems." Interestingly, the CRA examiners did not consider the fact that unscrupulous
mortgage brokers have increasingly been cited as key culprits in perpetrating predatory mortgage
lending practices on unsuspecting and unsophisticated low- and moderate-income, minority, and
elderly borrowers.

While examiners noted "inconsistencies in the bank's credit policies, procedures, and application
of the policies and procedures" during the Board's fair lending review, there is little discussion
of what these inconsistencies were or how they were applied. Examiners did note "no apparent
disparate or discriminatory effect based on any prohibited bases," but also commented that
Provident "was required to adopt new policies and procedures, implement an enhanced fair
lending training program, and establish fair lending review procedures." Clearly the examiners
raise more questions than they answer about the nature of the bank's treatment of protected
classes.

Given the many questions raised by Provident's inadequate CRA exam, NCRC pulled
Provident's March 30, 1999 and March 29, 2000 10-K statements on file with the Securities and
14 See Randall M. Scheessele, 1998 HMDA Highlights, Housing Finance Working Paper No. 9, Office of Policy
Development and Research, HUD, October 1999.

Exchange Commission to learn a little more about the bank's so-called "innovative and flexible
lending" and its efforts to truly meet the credit needs of low- and moderate-income communities.
We found that Provident had significantly increased its origination, securitization, and sale of
nonconforming residential loans, from a total of $264 million in 1996 to $2.3 billion by 1999.
From 1998 to 1999 alone, Provident increased its pool of nonconforming loan originations over
100 percent, from $1.1 billion to $2.3 billion. As Provident itself declared, its national
nonconforming mortgage operations "continue to represent a major business opportunity." But
the CRA exam and fair lending review did not scrutinize Provident's rapidly increasing subprime
lending operations.

As far as a reader of the Provident CRA exam can determine, Federal Reserve examiners failed
to take readily accessible information from 10-K statements into account during Provident's
CRA exam and, as a result, Provident's subprime and potentially abusive and discriminatory
lending activity continues to grow without real regulatory scrutiny. In fact, Provident's March
29, 2000 10-K statement, issued after the Board's July 1999 CRA performance evaluation,
reveals that the bank has increased its origination and sale of subprime loans.

Conseco Bank, Inc. – Federal Deposit Insurance Corporation

In its December 8, 1999 Community Reinvestment Act (CRA) performance evaluation, Conseco
Bank, Inc. received a satisfactory rating. At that time, FDIC examiners indicated that, among
other things, Conseco had "…an excellent distribution to individuals of different income levels,
including low- and moderate-income borrowers" and that bank management had "…identified
several meaningful ways to address the bank's CRA obligations." Curiously enough, in reaching
these conclusions, examiners failed to scrutinize or even note the fact that Conseco Bank is
significantly engaged in subprime manufactured home lending, having been clearly identified as
such an institution by the U.S. Department of Housing and Urban Development.15
FDIC examiners made no mention of interest rates or terms and conditions associated with
Conseco's mortgage loans. In addition, examiners seemed to brush aside the fact that Conseco
made just two percent of its residential mortgage loans in 1998 and the first three-quarters of
1999 to borrowers within its assessment area.

The examiner used the streamlined small bank exam procedures for Conseco's exam although
Conseco had assets above the $250 million threshold for small bank exams for about half of the
15 See HUD's 1999 list of subprime and manufactured home lenders.
National Community Reinvestment Coalition * 202-628-8866 * http://www.ncrc.org 26
exam period. Inadequate examination procedures most likely contributed to Conseco's failure
since its regulator, the FDIC, was not ensuring that credit needs were met in a safe and sound
manner.

OCC Exam of Chase Manhattan Bank USA: Completely Inadequate

More recent large bank exams also reveal inadequate regulatory scrutiny. The Office of the
Comptroller of the Currency's most current exam of Chase Manhattan Bank USA was in March
of 2003. Chase Manhattan Bank USA is a non-traditional lender that closed its only branch in
New Castle County, Delaware. It appears not to make loans directly, but "books" loans made by
an affiliate, Chase Manhattan Mortgage Corporation.

The OCC applied an unjustifiably narrow assessment area to Chase Manhattan Bank USA. The
only assessment area is New Castle County, Delaware although the bank closed its one branch in
New Castle County.

The bank made less than one percent of its home and small business loans in the New Castle
County assessment area. In other words, the exam only considered 1 percent of the bank's loans.
The examiner believes that the bank did well in New Castle County in reaching low- and
moderate-income borrowers and communities. The bank received an Outstanding on its lending
test. Is this really that hard to do, considering that the bank can concentrate its efforts in New
Castle County, and not worry too much about meeting the needs of low- and moderate-income
borrowers in the geographical areas that receive 99 percent of the bank's loans!

Despite the fact that HUD has classified Chase USA as a subprime lender, the OCC declined to
conduct a fair lending review to coincide with the CRA exam. The CRA exam states that based
on "an analysis of (public comments and consumer complaint information), the OCC decided
that a comprehensive fair lending examination would not need to be conducted in connection
(with the CRA exam)."

Additional evidence suggesting the need for a fair lending review is a recent prospectus
statement regarding loans sold by Chase Manhattan Bank USA and Chase Manhattan
Mortgage.16 In this prospectus, Chase was selling more than 2,000 high-interest loans to
investors. Thirty percent of the loans were "stated" income loans, meaning that Chase did not

16 See July 11, 2003 Mortgage Loan Assets-Backed Certificates, Series 2003-6 on the SEC web page,
http://www.sec.gov.

verify income levels. As stated above, the Federal Reserve Board has disallowed "no income"
verification loans for high cost loans covered by HOEPA. While most of the loans in Chase's
prospectus were not HOEPA loans, the great majority of these loans were subprime. NCRC
believes that the agencies should not allow "no income" verification loans. Certainly, subprime
lenders making a significant number of these loans ought to be subjected to rigorous fair lending
reviews that assess whether borrowers can afford the loans.

Cursory CRA Exams of Additional Subprime Depository Institutions

Additional CRA exams of subprime depository institutions expose more shortcomings. For a
number of years, the Department of Housing and Urban Development has classified Key Bank,
USA, NA; Bank One, Michigan; and Travelers Bank and Trust as either subprime or
manufactured home lenders. Despite HUD's classification of these lenders as high cost lenders,
their CRA exams were completely lacking except for Travelers.

In an OCC exam of Key Bank, USA during 2000, the OCC does not even discuss the subprime
nature of Key Bank, USA's lending activities. Under the product innovation and flexibility
section, the exam mentions flexible underwriting including "higher debt ratios" and the "use of
alternative credit histories to enable more low- and moderate-income applicants to qualify." The
fair lending review does not scrutinize if these "flexible" underwriting criteria are abusive.
Instead the fair lending review focuses on a traditional exam for a prime lender assessing if
minority applicants are denied more frequently than whites. Likewise, a Federal Reserve exam
conducted in 2001 of Bank One Michigan does not mention that HUD classified Bank One as a
manufactured home lender. The exam cites "innovative" products for borrowers without
established credit histories, but does not mention the interest rates and fees associated with those
products. The exam also lauds Bank One for excellent penetration of loans to low- and
moderate-income borrowers and census tracts. The fair lending review, however, does not assess
the loan terms and conditions available to low- and moderate-income borrowers and census
tracts.

Travelers Bank and Trust, Office of Thrift Supervision, 2001

The OTS' CRA exam of Traveler's is the only CRA exam reviewed by NCRC that approaches a
suitable review of a subprime lender. While shortcomings in the evaluation occurred, this CRA
exam included a fair lending review testing for abusive lending and attempted to compensate for
an inadequate assessment area.
Owned by Citigroup, Travelers Bank and Trust is a non-traditional lending institution
National Community Reinvestment Coalition * 202-628-8866 * http://www.ncrc.org 28
headquarted in New Castle County, Delaware. Travelers does not have a traditional branch
network. Instead more than 100,000 agents of Primerica Financial Services make loans as well
as selling insurance policies.

In order to ensure that Travelers subprime loans were not abusive, the CRA exam included a
rigorous fair lending review. The fair lending review consisted of statistical sampling and
qualitative reviews of loan files to detect if the lender was fee packing (loading up the loan with
high fees without the informed consent of the borrower). The review also probed for evidence of
equity stripping, flipping, abusive foreclosure procedures, or targeting minorities and other
protected classes with loans containing "less favorable terms than those which the borrower
qualifies."

The fair lending review also consisted of interviews with community groups.
Based on the fair lending review, the examiner concluded that Travelers subprime loans had
lower settlement fees than those associated with conventional loans and lower rates than those
assessed by other national subprime lenders. In addition, the examiner documented that
Travelers "Tangible Net Benefit Initiative" featured refinance loans that offered customers lower
rates than their previous loans, or provided customers with loans that did not include abusive
features such as large balloon payments. Without judging the validity of these conclusions, it is
clear that this CRA exam reviewed critical issues associated with subprime lending and thus
provides the general public with more assurances than the other CRA exams reviewed.

Because Travelers is a non-traditional lender, its assessment area is artificially restricted to its
two offices in the Wilmington, Delaware metropolitan area. Travelers made only 1 percent of its
loans in Wilmington. Per the Interagency Question and Answer guidance on CRA, the OTS
declared that this tiny portion of lending in Wilmington would not be sufficient for an institution
the size of Travelers. Since Travelers met credit needs in Wilmington, the OTS examined
Travelers performance in nine other major metropolitan areas in "reaching an overall evaluation"
of Travelers performance under the lending test.
The lending levels in the nine other areas still
added up to a minority of Travelers' overall lending levels; each metropolitan area contained
about 3 percent of the thrift's loans. NCRC would have preferred the OTS to select areas
constituting the majority of the thrifts' loans. At least, the OTS examined a considerably greater
portion of loans than other CRA exams of non-traditional lenders, including the OCC's
inadequate exam of Chase Bank, USA.

The combination of a comprehensive fair lending review and expanded assessment area
procedures in the Travelers' CRA exam gives the public some confidence that the regulator is
assessing whether the thrift is meeting community credit needs in a safe and sound manner.
Moreover, the fair lending review, which looked for equity stripping, flipping, and other abuses,
National Community Reinvestment Coalition * 202-628-8866 * http://www.ncrc.org 29
suggests that the federal agencies can surely construct a comprehensive anti-predatory standard
for all of their CRA exams.

CRA Exams Including Subprime Affiliates

As well as scrutinizing CRA exams of subprime depository institutions, NCRC looked at several
exams of prime lenders that had subprime affiliates. NCRC found that none of these exams
contained an adequate fair lending review. If a lender elects to include a subprime affiliate, the
regulator must conduct a fair lending review probing for abusive lending. Otherwise abusive
lending can easily count on the CRA exam. If the regulators followed NCRC's
recommendations, the affiliates would automatically be included on the exams, triggering
automatic rigorous fair lending reviews. At the very least, optional inclusion of a subprime
affiliate must trigger a comprehensive fair lending review of the lender and its affiliate.
NCRC's review identified the following CRA exams including subprime affiliates:
First Union, NA, OCC, September 2000 - subprime affiliate Money Store
National City affiliates, several OCC exams in 2000 – subprime lender Altegra Credit Company
Citibank, NA (OCC, 2000) and Citibank FSB (2001) – subprime Citifinancial
Chase Manhattan Bank, Federal Reserve, 2001& 2003 – subprime Chase Manhattan Bank USA
Bank One, NA, OH, OCC, March 2000 – subprime Banc One Financial Services
In the First Union exam, the examiner states that "neither the end of the Money Store's
operations nor its activities during its tenure as the bank's subsidiary" had an "adverse" impact
on First Union's CRA performance. Given the reputation of the Money Store and its closure, it
would be reasonable to expect a more thorough investigation. However, the seven-sentence
summary of the fair lending review does not indicate if the review included any anti-predatory
lending scrutiny.

In the case of the Citibank exams, the fair lending reviews did not probe into predatory lending
issues. The examiners reviewed comment letters and concerns expressed by community groups
during Citigroup's acquisition of the Associates. The examiners then listed a series of reforms
promised by Citibank. The examiners appeared to rely on Citibank's responses to the
community group concerns without independently confirming via fair lending exams.

The Citibank FSB exam states that affiliate subprime lending accounts for about 7 percent of the
total lending and the Citibank NA exam suggests that subprime affiliate lending accounts for a
much smaller percent of total bank lending. Although the subprime lending accounted for a
National Community Reinvestment Coalition * 202-628-8866 * http://www.ncrc.org 30
small percentage of the total, these exams followed the heels of the controversial acquisition of
Associates. The exams provided an important and missed opportunity to determine if Citigroup
was implementing its newly announced reforms. Moreover, the OTS and OCC combined
resources on the fair lending review of Citigroup's thrift and bank. With augmented resources,
the regulators still failed to conduct a strong far lending review.

The only positive element of these exams is that they attempted to document the quantity of
subprime lending in an effort to assess how they would treat the subprime lending. In addition,
the Citibank FSB exam states that the affiliate lending did not increase the percentage of loans to
low- and moderate-income borrowers and census tracts. This is at least an honest discussion of
the impacts of affiliates on the thrift's overall CRA and lending test rating. At the same time,
however, the examiner states that more than 18,000 Citibank FSB loans were "no income"
documentation loans lacking information about borrower incomes. This magnifies the
inadequacy of the fair lending review.

CRA Exams Excluding Subprime Affiliates

NCRC has also identified a number of depository institutions excluding their subprime affiliates
from their CRA exams. Per the current CRA regulation, the regulators will not conduct a fair
lending review of the excluded affiliates. The nature of these affiliates' subprime lending
remains hidden from scrutiny.

NCRC identified a number of CRA exams of large lenders that excluded subprime affiliates.
These include:

Bank of America – OCC, 2001 – excludes Nationscredit Financial Services
Charter One Bank, FSB – OTS, 2001 – excludes Charter One Credit Corporation
KeyBank, USA, NA – OCC, 2000 – excludes Champion Mortgage Company
National City – OCC exams of prime affiliates in 2000 – excludes First Franklin Financial
Washington Mutual – OTS exams of 2000 – excludes Long Beach Mortgage Company

Since these exams occurred between 2000 and 2001, NCRC conducted data analysis with the
year 2000 and found that the excluded affiliates made a significant amount of loans. In
KeyBank's case, Champion made 14,296 single-family loans in 2000 and 4,475 loans to lowand
moderate-income (LMI) borrowers. National City's First Franklin Financial Corp. made
35,915 single family loans overall and 9,741 to LMI borrowers. While National City excluded
First Franklin from its CRA exams, it included the subprime affiliate, Altegra Credit Corporation
that made a much smaller 1,163 loans during 2000.

Bank of America's and Washington Mutual's excluded subprime affiliates also made high
volumes of loans. Washington Mutual's subprime affiliate, Long Beach Mortgage Company,
made 30,629 loans overall and 11,249 to LMI borrowers in 2000. Bank of America's
Nationscredit Financial Services made almost 18,000 single-family loans to LMI borrowers and
39,197 loans overall during 2000.
The current regulatory procedure of optional treatment of affiliates results in the exclusion of
thousands of subprime loans from regulatory scrutiny. An anti-predatory standard is not
meaningful in this regime since it will not be applied to large-scale subprime lending operations.
Enhanced Data Disclosure: Good Proposal but Data Must be Utilized, Not Simply
Reported
The federal agencies propose that they will publicly report the specific census tract location of
small businesses receiving loans in addition to the current items in the CRA small business data
for each depository institution. This will improve the ability of the general public to determine if
banks are serving traditionally neglected neighborhoods with small business loans. Also the
regulators propose separately reporting purchases from loan originations on CRA exams and
separately reporting high cost lending (per the new HMDA data requirement starting with the
2004 data).
The positive aspects of the proposed data enhancements do not begin to make up for the
significant harm caused by the first two proposals. Furthermore, the federal agencies are not
utilizing the data enhancements in order to make CRA exams more rigorous. The agencies must
not merely report the new data on CRA exams, but must use the new data to provide less weight
on CRA exams to high cost loans than prime loans and assign less weight for purchases than
loan originations.
Use New Data to Weigh Prime Lending More Heavily on CRA Exams
Lending institutions must be encouraged to make as many prime loans as possible to LMI and
minority communities. NCRC believes that the prime lending market is not saturated; that is,
there are many more opportunities to lend to LMI and minority borrowers creditworthy for prime
loans. The subprime market suffers from widespread price discrimination and a segment of the
market is predatory. Evidence of price discrimination indicates that thousands of borrowers
stuck with subprime loans are creditworthy for prime loans. Analyzing ten large metropolitan
areas, NCRC's recent Broken Credit System report reveals that the portion of subprime loans in
National Community Reinvestment Coalition * 202-628-8866 * http://www.ncrc.org 32
neighborhoods increases as the number of minority and elderly residents increases, after
controlling for creditworthiness and housing stock characteristics. Another study conducted by a
Federal Reserve economist came to similar conclusions, looking at subprime lending in the
Philadelphia and Chicago metropolitan areas.17 The CRA regulations have an important role to
play in increasing prime lending for underserved customers and in cleaning up the subprime
industry.
The lending test's qualitative factors relating to innovative and flexible loan practices must be
further developed in the cases of subprime lenders. For example, consider the case of two
lenders of similar asset size and range of loan products operating in the same metropolitan area.
If one lender is making most of its loans with relatively high interest rates and fees to low- and
moderate-income communities and borrowers, it should not receive as high a CRA rating as the
other lender that is making mostly prime rate loans to these same communities. Clearly, the
lender that is making prime rate loans has been more effective in figuring out how to apply
flexible underwriting techniques to traditionally underserved borrowers.
NCRC recommends that any bank or thrift whose subprime lending exceeds a nominal amount
such as 5 percent of its total loan amount must have a separate prime and subprime lending test.
In particular if more than 5 percent of the bank's loans have APR's above the new HMDA
trigger, the bank must have separate lending tests. The subprime test would analyze loans with
APRs above the HMDA price trigger and the prime test would analyze loans below the price
trigger.
In other words, prime and subprime loans must be evaluated separately on CRA exams just like
home mortgage, refinance, and home improvement lending are currently. In order for a bank
that offers both prime and subprime lending to pass its lending test, it must receive at least a
satisfactory rating on the prime portion of its test. This proposal is similar to the requirement
that a bank must receive at least a low satisfactory on the lending test to pass its CRA exam. But
it would put more weight on the prime portion of the lending test since the lender would have to
perform in a satisfactory manner on that part of the test, not merely in a low satisfactory manner.
The prime and subprime lending test proposal is applicable even in situations when the great
majority of the lender's portfolio is subprime. Even if prime lending is just 25 percent of the
loan portfolio, the lender has to make efforts to ensure that a reasonable number of LMI
17 NCRC's Broken Credit System report is available via http://www.ncrc.org. Paul S. Calem, Kevin Gillen, and
Susan Wachter, The Neighborhood Distribution of Subprime Mortgage Lending, October 30, 2002. Available via
pcalem@frb.gov.
National Community Reinvestment Coalition * 202-628-8866 * http://www.ncrc.org 33
borrowers and residents of LMI neighborhoods receive the lender's prime loans. If the
distribution of prime loans to LMI borrowers and residents is poor, then the lender fails its CRA
exam, regardless of its performance on the subprime test. In the case of a lender with nearly 100
percent subprime lending (such as Travelers discussed above), the prime and subprime test
proposal cannot be readily implemented. Nevertheless, the regulatory agency in these cases
must subject a lender to a rigorous fair lending review that applies a stringent anti-predatory
standard. The Travelers CRA exam and fair lending review discussed above is a step in that
direction.
If federal agencies do not carefully consider how to evaluate prime and subprime lending, they
may unwittingly exacerbate the relative scarcity of prime loans and the saturation of subprime
loans in LMI and minority neighborhoods. A large body of research shows subprime lenders
tend to make a greater portion of their loans to LMI and minority borrowers and neighborhoods
than prime lenders. If the regulators continue to make no distinction between prime and
subprime loans on CRA exams, they will encourage banks to increase their subprime lending in
order to boost their percentage of loans to LMI borrowers and neighborhoods.
The lending patterns of Citigroup illustrate a likely trend if the federal agencies do not change
how they evaluate prime and subprime lending. In the year 2000 (which was just on the heels of
the Associates acquisition and also was a year covered by a number of exams of Citigroup
affiliates), Citigroup's prime affiliates made more than 77,000 single-family home loans to
borrowers in all income groups. Citigroup's subprime depository affiliate (Travelers) made
more than 20,000 loans. The non-depository subprime affiliates (various Citifinancial HMDA
reporters) issued 21,700 loans. Together, the subprime affiliates on CRA exams issued about
30,000 less loans than the prime affiliates. However, the recently acquired Associates affiliates
made more than 144,000 subprime loans during 2000.18
Citigroup included Citifinancial affiliates on its exams conducted on or near the year 2000. The
lender is likely to include the former Associates units in future exams. The total amount
subprime lending is likely to be much higher than the prime lending on future exams.
18 NCRC used CRA Wiz to analyze 2000 HMDA data. Citigroup prime affiliates that year included Citibank, NA;
Citibank, Nevada, NA; Citibank, FSB; and Citibank (New York State); and Citimortgage, Inc. Citigroup subprime
affiliates included Travelers Bank and Trust, FSB; various Citifinancial HMDA reporters; Associates Financial
Services; Associates Home Equity; and Associates Housing Finance. Many CRA exams occurred around 2000 –
Citibank (Nevada) NA in 1999; Citibank, FSB in 1999 & 2001; Citibank, NA in 2000; Citibank (NY state) in 2000;
Travelers in 2001.

While Citigroup is just one lender, it is the largest bank holding company in the United States in
terms of assets. It would be unfortunate indeed if the structure of CRA exams encouraged other
major lenders to emphasize subprime lending and thus narrow product choice further in LMI and
minority communities. It would be sadly ironic that a law passed to combat redlining ended up
intensifying reverse redlining and the dual lending market of predominantly high cost lending in
LMI and minority communities and prime lending in affluent neighborhoods. The federal
agencies have a critical opportunity that must not be missed to change the lending test so that
CRA remains an effective weapon against all forms of discrimination.
CRA Exam Consideration of Loan Purchases is Sloppy and Encourages Manipulation
NCRC recommends that CRA exams treat purchases in a manner similar to its recommendation
for treating subprime lending. More CRA points would be awarded to loan originations. For
example, a system could be established similar to HUD's affordable housing goals, in which
some activities are weighted more heavily. In the case of CRA exams, a point system could be
established in which loan originations count twice as much as purchases.
Alternatively, the lending test would consist of a series of tests. In the case of lenders making
both prime and subprime loans, the lending test would consist of separate prime and subprime
tests. Within these two separate tests, there would be a loan origination test and a loan purchase
test. A bank could not pass either its prime or subprime test if it did not pass its loan origination
test. Then, it could not pass its overall CRA exam if it did not score at least satisfactory on the
prime part of its test. In the case of banks making only prime or subprime loans, the lending test
would consist of two subparts – a loan origination and a loan purchase test.

When banks purchase many more loans to LMI census tracts and borrowers than they originate,
the fault lies less with the banks and more with the CRA regulatory regime. Purchasing loans is
useful but originating loans is the more difficult task, and should be counted more heavily on
exams. Moreover, the current regulatory treatment of purchased loans tempts banks to purchase
loans to LMI borrowers just before their CRA exams. In contrast, it is considerably more
difficult for banks to make a large amount of loans to LMI borrowers right before their CRA
exams. Lending, therefore, represents a more sustained commitment to serving credit needs than
purchasing. Finally, when banks skew their activities toward purchasing, the CRA regulations
have caused an unhealthy distortion of their CRA activities. The CRA regulations should
encourage a balanced mix of responding to credit needs, including residential and small business
lending, community development lending, and investing. When purchases count as much as loan
originations, CRA exams have and can continue to distort activity towards purchasing other
banks' loans.

An example of unbalanced activity is Bank of New York's 2000 home loan data and 2001 CRA
exam. During 2000, Bank of New York made just 13 loans to LMI census tracts but purchased
1,307 loans made in LMI census tracts. If the bank had not purchased these loans, it may have
failed its CRA exam. The Federal Reserve's CRA exam of 2001 gave the bank a Low
Satisfactory grade on the lending test, and stated that the home and small business lending
showed "poor responsiveness to the credit needs of the bank's assessment areas." At the time of
the CRA exam, Bank of New York had $74 billion in assets. Certainly, a bank of that size could
have put more effort in loan originations and in making loans to LMI borrowers and geographies.
Had purchases been weighted less than loan originations on the CRA exam, the bank most likely
would have put more effort into loan originations or else would have faced the likely possibility
of failing its exam. The examiner notes that the bank made efforts to improve its lending
activities, but the effort came "later in the examination period." The bank would have made an
effort to improve its performance earlier if CRA exams placed more emphasis on loan
originations than purchases.

While NCRC cited the Travelers OTS CRA exam for a thoughtful fair lending review, the same
CRA exam encourages the thrift to manipulate its purchasing activity. The CRA exam states that
Travelers purchased "300 loans from a Wilmington-based commercial bank during the CRA
evaluation period, totaling $25.1 million. Travelers paid a premium for these loans and requested
that they all be to low- and moderate-income borrowers." In the very next paragraph, the exam
reiterates that 100 percent of these 300 purchased loans were to LMI borrowers. Instead of
paying a premium for purchased loans, perhaps Travelers should have made more of an effort to
make loans to LMI borrowers in its assessment area of Wilmington.

Purchased loans are becoming like affiliates in that banks are starting to use them in an optional
manner on CRA exams. Banks are asking their examiners to discount purchases when purchases
hurt their performance on the lending test, but asking their examiners to include purchases when
it helps their lending test. For example, National City asked its CRA examiner to exclude
purchases from its CRA exam. The CRA examiner agreed, noting that including purchases
lowers the percent of National City's loans made to LMI borrowers and census tracts.19 In an
even more bizarre example, the CRA examiner of Fifth Third bank included purchases in the
category of lending to LMI census tracts but excluded it in the category of LMI borrowers.20
Purchases presumably increased the percentage of loans to LMI tracts, but lowered the
percentage of loans to LMI borrowers due to the considerable number of purchases without

19 OCC CRA Exam of National City Bank of Ohio, Charter #786, February 2000.
20 Federal Reserve Bank of Cleveland, CRA exam of Fifth Third Bank, Charter #723112, March 2001.

borrower income recorded.
The optional use of purchases increases possibilities of banks manipulating their CRA exams and
reporting poor HMDA data. One reason banks opt to exclude purchases made to LMI borrowers
is that banks are not required to report the income level of borrowers on loans they purchase.
Instead of requiring or encouraging institutions to include data on income level of borrowers in
their purchased loans, examiners have allowed banks to selectively include or exclude this data
on their exams.

The current regulation and procedures involving purchases only invites more manipulation and
poor HMDA data reporting. The end result is that ability of CRA exams to accurately measure
whether banks are meeting credit needs will diminish over time.

Conclusion

We are aware that some parties will be urging you to provide the streamlined test for banks with
up to $1 billion assets. NCRC reiterates that our comprehensive analysis reveals starkly the
devastating impacts of providing the streamlined test for any additional institutions.
Furthermore, some parties will be urging you to further dilute the CRA investment and
community development lending tests. The existing regulations, if anything, do not focus
enough on activities that substantially benefit low- and moderate-income communities. Any
dilution will reduce the number of critically needed and safe and sound community development
loans and investments for low- and moderate-income communities.

The proposed changes to CRA will directly undercut the Administration's emphasis on minority
homeownership and immigrant access to jobs and banking services.

The proposals regarding streamlined exams and the anti-predatory lending standard threaten CRA's statutory purpose of
the safe and sound provision of credit and deposit services.

The proposed data enhancements would become much more meaningful if the agencies update procedures regarding assessment
areas, affiliates, and the treatment of high cost loans and purchases on CRA exams.

CRA is simply a law that makes capitalism work for all Americans. CRA is too vital to be gutted by
harmful regulatory changes and neglect.

For tables, footnotes, sources, go to :
National Community Reinvestment Coalition * http://www.ncrc.org 34

Also: FDIC Summary of Deposit database
National Community Reinvestment Coalition * http://www.ncrc.org *
NCRC Analysis