Community banks and rural development: research relating to proposals to revise the regulations that implement the Community Reinvestment Act.

Since 1977, the Community Reinvestment Act (CRA) has required that federally insured banking institutions--commercial banks and savings associations--be evaluated on their records of helping to meet the credit needs of their local communities, including low- and moderate-income (hereafter, lower-income) neighborhoods. In 1995, the four federal agencies responsible for bank supervision substantially revised the regulations that implement the CRA. (1) The revisions were intended to emphasize performance rather than process, to reduce unnecessary regulatory burden, and to increase consistency in CRA evaluations.

Under the 1995 regulations, "large" institutions, generally those with assets of $250 million or more, have been evaluated under a three-part test, whereas "small"' institutions, generally those with assets of less than $250 million, have been subject to comparatively streamlined evaluations. Large institutions have been required to report data annually on certain types of CRA-related loans (small-business, smallfarm, and community development loans) and on the geographic areas (for example, census tracts) that constitute their local communities, whereas small institutions have been exempt from such reporting.

In 2001, the agencies began reviewing the CRA regulations to determine whether they were successful in meeting the objectives that the agencies set forth in 1995. The review focused in part on the possibility of extending the eligibility for streamlined examinations and the exemption from data reporting to more institutions. In 2004 and 2005, the agencies put forth several proposals to implement these changes by raising the asset-size threshold from $250 million to $500 million or $1 billion. The proposals, and the public's comments on them, paid particular attention to how and when to evaluate the community development performance of banking institutions with assets of less than $1 billion, especially in rural areas, where such institutions have a proportionately larger presence than in urban areas. A related but separate issue that the agencies presented for public comment was how to define which bank activities in rural areas should be considered community development in CRA evaluations.

We have evaluated a large amount of data to gain insight into the potential effects of these proposals, and in this article we report the key findings of our research. Our intent is to inform deliberation over the recent proposals, not to advocate any particular view.

BRIEF DESCRIPTION OF THE CRA

The CRA encourages federally insured banking institutions to help meet the credit needs of their communities, including lower-income neighborhoods, in a way that is consistent with the safe and sound operation of those institutions. (2) In particular, the CRA directs the federal agencies responsible for bank supervision (1) to assess through examinations every institution's record of meeting such community credit needs and (2) to consider the institution's CRA record when evaluating its application for deposit insurance or for a charter, branch or other deposit facility, office relocation, or merger or acquisition.

The CRA gives the agencies broad discretion to implement the law. For example, the act does not define "low- or moderate-income neighborhood" or a banking institution's "community"; rather, the act leaves those definitions to the agencies. The act also leaves to the agencies the establishment of criteria for rating an institution's record of meeting its community's credit needs. Each agency has separate rulewriting authority for the institutions it supervises; but with one recent exception, the four agencies have adopted identical regulations. (3)

The 1995 regulations establish objective standards for measuring performance. Rather than providing specific lending thresholds for particular CRA ratings, however, the standards are flexible and are applied in the context of information about an institution, its community, and its competitors (broadly referred to as the institution's "performance context"). Moreover, the standards relate not only to the quantity of an institution's activities (for example, the dollar amount of mortgage loans extended) but also to the quality of those activities (that is, their correlation with the community's needs for credit).

Examiners evaluate institutions primarily on their performance in their local communities, which the regulations define as the institutions' "assessment areas." Assessment areas encircle an institution's deposit-taking facilities, such as its branches and, if applicable, its automated teller machines (ATMs). Assessment areas are composed of census tracts or aggregations of census tracts, such as counties or metropolitan statistical areas. Examiners consider an institution's performance outside its assessment area only in limited circumstances.

Transparency is an important aspect of the regulations. Every institution's CRA rating--either "outstanding," "satisfactory," "needs to improve," or "substantial noncompliance"--is made public, as is a written evaluation that explains the basis of the rating. (4) Moreover, large banking institutions must report loan data to their supervisory agencies, which make the data publicly available. Large institutions must report the number and dollar amount of their small-business and small-farm loans by individual census tract. They must also report the total number and dollar amount, but not the geographic distribution, of their community development loans. In addition, if an institution is subject to the reporting requirements of the Home Mortgage Disclosure Act (HMDA), it is required to disclose detailed information about its mortgage loans; if the institution is also large for CRA purposes, it must report geographic information for rural mortgage loans, which it otherwise does not need to report. (5)

The criteria in the 1995 regulations for evaluating an institution's performance incorporate four key distinctions. First, the criteria distinguish large banking institutions from small ones. Large banking institutions are subject to a three-part test that looks at lending, investments, and services, whereas small banking institutions face a streamlined test that concentrates on lending (see boxes "The Large-Institution Evaluation" and "The Small-Institution Evaluation"). Moreover, large banking institutions must report data to the agencies; small banking institutions need not do so.

Second, the criteria distinguish among types of banking activity: lending, investing, and providing services. The regulations require the agencies to give large banking institutions explicit sub-ratings on each of these types of activity. Although small banking institutions are not usually evaluated on their investments or services, they may improve their chances of receiving an "outstanding" CRA rating if they elect to be evaluated in those areas.

Third, the evaluation criteria reflect a distinction between area-based and recipient-based measures of performance. The CRA's measure of area is the census tract. Key area-based criteria in CRA evaluations include the proportion of an institution's retail loans, and the proportion of its branches, in lower-income census tracts. Categories of census tract income are determined by the ratio of a census tract's median family income to the median family income of the relevant surrounding area as established at the most recent decennial census. The ranges are 0-49 percent (low), 50-79 percent (moderate), 80-119 percent (middle), and 120 percent or more (upper). For a census tract in a metropolitan (urban) area, the relevant surrounding area is the metropolitan area. For a census tract in a nonmetropolitan (rural) area, the relevant surrounding area is the entire nonmetropolitan region of the state. Baseline classifications of census tract income change every ten years with the release of the census. (6)

In addition to area-based measures of performance, CRA evaluations use analogous recipient-based measures. Examples include the proportion of an institution's loans extended to lower-income borrowers (in the case of mortgage and consumer loans) and to enterprises of different sizes (in the case of small-business and small-farm loans) and the proportion of an institution's services offered to lower-income individuals. The evaluation criteria classify borrowers by income in relation to the median family income of the relevant surrounding area. In doing so, the criteria use the same percentage breakdowns used to classify census tracts by income and the same metropolitan-nonmetropolitan distinction to construct the baseline. The main difference between the classifications for borrowers and those for census tracts is that baseline classifications for borrowers are updated every year, when the Department of Housing and Urban Development publishes the estimates of area family income, whereas those for census tracts are updated every ten years.

Fourth, the evaluation criteria distinguish between retail activities, which are often regarded as the traditional business of a banking institution, and community development activities, which are intended primarily to improve the welfare of lower-income people or areas. The regulation recognizes four categories of community development activity, three of which (affordable housing, community services, and economic development through small-business or small-farm financing) target certain recipients--lower-income people, small businesses, or small farms--and one of which (revitalization and stabilization) targets certain areas--lower-income census tracts. For a large institution, community development performance is a factor in the CRA sub-rating on each of the three activity-based tests (lending, investment, and service). In the case of the investment test, the sub-rating depends entirely on the institution's record of making community development investments, whereas in the case of the lending and service tests, the sub-rating depends, respectively, on the institution's record of providing retail and community development loans and on its record of providing retail and community development services.

For a small institution, unlike for a large one, community development performance is not a mandatory part of the evaluation. But a small institution may choose to be evaluated on its community development loans, investments, or services as a basis for possibly boosting the institution's rating from "satisfactory" to "outstanding."

THE AGENCIES' PROPOSALS TO AMEND THE CRA REGULATIONS

In 1995, when the four banking agencies adopted major amendments to the regulations that implement the CRA, they committed themselves to reviewing the amended regulations to assess the regulations' effectiveness in emphasizing performance over process, promoting consistency in evaluations, and eliminating unnecessary regulatory burden. (7) They began that review in July 2001 with the publication in the Federal Register of an advance notice of proposed rulemaking. (8) Since early 2004, the agencies have issued several proposals.

Recent CRA Proposals

In February 2004, the banking agencies issued identical proposals to amend their respective CRA regulations to increase the number of institutions classified as small. (9) Under the 1995 regulations, an institution is defined as small if it has less than $250 million in assets and is not a member of a holding company that has $1 billion or more in assets. Institutions not defined as small are classified as large. (10) The four agencies proposed to expand the definition of "small institution" to cover those institutions with assets of up to $500 million and to eliminate the holding company criterion.

Commenters on that proposal were deeply split. Industry commenters, seeking to reduce their regulatory burden, wanted to raise the large-institution threshold higher than was proposed (as high as $2 billion). But community groups opposed any increase in the threshold, asserting that an increase would lead institutions newly classified as small to reduce their investments in community development.

In July 2004, the OTS announced that it would raise the large-institution threshold for savings associations to $1 billion, the OCC announced that it would refrain from adopting the February proposal, and the Board stated that it would formally withdraw the proposal from consideration. (11) The Board explained that raising the large-institution threshold to $500 million was not guaranteed to yield significant cost savings for institutions and that it might significantly reduce investments in community development in some rural communities. (12)

A month later, the FDIC issued a new proposal to raise the large-institution threshold to $1 billion for FDIC-supervised institutions and to continue to evaluate institutions with assets between $250 million and $1 billion on their community development records but on a modified basis. (13) Again, commenters were divided over the proposal. Many industry commenters opposed evaluating these institutions on their community development records; many community group commenters contended that the proposed evaluation was not rigorous.

Also in August 2004, the FDIC proposed that a bank activity that benefits an individual or a community in a rural area be considered community development under the CRA, even if neither the individual nor the community is of lower income. Commenters also split on that proposal. Some expressed concern that the agency would give CRA recognition to bank investments in affluent rural areas. Some supported the proposal, however, because they favored recognizing institutions' support of infrastructure, business development, and other needs in rural areas as community development.

In November 2004, the OTS, too, proposed to recognize as community development a bank activity that benefits an individual or a community in a rural area, even if neither the individual nor the community is of lower income. (14)

The Three-Agency Proposal of February 2005

In February 2005, the Board, the OCC, and the FDIC published for public comment a joint proposal, which again addressed the definitions of "small institution" and "community development" in rural areas. The proposal would modify the CRA regulations in three ways:

1. It would raise the asset threshold for a large institution from $250 million to $1 billion and would eliminate the holding company criterion. Thus, all banking institutions with less than $1 billion in assets would be exempt from CRA data reporting obligations.

2. It would create a subcategory of small institutions called "intermediate small institutions," those with assets between $250 million and $1 billion, and would subject such institutions to a two-part evaluation. (15)

* One part would evaluate the institution's retail lending. The evaluation would use the criteria now used for small institutions (those with less than $250 million in assets)--for example, the ratio of overall loan dollars to deposits and the distribution of loans across borrowers and areas of different relative incomes. Those criteria differ little in substance from the criteria applied to the retail lending of large institutions, but the evaluation of small institutions' retail lending is, in practice, more streamlined because of their exemption from the requirement to collect or report data on loans and assessment areas.

* A second part, given equal weight in assigning an overall CRA rating, would evaluate an intermediate small institution's community development record. Instead of considering that record in three separate tests (as does the large-institution evaluation, which now applies to intermediate small institutions), the evaluation would gather into one test all community development activities, regardless of type, including lending, investing, and providing services.

3. It would revise the definition of "community development" in rural areas--for institutions of any size. The definition in the 1995 regulations imposes a lower-income restriction on bank activities that may be credited as community development in CRA evaluations: Such activities must primarily benefit either lower-income people or lower-income areas. The agencies proposed to relax that restriction in rural areas.

* Under the proposal, bank activities would be considered community development if they revitalized or stabilized any "underserved rural area" or provided affordable housing for any individual in any such area, even if the area was not defined as "low or moderate income." The agencies sought comment on how to identify underserved rural areas not already classified as lower-income tracts. The agencies specifically sought comment on criteria adapted from the Community Development Financial Institutions Fund's definition of an "investment area." The criteria, as adapted, would identify as underserved an area that has at least one of the following characteristics: (1) an unemployment rate of at least 1.5 times the national average, (2) a poverty rate of 20 percent or more, or (3) a population loss of 10 percent or more between the previous and most recent decennial censuses or a net migration loss of 5 percent or more over the five-year period preceding the most recent census.

* The agencies also sought comment on an alternative proposal to liberalize the definition of a "low or moderate income" rural census tract in one of two ways, at least for the purpose of determining which area-based activities in rural areas are considered community development: (1) change the baseline for defining rural tract incomes from the nonmetropolitan state median income to the statewide median income, which is the higher of the two statistics in all but one state, or (2) raise the "low or moderate income" limit from its current level of 80 percent.

EFFECTS OF RAISING THE ASSET-SIZE THRESHOLD

As noted earlier, the 2005 proposal would raise the asset-size threshold for a large institution from $250 million to $1 billion and would eliminate the holding company criterion. Institutions with asset sizes below the $1 billion threshold would be subject to a streamlined CRA lending test equivalent to that now used for small institutions; they would also be exempt from the evaluation of branching under the service test now applied to institutions with assets of more than $250 million. (16) The proposal would also create a new community development test for intermediate small institutions.

In the first part of this section, we analyze several issues related to this portion of the proposal. (17) First, we identify and describe the institutions and banking markets that would be affected by raising the threshold to $1 billion and eliminating the holding company criterion. Second, we examine the potential effect of using a streamlined version of the CRA lending test and eliminating the service (branching) test for institutions with asset sizes below the threshold. Specifically, we examine the effect of the current $250 million threshold on the retail lending and branching activities of institutions within a narrow range of the current threshold. Third, we consider whether the role of community development lending in CRA ratings has been significant.

Parties Affected by Raising the Threshold and Eliminating the Holding Company Criterion

Raising the threshold and eliminating the holding company criterion would affect both banking institutions and the communities they serve. Using 2003 as a test year, we looked at the characteristics of institutions that would have been subject to a different CRA evaluation process had the regulations proposed in February 2005 been in effect. (18) We also identified the local banking markets that might have been most affected had the threshold been raised.

Banking Institutions

As of December 31, 2003, 9,095 banking institutions were subject to the CRA (table 1). We estimate that, of those, 1,621 institutions that were considered large as of that date would be considered intermediate small or small under the agencies' 2005 CRA proposal (columns 1-3). These "status-changing" institutions constituted 18 percent of all institutions subject to the CRA, and they held 13 percent of the deposits held by all such institutions. Intermediate small institutions consisted of 1,264 institutions that had between $250 million and $1 billion in assets (column 2) and 116 institutions that had more than $1 billion in assets but which would not be considered large under the proposal because of a proposed requirement to exceed the asset threshold for two consecutive years (column 3). The "newly small" institutions consisted of the 241 institutions that had assets of less than $250 million but which nonetheless were subject to the large-institution CRA examination in 2003, generally because they were part of a bank holding company with assets of more than $1 billion (column 1). (19) These institutions would be considered small under the 2005 proposal. (20)

The 1,621 status-changing institutions differ from other CRA-covered institutions along a number of dimensions. First, 28 percent of the status-changing institutions had headquarters in nonmetropolitan areas, compared with 7 percent of institutions with assets exceeding $1 billion and 52 percent of small banking institutions. Of the status-changing institutions with headquarters in nonmetropolitan areas, 60 percent had headquarters in exurban counties (nonmetropolitan counties adjacent to metropolitan areas) and 40 percent in remote counties (counties not adjacent to metropolitan areas). (21) Second, 14 percent of the status-changing institutions had "outstanding" CRA performance ratings at their last examinations, compared with 37 percent of larger institutions and 11 percent of small institutions.

Local Banking Markets

Another way to look at the effect of raising the threshold is in terms of local banking markets. There is no universally accepted geographic definition of local banking markets, but the Federal Reserve Banks have constructed a list of local banking market definitions for reviews of the competitive effects of proposed mergers and acquisitions, and we used these definitions in our analysis. (22) Not all parts of the country have been defined for this purpose; however, the 1,873 defined markets account for 96.7 percent of the branches and 97.7 percent of the deposits held by banking institutions nationwide (table 2). Seventy percent of banking markets are rural, but such markets account for a relatively small proportion of banking deposits nationwide (about one-eighth--data omitted from table) because most people and businesses are located in metropolitan areas.

For the market analysis, we focused on markets in which status-changing institutions play a significant role. We used two methods to characterize the roles of status-changing institutions in their markets: the percentage of a market's deposits held by status-changing institutions and the determination of whether a status-changing institution is the largest institution in the market.

The first method, the "market-share method," classifies markets by the percentage of each market's deposits held by status-changing institutions. This method assumes that an institution's propensity to invest in its market is directly related to its share of the market's deposits. (Here, we use the word "invest" in its broadest sense to include extensions of credit, services, grants, and equity investments.) The method further assumes that all institutions that shift from a large-institution examination to a small-institution examination experience the same proportional change in their propensity to invest in the market. These assumptions imply that the markets with the greatest presence (as measured by share of market deposits) of status-changing institutions will experience the largest proportional changes in banking institutions' lending and investing in the market.

The second method of market analysis, the "largest-institution method," classifies markets by the size of the largest institution with a presence (office) in the market regardless of the market share of that institution. This method assumes that if raising the threshold has an effect, the effect is particularly large in those markets that go from having one or more institutions that are subject to the three-part large-institution examination to having no such institutions.

As noted earlier, the market-share method identifies the share of market deposits held by status-changing institutions (table 3). As of December 31, 2003, 28 percent of the nation's 1,873 banking markets (with 3 percent of nationwide deposits) had no status-changing institution located within the market (column 1). Under either the marketshare or the largest-institution method, those markets would presumably be unaffected by raising the threshold to $1 billion. In another one-third of markets (with 82 percent of nationwide deposits), status-changing institutions held less than 20 percent of deposits and thus, under the market-share method, would likely not see major effects (columns 2 and 3). But in roughly 10 percent of markets (with 2 percent of nationwide deposits), status-changing institutions held more than 50 percent of market deposits and, consequently, those markets have the potential to be most affected (column 5).

The largest-institution method yields a different group of potentially most affected markets (table 4). The largest institution was a status-changing institution in about 8 percent of all markets (with less than 1 percent of nationwide deposits) (column 2).

Although there is little overlap between the groups of potentially most affected markets defined by the two methods, the groups have several characteristics in common (table 5). In both groups, the markets are overwhelmingly rural, are served by few banking institutions (data omitted from table), have unemployment rates near the national rural average, and have bank branches per capita similar to the national rural average. Yet the groups of markets differ in key respects. Whereas population growth is much higher than the national rural average in the potentially most affected markets identified by the market-share method, it is notably lower than the national rural average in such markets identified by the largest-institution method. Moreover, under the market-share method, the potentially most affected markets are scattered throughout the country (figure 1), but under the largest-institution method, such markets are concentrated substantially in the Great Plains region, with much smaller concentrations in Iowa, Louisiana, Kentucky-Tennessee, and southern Georgia-northern Florida (figure 2). (23)

[FIGURES 1-2 OMITTED]

Results of Threshold Tests

The 2005 proposal would subject intermediate small banking institutions to the streamlined lending test currently applied to smaller institutions and would eliminate the service (branching) test for intermediate small institutions. Testing directly to determine how those changes would affect activities of intermediate small institutions is impossible. However, an inference might be drawn from the effect of the current $250 million threshold on the retail lending and branching activities of institutions with assets near this threshold. Of particular interest are the retail lending activities covered by both the large-institution and the streamlined lending tests. Our tests compare the retail lending and branching of institutions just above and just below the current large-institution threshold of $250 million (table 6). Institutions "just below" the threshold are defined here as those that had between $150 million and $250 million in assets as of December 31, 2003, and were deemed small as of that date and for the purposes of their most recent CRA performance evaluation (group 1); institutions "just above" the threshold are defined as those that had between $250 million and $350 million in assets and were deemed large as of that date and for the purposes of their most recent examination (group 2).

Banking institutions in these two groups were restricted to institutions that were independent of multibank holding companies, that had a CRA examination completed between January 1, 1999, and June 30, 2004, and that received a "satisfactory" CRA performance rating on their most recent examination in that period. Institutions with "outstanding" ratings were excluded to control for the possibility that such institutions were influenced less by the nature of their CRA examinations and more by other factors, such as institution philosophy, than were institutions with "satisfactory" ratings. Institutions with less than "satisfactory" ratings were excluded for similar reasons. Institutions with headquarters in U.S.-affiliated areas were also excluded.

The threshold test relies on sources of data that provide the same types of information for institutions just above the threshold as for institutions just below. The information consists of five balance sheet ratios constructed from dollar values provided in Call Report data supplied to federal banking agencies. (24) The ratios compose two categories: loan dollars to deposits (overall; consumer; and business, including small commercial and industrial [C&I], small commercial real estate [CRE], and small farm) and mortgage dollars to deposits (one- to four-family and multifamily).

In addition, five measures of retail lending to lower-income populations were constructed from filings pursuant to HMDA: the percentage of an institution's home-purchase and home-improvement loans extended to lower-income borrowers or census tracts and a comparable calculation for loans extended for multifamily housing in lower-income census tracts. Each of the HMDA-based measures was expressed as the difference between the percentage of the institution's retail loans made to lower-income borrowers (or borrowers that live in lower-income tracts) and the percentage of the families that live in the areas served by the institution who have lower incomes (or live in lower-income tracts). Because of limitations on reporting requirements for rural institutions under HMDA, the comparisons that use the HMDA-based lower-income lending measures were restricted to retail lending activities in urban areas.

We present means of these metrics for the groups above and below the threshold, adjusted to remove effects related to state, institution type (savings association or commercial bank), and headquarters location (center city, suburban, exurban, remote) as rough controls for economic and demographic factors. (25) The results are of three types. First, overall C&I lending, CRE lending (overall or small), multifamily housing, and the HMDA data measures show no statistically significant differences. Second, the other business loan categories do show differences that are statistically significant; however, the direction of the differences is the opposite of what would have been expected had the differences been caused by tougher evaluation criteria in the large-institution evaluation.

Third, statistically significant differences exist in the groups' ratios of overall loan dollars to deposits, consumer loan dollars to deposits, and one- to four-family mortgage dollars to deposits, and these differences go in the direction that might suggest a threshold effect. To confirm that this result reflects differences in CRA evaluation criteria and not merely in asset size, we conducted an additional comparison test. We constructed a third group of banking institutions that had between $350 million and $450 million in assets and that otherwise met the same requirements as the institutions in group 2 (group 3). A comparison of adjusted means for group 3 with those for group 2 isolates the effects of size differences because banking institutions in both groups are subject to the same type of CRA evaluation. Institutions in group 3 have lower ratios of overall loan dollars to deposits, consumer loan dollars to deposits, and one- to four-family mortgage dollars to deposits than have institutions in group 2 (and, in two out of three cases, lower than those of group 1 institutions), an indication that the difference between groups 1 and 2 may be caused by a factor other than the difference in CRA examination types.

We also compared adjusted means for groups 1 and 2 (and for groups 2 and 3) on three measures of branching activity: (1) the number of branches per $100 million of deposits, (2) the difference between the percentage of branches in lower-income census tracts and the percentage of the population that lives in lower-income census tracts in the areas that the institutions serve, and (3) the five-year change in the percentage of branches in lower-income areas. None of the three measures shows a significant difference among any of the groups.

As a further test for the effects of differences in CRA evaluation types, we compared independent institutions that had between $150 million and $250 million in assets (group 1) with similarly sized institutions that were subject to the large-institution evaluation criteria because of their affiliation with holding companies with assets of $1 billion or more (group 4). The ratios of overall loan dollars to deposits, consumer loan dollars to deposits, and one- to four-family mortgage dollars to deposits are the only measures with a statistically significant difference: Group 4 has higher ratios than does group 1. One should interpret these results cautiously, as they may mean only that banking institutions in holding companies are more likely than independent banking institutions to raise funds through wholesale, nondeposit markets and to be institutions focused on retail lending.

A final test for the effects of differences in CRA evaluation types examined whether banking institutions that passed the $250 million threshold measurably changed their retail lending and branching activities. Specifically, one hundred institutions covered by the large-institution CRA evaluation (though not necessarily yet evaluated as large institutions) at the end of 2002 and at the end of 2003 had been subject to the small-institution evaluation in 2001. This test, unlike the other tests, looked for any change in an individual institution's behavior induced by a change in evaluation type. We restricted the comparison to institutions covered by the large-institution examination for both 2002 and 2003 to ensure that ample time had elapsed for behavioral changes to result in measurable changes in balance sheet variables. We found only two statistically significant changes in retail lending or branching behavior as an institution passed through the $250 million threshold (compare columns 5 and 6), but the changes were in opposite directions: Consumer lending fell, and overall CRE lending rose. (26)

Taken together, the threshold tests provide little evidence that the nature of the CRA examination influences the retail lending and branching activities of banking institutions in the size range near the $250 million threshold. However, the threshold tests have an important limitation. The tests are limited to inferences about the behavior of institutions around the margin of the current threshold, $250 million. They suggest that raising the threshold some amount above $250 million would not have a significant effect on retail lending or branching. However, they fail to reveal what amount of increase in the threshold, if any, would result in a significant effect.

The Role of Community Development Lending

The 1995 regulations require that, for a large institution, community development lending be evaluated as only one component of the CRA lending test, which includes a wide range of other, retail types of lending (see box "The Large-Institution Evaluation"). Under the 2005 proposal, intermediate small institutions would be subject to a new community development test. Instead of considering community development loans, investments, and services in three separate tests, the proposal is for the three types of activity to be considered in a single test. The proposal responds in part to the argument that community development lending is more like community development investments--both are primarily for the benefit of lower-income people or areas--than like retail lending. It also responds to the argument that evaluating community development investments separately from retail lending places too much emphasis on investment vehicles, especially for smaller institutions that have limited experience with and opportunities for investments and substantially more experience with and opportunities for retail lending.

We could not test for the effects of adopting a community development test on community development loans, investments, or services. We could, however, consider whether the number or dollar amount of community development loans have played a significant role in CRA ratings. Our sample was restricted to institutions examined under the large-institution examination between January 1, 2001, and December 31, 2003. We looked at the institutions' community development lending records over the same period (table 7). The data suggest that an institution's community development lending record is largely unrelated to its overall CRA rating. The lack of relationship is most apparent among institutions with assets of more than $5 billion: Nearly one-half of the institutions with "outstanding" ratings had community development lending activity in the bottom half of their asset-size group. Among intermediate small institutions, those with "outstanding" ratings were a little more likely than their counterparts with "satisfactory" ratings to do community development lending. However, fully one-fourth of the institutions rated "outstanding" in this category did no community development lending, and about 40 percent had community development lending activity in the bottom half of their asset-size group. Although the data provide no information about how community development lending should be treated in CRA evaluations, they do suggest that such lending is not currently critical in overall CRA ratings. The likely explanation is that, because examiners consider community development loans as part of a comprehensive lending test, other types of lending may have compensated for an institution's lack of community development loans. Another possibility is that, despite the mandate of the regulations to treat community development loans and community development investments separately, examiners implicitly treat them as substitutes. (27)

COMMUNITY DEVELOPMENT AND RURAL AREAS

Another part of the agencies' 2005 proposal would expand the definition of "community development" in rural, though not urban, areas. This part of the proposal would cover banking institutions of all sizes, not just intermediate small institutions.

The Problem and the Agencies' Proposed Solution

The regulations' current definition of "community development" is identical for urban and rural areas. As noted earlier, the definition covers four categories of activity, three of which (affordable housing, community services, and economic development) are defined in terms of the activity's targeting of certain recipients (lower-income people, small businesses, or small farms) and the fourth of which (revitalization and stabilization) is defined in terms of the activity's targeting of certain areas--namely, lower-income census tracts.

Some have said that the lower-income-area limitation in the fourth category, revitalization and stabilization, may unduly constrain the effectiveness of the regulations in promoting community development activities in rural areas. In response to such concerns, the agencies proposed to expand the definition of "community development" to include revitalizing or stabilizing activities in underserved rural areas. No such change was proposed for urban areas.

The problem that the agencies sought to address stems in part from the way rural census tracts are classified. As applied to rural areas, the 1995 regulations' system for classifying census-tract income has two defining characteristics. The first characteristic is that the system ignores the fact that rural areas are generally poorer than urban areas. Forty-three percent of rural census tracts in the United States (containing 40 percent of the rural population) have a median family income below 80 percent of the median family income of the state in which the tracts are located; in contrast, 30 percent of urban census tracts (containing 26 percent of the urban population) have a median family income below 80 percent of the statewide median (table 8, "Total rural" and "Total urban" categories). But the 1995 regulations classify rural census tracts relative only to a state's rural median income, not relative to the median income of the entire state, including its urban areas. Thus, the current rule classifies only 15 percent of rural tracts, not 43 percent, as lower income. In contrast, despite the higher absolute incomes of urban areas, double the proportion of urban tracts (31 percent), which are classified relative to the relevant metropolitan area income, are currently classified as lower income.

The second characteristic is that the census tract identifies pockets of lower-income populations less effectively in rural areas than in urban areas. Compared with urban census tracts, rural tracts are drawn over relatively large geographic areas, have lower population densities, and often have relatively heterogeneous populations that, when averaged, tend toward the middle (table 9). Indeed, 73 percent of all rural tracts are defined as middle income; in contrast, 44 percent of urban tracts are defined as such (percentages derived from table 8).

The large size of rural census tracts and the relative heterogeneity within them have another consequence: uneven distribution of lower-income tracts among areas that define banking institutions' markets, such as counties and assessment areas. Most rural counties (almost 60 percent) have no tracts that are classified as lower income under the current definition; in contrast, only 18 percent of urban counties are without any such tracts (table 9). About 44 percent of the rural assessment areas that large institutions reported under the CRA regulation in 2003 lacked any tracts classified as lower income, whereas only 14 percent of the urban assessment areas that these institutions reported lacked any such tracts. Small institutions do not report their assessment areas, though the areas are described in their performance evaluations. A rough approximation of a small institution's assessment areas--one that uses the counties in which its branches are located--suggests that 54 percent of small institutions also lack any lower-income tracts in rural areas.

The relative lack of lower-income tracts in rural areas could have different consequences. Banking institutions might invest less, or less efficiently, in community-improving activities in rural areas than they might under a standard more appropriate for rural community development. Or they might shift more of their community-improving loans or investments to urban areas than they might under a standard that would give more equal area-based CRA consideration in urban and rural areas. A third possibility is that, even if the first and second possibilities failed to occur, banking institutions might receive inadequate recognition of their community-improving activities in rural areas because the activities did not meet the exact requirements to qualify as community development.

Perhaps to address these possible consequences, the 2005 proposal would expand, in two ways, the criteria under which banking institutions receive CRA consideration for community-improving activities in rural areas. First, CRA consideration would be available for activities that revitalize or stabilize "underserved rural areas," in the words of the proposal, even if the areas lack lower-income tracts. Second, the proposal would extend CRA consideration to affordable housing for any individual in an underserved rural area. The 1995 regulations limit affordable housing consideration to housing for lower-income individuals; consideration does not depend on where the lower-income individuals reside. The proposal would leave unchanged the recognition of community development activities in urban areas and the non-community-development CRA measures.

In this section, we analyze several issues related to the agencies' proposal to expand the criteria for recognizing rural community-improving activities as community development. First, we test the proposal's premise that the 1995 regulations "disfavor" community-improving activities in rural areas relative to those in urban areas. Second, we explore the implications of various options to revise the regulations on which the agencies sought public comment.

Concern about Whether the 1995 Regulations "Disfavor" Rural Areas

Research on the question of whether the 1995 regulations disfavor rural areas is constrained by the difficulties in gathering comprehensive data on community-improving activities that fail to qualify for CRA consideration and by the lack of geographic data on community development loans, investments, and services. However, geographic data are available for all other CRA-related loan products and branches, such as loans to purchase or improve homes or to finance small businesses or small farms. We used these data to test whether large institutions whose assessment areas include both urban and rural areas ("urban-rural institutions") appear to favor urban areas over rural areas in retail lending and branching activities. We also tested whether large institutions with headquarters in rural areas ("rural institutions") were less likely than similarly situated institutions with headquarters in urban areas ("urban institutions") (1) to receive "outstanding" CRA ratings or (2) to engage in community development lending.

The first test was based on the distribution of urban-rural institutions' activities between urban and rural parts of their assessment areas. We restricted the analysis to institutions covered by the large-institution examination as of December 31, 2003, because such institutions are required to report the geographic location of most of their CRA-related loans, with the notable exception of community development loans. The test involved two distinct comparisons. First, the test compared the distribution of numbers of retail loans between urban and rural parts with the distributions of offices, populations (families), and housing structures (owner-occupied or multifamily) between the parts. This comparison tested whether urban-rural institutions extend retail loans in the same proportion to the offices, populations, and housing structures of those areas (table 10). (28) Second, the test compared, for various types of retail loan in those institutions' assessment areas, the distribution of loan dollars between urban and rural parts with the distribution of deposits between the parts. The comparison tested whether urban-rural institutions extend loan dollars in the same proportion that they receive deposits in rural and urban areas (table 11). We conducted both comparisons separately for banking institutions in four asset-size categories: $250 million to $500 million, $500 million to $1 billion, $1 billion to $5 billion, and more than $5 billion.

We found that remote areas receive more retail loans as measured against the distributions of offices, populations, and housing structures and more loan dollars as measured against the distribution of deposits than do urban areas in the aggregate for banking institutions of every size category and for retail loans of almost every type considered (tables 10 and 11). For example, urban-rural institutions with assets between $500 million and $1 billion received 13.6 percent of their deposits from branches in remote areas and extended 18.0 percent of their home-purchase loans, 20.6 percent of their home-improvement loans, and 23.9 percent of their smallfarm or small-business loans in such areas. (29)

The data for exurban areas are more difficult to interpret than are the data for remote areas. Generally, urban-rural institutions make more retail loans per family in exurban areas than in urban areas (data derived from table 10). But such institutions extend fewer retail loan dollars per deposit dollar in exurban areas than in urban areas (data derived from table 11). This apparent inconsistency may be explained by the fact that the majority of the urban-rural institutions in our sample, particularly the smaller ones, have headquarters in exurban areas. The deposit data may reflect a practice by some of those institutions of booking deposits to their headquarters regardless of the locale from which deposits originated.

The tendency of urban-rural institutions to make more retail loans to their rural components than to their urban components also holds true at the level of the individual institution. With one exception, more than one-half of the institutions in every size category extended more retail loans per family, per owner-occupied housing structure, or per multifamily housing structure to the rural parts of their assessment areas than to the urban parts (table 12). The exception was multifamily loans for institutions in the smallest size category. When measured in terms of retail loan dollars per deposit dollar, the results were somewhat mixed. For example, the rural parts appeared to get more home-improvement loans but fewer home-purchase loans than did the urban parts. (30)

The second test compared rural institutions with similarly situated urban institutions in two respects: the likelihood of receiving an "outstanding" CRA rating and the level of engagement in community development lending. The sample in this test used the same size categories as the sample in the first test and was also restricted to institutions covered under the large-institution evaluation procedures. The second test, however, eliminated the requirement that an institution have both urban and rural parts in its assessment areas.

The evidence suggests that rural banking institutions with assets of less than $1 billion are not less likely to receive "outstanding" ratings than are urban institutions with assets of less than $1 billion (table 13). Exurban institutions with assets between $250 million and $500 million are somewhat less likely to receive "outstanding" ratings than are their urban counterparts, but exurban institutions with assets between $500 million and $1 billion are significantly more likely to do so. Few institutions with assets exceeding $1 billion have headquarters in rural areas; those in that category are less likely to receive "outstanding" CRA ratings than are institutions that have assets exceeding $1 billion and headquarters in urban areas.

The evidence offers modest support for the conclusion that rural institutions do less community development lending than do similarly sized urban institutions. In each asset-size category under $5 billion, the percentage of rural institutions that reported no community development lending in 2003 was comparable to the percentage of similarly sized urban institutions that did so (data derived from table 13). However, for every asset-size category, with one exception, rural institutions that reported community development lending for 2003 made a smaller dollar amount of community development loans than did urban institutions. The exception was remote institutions with assets between $500 million and $1 billion. These institutions had higher community development loan dollar amounts than did the combination of same-sized center-city and suburban institutions.

This evidence on community development lending, however, is indirect and inconclusive. For example, it excludes any measure of community development investments or services. Moreover, because we lack information about the location of community development loans, inferences drawn from locations of institutions' headquarters are subject to dispute.

In sum, the retail lending and branching measures used here provide little evidence that banking institutions collectively or individually underserve rural areas, with the possible exception of community development lending. Moreover, there is no evidence that a lower percentage of rural-based institutions receive "outstanding" CRA performance ratings (at least for such institutions with less than $1 billion in assets).

Rural Areas That Would Be Affected by the Agencies' Proposed Options

The agencies sought comment on several alternative definitions of CRA-eligible rural census tracts. Each alternative satisfies five basic principles. First, each alternative would permit an institution to know, when it decided to make a loan or investment, whether or not the loan or investment would qualify as community development. Second, each alternative would rely on measures that change no more often than annually and in most cases change much less frequently than that. Third, each alternative would rely on purely objective statistical criteria that could be applied mechanically and without judgment. Fourth, each alternative would be easy to apply: Any required calculations would be straightforward or would be obviated by the government's publication of a list of eligible areas. Fifth, each alternative would rely on readily available, government-produced data.

The three alternatives that we considered were (1) moving the income threshold for CRA-eligible rural tracts from 80 percent to 90 percent or 100 percent of the statewide nonmetropolitan median family income, (2) changing the baseline for determining the CRA eligibility of rural tracts from the statewide nonmetropolitan median family income to the statewide median family income, and (3) adopting a modified version of the criteria used by the Community Development Financial Institutions Fund (CDFI Fund) to identify "investment areas."

The fund uses four alternative criteria of interest here to classify geographic areas (tracts, counties, or other aggregations) as investment areas. According to the fund, an area qualifies as an investment area if it has (1) a median family income that is less than 80 percent of the relevant metropolitan median family income or the national metropolitan median family income, whichever is higher, in the case of a metropolitan area, or a median family income that is less than 80 percent of the relevant statewide nonmetropolitan median family income or the national nonmetropolitan median family income, whichever is higher, in the case of a nonmetropolitan area; (2) an unemployment rate of at least 1.5 times the national average; (3) a poverty rate of 20 percent or more; (4) a population loss of 10 percent or more between the previous and most recent censuses or a net migration loss of 5 percent or more over the five-year period preceding the most recent census. (31) Data for unemployment, poverty, and population are updated annually at the county level and decennially at the tract level.

To permit comparison with the current rule, we modified the fund's criteria. Instead of using the fund's income criterion, we used the CRA's. That is, we treated as CRA-eligible any tracts currently classified as lower income (using, in rural areas, the current CRA baseline of the nonmetropolitan statewide median family income) and any tracts currently classified as middle income that are located in a county that meets any of criteria 2 through 4. Thus, in this article, when we refer to the "modified CDFI Fund criteria," we use a modification of the first fund criterion, the one based on income.

There are two key differences between the fund's criteria, which use non-income measures of community need, and the other alternatives, each of which relies solely on a relative tract-income criterion. First, the fund's criteria use measures for which data are at the county level, not the tract level. Second, the fund's county-level criteria use measures that are updated annually; income data at the tract level, in contrast, are updated only every ten years. (32) Consequently, the way in which the fund's criteria identify CRA-eligible areas is different from that in which the income-based alternatives do, and the difference can result in different outcomes.

Our analysis expands on the agencies' proposal in two main respects. The agencies proposed to apply the alternatives outlined earlier only to rural areas and only for the purpose of qualifying activities as community development. Our analysis evaluates the alternatives on those terms but goes beyond those terms. In particular, we show the implications of adopting these alternatives in urban areas, divided into central-city and suburban components, and in rural areas, divided into exurban and remote components. We also show the implications of adopting the alternatives for the purpose of evaluating other CRA-related activities, such as retail lending (table 14).

We computed the effects of the alternatives on the coverage of rural and urban census tracts and on the retail lending activities that would have counted as CRA-related if the alternatives had been in effect in 2003 (we assumed that banking institutions had not altered their behavior). We compared each alternative with actual 2003 retail lending activities adjusted for changes, implemented in 2004, in the definitions and boundaries of metropolitan statistical areas. (33)

Each of the proposals would substantially increase the number of rural tracts that are CRA-eligible--that is, eligible for area-based community development activities. Currently, 14.6 percent of rural census tracts are classified as CRA-eligible; these tracts contain 12.6 percent of the rural population. Raising the threshold to 90 percent for rural areas would roughly equate the percentages of urban and rural tracts classified as CRA-eligible, at about 31 percent; raising the threshold to 100 percent would qualify 55 percent of rural tracts as CRA-eligible. Similarly, changing the baseline for classifying rural tracts to the statewide median but retaining the 80 percent threshold-would qualify 43 percent of rural tracts as CRA-eligible. (34)

We also calculated the effects of adopting the modified CDFI Fund criteria. Using the criteria to identify middle-income tracts that would be CRA-eligible would classify 33 percent of rural census tracts as lower income, a proportion nearly equal to the 31 percent of urban census tracts currently classified as CRA-eligible (see "Current rule" and "Modified CDFI Fund criteria" categories). When each of the unemployment, poverty, and population loss criteria is applied separately, the proportion of rural tracts classified as CRA-eligible is 25 percent, 22 percent, and 19 percent respectively. Applying the modified CDFI Fund criteria to urban tracts would have a comparatively modest effect, increasing the number of urban tracts classified as CRA-eligible from 31 percent to 36 percent. The general patterns described in this paragraph and in the previous one are also found when the unit of analysis is the proportion of population in CRA-eligible tracts.

We also profiled the economic and demographic characteristics of the tracts now classified as lower income and of the additional tracts that would be CRA-eligible under each of the alternatives (table 15). The profile reveals that, in rural areas, the exurban and remote tracts currently classified as lower income have similar average characteristics along most dimensions (although the number of tracts in exurban and remote areas is different). In urban areas, however, the center-city and suburban tracts with this classification are largely dissimilar. Further, a comparison of lower-income tracts in urban and rural areas reveals differences in most characteristics. The differences between suburban and exurban tracts are a case in point: Exurban tracts have higher poverty rates, vacancy rates, and unemployment rates and lower absolute incomes, house values, and population growth rates than do suburban tracts.

Although the specific tracts added by each alternative are different, their economic and demographic characteristics are similar. Under any of the alternatives for expanding the class of rural CRA-eligible tracts, the rural tracts that would be newly classified as CRA-eligible show more-favorable economic characteristics than do the rural tracts currently classified as such. The relationship of the newly classified rural CRA-eligible tracts to urban tracts currently classified as CRA-eligible is complicated. Under any of the alternatives, the newly added rural CRA-eligible tracts would have lower poverty, unemployment, and population growth rates and higher owner-occupancy and vacancy rates than would the current urban CRA-eligible tracts; median incomes for both types of tract would be about the same. Moreover, the rural tracts that would be added under the alternatives that contribute the most rural tracts (100 percent of median family income and 80 percent of statewide median family income) show, not unexpectedly, the most-favorable economic characteristics.

When compared with the current rule (figure 3), each alternative adds a different set of newly CRA-eligible rural tracts with significantly different geographic distributions. That is, with one exception, each alternative--raising the threshold from the current level to 90 percent (figure 4) or 100 percent (figure 5), changing the baseline to the statewide median income (figure 6), and adding the CDFI Fund's non-income criteria to the current 80 percent income rule (figure 7)--adds a set of tracts that differs from the other sets in terms of composition (the tracts that make it up), economic and demographic characteristics, and location. For example, of the 3,559 rural tracts added by adopting the modified CDFI Fund criteria or raising the threshold to 90 percent, only 760 (one-fifth) would be added by both options (data omitted from tables). Moreover, the modified CDFI Fund criteria themselves would add largely dissimilar sets of tracts: 18 percent of rural tracts that meet one or more of the criteria meet two or more of them, and less than 2 percent of rural tracts that meet one or more of the criteria meet all three criteria. (35) The exception to the pattern is that substantial overlap exists between raising the threshold to 100 percent and using the statewide median income as the baseline. Of the 5,188 rural tracts that would be added by either alternative, 64 percent would be added by both options.

[FIGURES 3-7 OMITTED]

The alternatives can also be evaluated from the perspective of banking institutions. For example, 30 percent of large institutions with at least one branch in a rural area currently have no CRA-eligible tracts in any of their rural assessment areas (table 9). (36) Under each of the three income-based alternatives (raising the threshold to 90 percent or 100 percent or changing the baseline to the statewide median income), more than one-half of those institutions would have at least one CRA-eligible tract in at least one of their rural assessment areas (table 16). In contrast, under the alternative of the modified CDFI Fund criteria, only 16 percent of those institutions now without any rural CRA-eligible tracts would have at least one; however, the 16 percent would on average have 54 percent of the rural tracts they serve classified as CRA-eligible. Although the income-based measures affect many more institutions, the average effect on each institution is much smaller. For example, the typical institution that experienced a change under the statewide-median-income alternative would end up with 30 percent of its tracts classified as CRA-eligible. The difference arises from the operation of the modified CDFI Fund criteria at the county level: In our analysis, if a county meets a criterion, then all of its middle-income tracts become CRA-eligible. Each of the other, income-based alternatives is likely to affect only a portion of the middle-income tracts in a given county.

SUMMARY OF FINDINGS

The data and the analyses reported in this article may be useful in evaluating recent proposals to revise the CRA regulations. Because of data limitations, much of the analysis uses indirect rather than direct tests. From these tests, several findings emerge.

First, we found little evidence of differences in retail lending or branching between institutions just below and just above the $250 million threshold that currently distinguishes institutions with small-institution evaluations from those with large-institution evaluations. Nor did we find evidence that institutions graduating from the small-institution evaluation to the large-institution evaluation significantly change their retail lending or branching behavior, at least in the first two years in which they are covered by the large-institution evaluation. However, the analysis was limited to inferences about the behavior of institutions around the margin of the current threshold, $250 million. Although the evidence suggests that raising the threshold some amount above $250 million would not have a significant effect on retail lending or branching, it fails to reveal what amount of increase in the threshold, if any, would result in a significant effect.

Second, in our investigation of the relationship of community development lending to overall CRA ratings for institutions examined under the large-institution examination, we found fairly consistent evidence that such lending plays a relatively limited role in determining overall CRA ratings. Indeed, a significant minority of institutions received "outstanding" ratings and reported no community development loans for three years; this finding holds true in each of several categories of institution asset size.

Third, we found little evidence to support the hypothesis that rural areas receive fewer retail loans or branches from CRA-covered institutions than do urban areas or that rural institutions have more difficulty in achieving "outstanding" ratings. Indeed, smaller rural institutions are equally or more likely to receive "outstanding" ratings than are smaller urban institutions. However, we found modest evidence that rural institutions are somewhat less likely to do any community development lending than are comparable urban institutions and that they make a lower volume of community development loans. These facts support the agencies' restriction of proposed revisions in the criteria for area-based CRA consideration to community development activities.

Finally, in our comparison of several proposals to expand area-based CRA consideration in rural areas, we found that all of the proposals would raise the number of CRA-eligible tracts in rural areas to the same percentage (or higher) as in urban areas. And all would add tracts with better economic characteristics than the tracts classified as lower-income under the 1995 regulations. However, each proposal adds a different set of tracts, affects a different number of banking institutions, and, in the case of institutions that are affected, affects them to different degrees.

1. Banking institutions covered by the CRA, grouped by
selected characteristics and distributed by asset size,
as of December 31, 2003

Number except as noted

                                    Type of institution, by asset
                                      size (millions of dollars)

                                        Large institution (1)

                                                    More than 1,000

                                  Less
                                  than     250-    Recent
Characteristic                   250 (2)   1,000    (3)     Nonrecent

Location (headquarters)
Urban (metropolitan area)
  Center city                       71       463     53         316
  Suburban                          81       456     47         101
Rural (5)
  Exurban                           61       199     10          23
  Remote                            28       144      5           8
U.S.-affiliated area (6)             0         2      1           9

Rating on most recent CRA exam
Outstanding                         34       164     26         168
Satisfactory                       185     1,050     84         282
Needs to improve                     0         7      1           0
Substantial noncompliance            0         0      0           1
None (no exam in 5 years)           22        43      5           6

Type of most recent CRA exam
Large-institution                  129       909     95         412
Small-institution                   79       291      9           3
Other (7)                           11        21      7          36
None (no exam in 5 years)           22        43      5           6

Current regulator (8)
Board                               20       162     14          62
FDIC                               136       639     57         163
OCC                                 70       298     22         145
OTS                                 15       165     23          87

All
Number                             241     1,264    116         457
Percent                            2.7      13.9    1.3           5

MEMO
Median number of days
  between exams                  1,703      963     927       1,035
Share of deposits (percent)         .5      9.6     2.9        75.6

                                 Type of institution, by asset
                                 size (millions of dollars)

                                 Small institution (1)

                                 Less             More
                                 than    150-     than
Characteristic                    150     250    250 (4)

Location (headquarters)
Urban (metropolitan area)
  Center city                      766     294     121
  Suburban                       1,633     415     153
Rural (5)
  Exurban                        1,649     273      67
  Remote                         1,417     182      41
U.S.-affiliated area (6)             6       0       0

Rating on most recent CRA exam
Outstanding                        573     165      50
Satisfactory                     4,405     918     293
Needs to improve                    27       0       1
Substantial noncompliance            3       0       0
None (no exam in 5 years)          463      81      38

Type of most recent CRA exam
Large-institution                    6       2       2
Small-institution                4,973   1,075     339
Other (7)                           29       6       3
None (no exam in 5 years)          463      81      38

Current regulator (8)
Board                              497     124      46
FDIC                             3,413     627     222
OCC                              1,100     280      77
OTS                                461     133      37

All
Number                           5,471   1,164     382
Percent                           60.1    12.8     4.2

MEMO
Median number of days
  between exams                  1,734   1,734   1,657
Share of deposits (percent)        6.0     3.6     1.9

                                      Total

                                                      MEMO:
                                                      Share
                                                       of
                                                    deposits
Characteristic                   Number   Percent   (percent)

Location (headquarters)
Urban (metropolitan area)
  Center city                    2,084     22.9       74.7
  Suburban                       2,886     31.7       15.0
Rural (5)
  Exurban                        2,282     25.1        5.8
  Remote                         1,825     20.1        3.7
U.S.-affiliated area (6)            18       .2         .8

Rating on most recent CRA exam
Outstanding                      1,180       13       54.3
Satisfactory                     7,217     79.4       42.9
Needs to improve                    36       .4         .1
Substantial noncompliance            4       .0         .0
None (no exam in 5 years)          658      7.2        2.7

Type of most recent CRA exam
Large-institution                1,555     17.1       80.2
Small-institution                6,769     74.4       12.8
Other (7)                          113      1.2        4.3
None (no exam in 5 years)          658      7.2        2.7

Current regulator (8)
Board                              925     10.2       17.4
FDIC                             5,257     57.8       24.4
OCC                              1,992     21.9       44.1
OTS                                921     10.1       14.2

All
Number                           9,095      100        100
Percent                            ...      100        100

MEMO
Median number of days
  between exams                  1,645      ...        ...
Share of deposits (percent)        ...      100        100

NOTE. Here and in subsequent tables, "CRA" means Community
Reinvestment Act, and components may not sum to totals because
of rounding.

(1.) Large institutions are banking institutions that reported
2003 data on small-business, small-farm, or community development
lending as required of large institutions under the CRA. All
other institutions are small institutions.

(2.) These institutions are generally part of multibank holding
companies with assets of more than $1 billion and are currently
covered by the large-institution CRA exam.

(3.) "Recent" institutions are banking institutions that had more
than $1 billion in assets as of December 31, 2003, but not in the
two consecutive years before 2003. If the asset-size threshold had
been raised to $1 billion as of year-end 2003, these institutions
would not yet have qualified for the large-institution CRA exam.

(4.) These institutions had more than $250 million in assets as
of December 31, 2003, but failed to qualify for the large-institution
CRA exam because they had not held this amount of assets for two
consecutive years.

(5.) Exurban areas are counties adjacent to metropolitan areas;
remote areas are counties not adjacent to metropolitan areas.

(6.) In this article, U.S.-affiliated areas consist of American
Samoa, Guam, the Commonwealth of the Northern Mariana Islands,
Puerto Rico, and the U.S. Virgin Islands.

(7.) "Other" exams cover strategic-plan, wholesale, and
limited-purpose institutions. A strategic-plan institution
develops its own plan, subject to the approval of a supervising
agency, for evaluating its CRA performance. A wholesale institution
does not extend home mortgage, small-business, small-farm, or
consumer loans to retail customers; a limited-purpose institution
offers a narrow product line, such as one composed of credit card
or motor vehicle loans, to a regional or broader market. Exams for
wholesale and limited-purpose institutions are limited to community
development activities.

(8.) Current regulators are the Board of Governors of the Federal
Reserve System (Board), the Federal Deposit Insurance Corporation
(FDIC), the Office of the Comptroller of the Currency (OCC), and
the Office of Thrift Supervision (OTS).

... Not applicable.

SOURCES. Here and in subsequent tables, except as noted, analyses
incorporate data from one or more of the following sources:
unemployment, Bureau of Labor Statistics (2002); assets and
business loans (as of June 30, 2003; data adjusted through
December 31, 2003, to reflect changes in banking institution
structure), Consolidated Reports of Condition and Income (Call
Report), Federal Deposit Insurance Corporation (2003); branches
and deposits (as of June 30, 2003; data adjusted through December
31, 2003, to reflect changes in banking institution structure),
Summary of Deposits, Federal Deposit Insurance Corporation (2003);
filings under the Home Mortgage Disclosure Act and the Community
Reinvestment Act, Federal Financial Institutions Examination
Council (2003); metropolitan statistical areas, Office of Management
and Budget (2004); census tracts, U.S. Census Bureau (2000); net
migration, Estimated Components of Population Change, Population
Estimates program, U.S. Census Bureau (2000); poverty, Small Area
Income and Poverty Estimates program, U.S. Census Bureau (2002);
rural area designations, Urban Influence Codes, Economic Research
Service, U.S. Department of Agriculture (2003).

2. Bank branches and deposits, grouped by location
and distributed by market status of location,
as of December 31, 2003

                                        Rural

                                                       U.S.-
                                                       affi-
                                                       liated
        Item               Urban    Exurban   Remote    area    Total

Branches
Number
  In defined markets       66,815   10,234    7,806       0     84,855
  Not in defined
      markets                 612    1,034      606     641      2,893
    Total                  67,427   11,268    8,412     641     87,748
Percent
  In defined markets         99.1     90.8     92.8       0       96.7
  Not in defined
      markets                  .9      9.2      7.2     100        3.3
    Total                     100      100      100     100        100

Share of deposits
(percent)
In defined markets           99.4     91.8     92.7       0       97.7
Not in defined
    markets                    .6      8.2      7.3     100        2.3
  Total                       100      100      100     100        100

MEMO
Number of defined
  markets                     563      693      617       0      1,873
Share of deposits in
   defined markets
  (percent)                  88.7      6.9      4.4       0        100
Share of deposits not in
  defined markets
  (percent)                  21.1     26.3     14.9    37.7        100
Average number of
  banking institution
  headquarters per
  defined market             19.1     7.0       6.3       0       10.4

NOTE. In this article, markets are those defined as local banking
markets by the Federal Reserve Banks; these defined areas do not
cover all parts of the country. For markets that span more than one
type of area, location is determined by the area with the largest
percentage of market deposits. For definitions of "exurban" and
"remote," see table 1, note 5. For definition of "U.S.-affiliated
area," see table 1, note 6.

3. Banking markets grouped by location and distributed by share
of market deposits held by institutions that would shift from
large-institution to small-institution CRA examinations under
the agencies' 2005 proposal, as of December 31, 2003

Percent except as noted

                               Share of market deposits
                                  affected (percent)

     Location of
     market (1)          0     1-10   11-20   21-50   51-100   Total

Urban
Number                    88    145     123     173      34      563
Percent                 15.6   25.8    21.9    30.7     6.0      100
Percent weighted by
  market deposits         .7   60.8    26.3    11.1     1.1      100

Rural
Exurban
  Number                 215    100      95     196      87      693
  Percent               31.0   14.4    13.7    28.3    12.6      100
  Percent weighted by
    market deposits     16.9   28.9    14.1    28.5    11.5      100
Remote
  Number                 218     61      88     185      65      617
  Percent               35.3    9.9    14.3    30.0    10.5      100
  Percent weighted by
    market deposits     17.7   15.7    17.3    39.6     9.8      100

All
Number                   521    306     306     554     186    1,873
Percent                 27.8   16.3    16.3    29.6     9.9      100
Percent weighted by
  market deposits        2.6   56.6    25.0    13.6     2.2      100

NOTE. See note to table 2.

(1.) The weighting factor for the weighted percentages is the
amount of deposits in the market location as a
share of total deposits.

4. Banking markets, grouped by location and distributed by
change in CRA reporting status of largest banking institution
with an office in the market, as of December 31, 2003

Percent except as noted

                                    Large
                                   changes
                         Remains     to      Remains
  Location of market      small     small     large    Total

Urban
Number                      25        19       519       563
Percent                    4.4       3.4      92.2       100
Percent weighted
  by market deposits        .1        .1      99.9       100

Rural
Exurban
  Number                    51        50       592       693
  Percent                  7.4       7.2      85.4       100
  Percent weighted
    by market deposits     2.2       3.0      94.8       100

Remote
  Number                    98        89       430       617
  Percent                 15.9      14.4      69.7       100
  Percent weighted
    by market deposits     4.3       7.5      88.2       100

All
Number                     174       158     1,541     1,873
Percent                    9.3       8.4      82.3       100
Percent weighted
  by market deposits        .4        .6      99.0       100

NOTE. See notes to table 3.

5. Characteristics of counties in markets considered potentially
most affected by an increase in the large-institution threshold
to $1 billion, by method of market analysis

Percent except as noted

                                      Method of market analysis and
                                          location of market (1)

                                                        Largest-
                                   Market-share       institution

         Characteristic            Urban    Rural    Urban    Rural

Demographic
Poverty rate, 2002                   12.6     14.0     12.5     16.0
Income per capita,
  2001 (dollars)                   24,304   25,481   21,827   21,040
Real income growth rate
  1996-2001                           6.8      7.0      6.2      5.7
  1981-2001                          33.3     33.0     30.0     26.9
Unemployment rate, 2001               5.2      5.4      5.4      5.4
Population growth rate
  1996-2001                           5.8      2.8      4.6      -.1
  1981-2001                          22.6     22.5     13.5       .1
Net migration rate, 1995-99 (2)       2.4      1.6      4.5       .7

Banking
Branches per 10,000 persons
  Number, 2003                        4.0      5.0      4.6      6.2
  Change, 1998-2003                   -.3      -.2      -.3       .0
Deposits
  Per capita (thousands of
    dollars), 2003                   12.8     14.5     10.9     15.2
  Change, 1998-2003                   -.1      -.2      -.1       .0

MEMO: Counties (data as of 2000)
Number                                 56      199       14      160
Percent with no lower-income
  tracts                             21.4     48.2     50.0     55.0
Percent with only lower-income
  tracts                               .0      3.0      7.1      7.5

                                   National
                                   average
                                   for rural
         Characteristic            counties

Demographic
Poverty rate, 2002                   15.0
Income per capita,
  2001 (dollars)                   21,908
Real income growth rate
  1996-2001                           1.7
  1981-2001                          28.1
Unemployment rate, 2001               5.7
Population growth rate
  1996-2001                           1.7
  1981-2001                          10.5
Net migration rate, 1995-99 (2)       1.2

Banking
Branches per 10,000 persons
  Number, 2003                        5.4
  Change, 1998-2003                   -.1
Deposits
  Per capita (thousands of
    dollars), 2003                   14.0
  Change, 1998-2003                   -.1

MEMO: Counties (data as of 2000)
Number                              2,051 (3)
Percent with no lower-income
  tracts                             48.2 (3)
Percent with only lower-income
  tracts                              5.5 (3)

(1.) For markets that span more than one type of area, location is
determined by the area with the largest percentage of market deposits.
Hence, rural counties may be located in urban markets. When market
boundaries do not correspond to county boundaries, the county is
assigned to the market with the largest share of deposits. Under
the market-share method, potentially most affected markets are
markets in which status-changing institutions (those with assets
between $250 million and $1 billion) held more than 50 percent of
market deposits. Under the large-institution method, potentially
most-affected markets are markets in which the largest institution
with an office in the market was a status-changing institution
(one with assets between $250 million and $1 billion).

(2.) Net migration rate is calculated as the difference in
migration between 1999 and 1995 relative to the estimated
population in 1997.

(3.) U.S. total.

6. Lending and branching activities of banking institutions with
asset sizes close to the current large-institution threshold for
CRA exams, as of December 31, 2003

                                           Asset size
                                     (millions of dollars)

                                      150-250       250-350
           Item                      (Group 1)     (Group 2)

Ratio of loan dollars to
  deposits, by type of loan
All                                78.6 (a), (B)   83.7 (c)
Consumer                            5.6 (a), (B)    7.0 (c)
Business (3)
  Commercial and industrial
    Overall                        11.2             9.9
    Small                           9.3 (a)         7.5
  Commercial real estate
    Overall                        19.0            20.9
    Small                          13.9            13.1
  Farm (4)
    Overall                        5.5 (a)          3.3
    Small                          6.6 (a)          4.1

Ratio of mortgage dollars to
  deposits, by type of mortgage
1-4 family                         27.0 (a), (b)   29.9
Multifamily                         2.2             2.6

Loans to lower-income borrowers
  (percentage points) (5)
Home-purchase                      -3.7            -3.3
Home-improvement                   10.0            10.8

Loans in lower-income areas (5)
Home-purchase                      -7.8           -10.6
Home-improvement                   -4.5            -7.5
Multifamily                         1.6             7.3

Branching activity (6)
Branches per $100 million of
  deposits (number)                 2.9             2.7
Branches in lower-income areas
    (percentage points)            -1.7             1.0
  5-year change in such branches     .0            -1.0

Number of institutions              646              72

                                            Asset size
                                       (millions of dollars)

                                    350-450    Less than 250 (2)
           Item                    (Group 3)       (Group 4)

Ratio of loan dollars to
  deposits, by type of loan
All                                  78.5             88.5
Consumer                              5.0              8.6
Business (3)
  Commercial and industrial
    Overall                          10.3             12.3
    Small                             7.5              9.7
  Commercial real estate
    Overall                          19.8             20.4
    Small                            11.8             13.2
  Farm (4)
    Overall                           3.8              6.0
    Small                             5.9              5.6

Ratio of mortgage dollars to
  deposits, by type of mortgage
1-4 family                           27.6             31.8
Multifamily                           2.5              2.2

Loans to lower-income borrowers
  (percentage points) (5)
Home-purchase                        -3.2             -4.5
Home-improvement                      7.6              4.6

Loans in lower-income areas (5)
Home-purchase                        -4.8            -14.2
Home-improvement                     -6.9              -.7
Multifamily                           8.1              -.6

Branching activity (6)
Branches per $100 million of
  deposits (number)                   2.7              3.0
Branches in lower-income areas
    (percentage points)              -3.5             -1.8
  5-year change in such branches     -2.0             -1.5

Number of institutions                142               49

                                            MEMO:
                                     Crossed threshold
                                   after end of 2001 (1)

           Item                       2001      2003

Ratio of loan dollars to
  deposits, by type of loan
All                                  85.3       85.4
Consumer                              6.4 (d)    5.1
Business (3)
  Commercial and industrial
    Overall                          11.4       10.7
    Small                             9.3        8.0
  Commercial real estate
    Overall                          17.9 (d)   20.9
    Small                            12.8       13.5
  Farm (4)
    Overall                           6.9        6.4
    Small                             8.4        7.3

Ratio of mortgage dollars to
  deposits, by type of mortgage
1-4 family                           33.0       29.9
Multifamily                           4.0        4.7

Loans to lower-income borrowers
  (percentage points) (5)
Home-purchase                         ...        ...
Home-improvement                      ...        ...

Loans in lower-income areas (5)
Home-purchase                         ...        ...
Home-improvement                      ...        ...
Multifamily                           ...        ...

Branching activity (6)
Branches per $100 million of
  deposits (number)                   3.1        2.9
Branches in lower-income areas
    (percentage points)               ...        ...
  5-year change in such branches      ...        ...

Number of institutions                100        100

NOTE. Data are group means adjusted for state, institution (savings
association or commercial bank), location (center city, suburban,
exurban, or remote), and charter effects. Analysis is restricted to
institutions that were examined in the past five years, that were in
existence for at least one year, and that received a "satisfactory"
rating on the small- or large-institution exam. Data exclude
institutions with headquarters in U.S.-affiliated areas and
strategic-plan, wholesale, and limited-purpose institutions (see
table 1, note 7).

(1.) Data are for the 100 banking institutions that were subject to
the small-institution evaluation in 2001 but were subject to the
large-institution evaluation in 2002 and 2003. Differences are omitted
for retail loans extended to lower-income borrowers, retail loans
extended in lower-income areas, and some categories of branching
activity because the lower-income classifications of 2001 were based
on the 1990 census, whereas those of 2003 were based on the 2000
census.

(2.) These institutions are part of multibank holding companies with
assets of more than $1 billion and are currently covered by the
large-institution CRA exam.

(3.) Business loan ratios are calculated as of June 30, 2001, or
June 30, 2003, for comparability with small-loan data. Data have
been adjusted through December 2001 or December 2003 to reflect
changes in banking institution structure.

(4.) Farm lending is measured only for rural commercial banks.
Small farm contains some loans not in overall farm.

(5.) Data cover only urban tracts and institutions that report data
under the Home Mortgage Disclosure Act. Data are the difference
between the average percentage of lending to borrowers in lower-income
census tracts and the average percentage of families that live in
lower-income census tracts in the areas that the institutions serve.

(6.) Branch data are measured as of June 30, 1998; June 30, 2001;
or June 30, 2003. Data have been adjusted through December 1998,
December 2001, or December 2003 to reflect changes in banking
institution structure. Data on lower-income areas are the difference
between the percentage of branches in lower-income census tracts and
the percentage of families that live in lower-income census tracts in
the areas that the institutions serve.

(a.) Difference between group 1 and group 2 is statistically
significant at the 10 percent level.

(b.) Difference between group 1 and group 4 is statistically
significant at the 10 percent level.

(B.) Difference between group 1 and group 4 is statistically
significant at the 1 percent level.

(c.) Difference between group 2 and group 3 is statistically
significant at the 10 percent level.

(d.) Difference between ratio in 2001 and that in 2003 is
statistically significant at the 10 percent level.

... Not applicable.

7. Rating on most recent CRA exam and community
development lending during 2001-03 at large
institutions, by rating and asset size of institution,
as of December 31, 2003

Percent except as noted

                                        Asset size of institution
                                          (millions of dollars)

      Number of institutions
      and status of community                                 More
       development lending,           250-   500-    1,000-   than
           by CRA rating              500    1,000   5,000    5,000

Outstanding
Number of institutions                  23      50       92      65
Made no loans                         26.1    22.0      8.7     6.2
Ranked in bottom half of asset-size
  class                               39.1    44.0     38.0    46.2

Satisfactory
Number of institutions                 241     316      254      69
Made no loans                         29.9    19.0     11.0     5.8
Ranked in bottom half of asset-size
  class                               51.0    51.0     54.3    53.6

NOTE. Analysis is restricted to institutions that were subject to
the large-institution CRA exam each year from 2001 through 2003,
that were in existence for at least one year, that received an
"outstanding" or "satisfactory" rating on the exam, and that had
assets of more than $250 million in 2003. Data exclude strategic-plan,
wholesale, or limited-purpose institutions (see table 1, note 7) and
institutions with headquarters in U.S.-affiliated areas.

8. Distribution of census tracts, population, and families,
by location of tract and tract income relative to wider
area, as of December area, as of December

Percent except as noted

                         Census tracts

                                                                Memo:
                                                               Families
                                                                 with
                                                               incomes
Census tract location                      Popu-                 less
and percent of median   Number   Percent   lation   Families   than 80
    family income                                              percent
     in area (1)                                                of MSA
                                                               or non-
                                                               MSA me-
                                                               dian (2)

Urban
Center city tracts
  Income relative to
  MSA
    Less than 50         3,437      13.1     10.0        8.7     78.2
    50-79                8,004      30.5     29.7       27.7     60.5
    80-90                2,622        10     10.4       10.3     45.8
    90-100               2,503       9.5     10.0       10.2     39.4
    100-119              3,898      14.8     16.3       17.4     31.6
    120 or more          5,814      22.1     23.6       25.7     19.0
      Total             26,278     100      100        100       42.7
  Income relative to
  state
    Less than 80        10,938      41.6     37.8       34.8     64.8
    80-99                4,851      18.5     19.2       19.3     42.9
    100-119              3,900      14.8     15.9       16.7     31.8
    120 or more          6,589      25.1     27.0       29.2     18.0
      Total             26,278     100      100        100       41.4
Suburban tracts
  Income relative to
  MSA
    Less than 50           444       1.7      1.2        1.0     76.9
    50-79                4,456      16.9     15.9       14.6     57.9
    80-90                3,384      12.9     12.5       12.3     45.6
    90-100               4,069      15.5     15.5       15.5     38.9
    100-119              6,563      25.0     25.6       26.2     30.7
    120 or more          7,382      28.1     29.2       30.4     18.5
      Total             26,298     100      100        100       34.5
  Income relative to
  state
    Less than 80         4,592      17.5     15.9       14.4     59.7
    80-99                6,333      24.1     23.5       23.3     42.2
    100-119              6,100      23.2     23.8       24.1     30.8
    120 or more          9,273      35.3     36.8       38.2     17.6
      Total             26,298     100      100        100       32.6

Rural
Exurban tracts
  Income relative to
  MSA
    Less than 50            57        .7       .5         .4     72.9
    50-79                  912      11.9     10.5        9.8     55.1
    80-90                1,157      15.1     14.4       14.1     45.8
    90-100               1,780      23.2     22.7       22.8     39.8
    100-119              2,720      35.5     36.9       37.4     32.6
    120 or more          1,035      13.5     15.0       15.5     23.7
      Total              7,661     100      100        100       37.1
  Income relative to
  state
    Less than 80         3,051      39.8     37.5       36.3     57.4
    80-99                3,394      44.3     45.0       45.7     43.4
    100-119              1,006      13.1     14.5       14.9     33.0
    120 or more            210       2.7      3.0        3.1     22.9
      Total              7,661     100      100        100       46.3
Remote tracts
  Income relative to
  MSA
    Less than 50            39        .8       .6         .4     72.0
    50-79                  794      17.1     15.0       14.4     55.8
    80-90                  927      19.9     18.7       18.7     46.1
    90-100               1,136      24.4     23.4       23.8     39.9
    100-119              1,285      27.6     29.6       30.0     32.9
    120 or more            469      10.1     12.7       12.8     23.3
      Total              4,650     100      100        100       39.3
  Income relative to
  state
    Less than 80         2,240      48.2     44.6       43.9     58.6
    80-99                1,764      37.9     38.4       38.9     43.7
    100-119                487      10.5     12.5       12.6     33.1
    120 or more            159       3.4      4.6        4.5     23.5
      Total              4,650     100      100        100       48.0

Total urban
Income relative to
MSA
  Less than 50           3,881       7.4      5.4        4.4     78.0
  50-79                 12,460      23.7     22.4       20.4     59.5
  80-90                  6,006      11.4     11.5       11.4     45.7
  90-100                 6,572      12.5     12.9       13.1     39.1
  100-119               10,461      19.9     21.3       22.3     31.0
  120 or more           13,196      25.1     26.6       28.3     18.7
    Total               52,576     100      100        100       38.2
Income relative to
state
  Less than 80          15,530      29.5     26.2       23.5     63.1
  80-99                 11,184      21.3     21.5       21.5     42.5
  100-119               10,000      19.0     20.1       20.8     31.2
  120 or more           15,862      30.2     32.2       34.2     17.7
    Total               52,576     100      100        100       36.5

Total rural
Income relative to
MSA
  Less than 50              96        .8       .5         .4     72.6
  50-79                  1,706      13.9     12.1       11.4     55.4
  80-90                  2,084      16.9     15.9       15.7     45.9
  90-100                 2,916      23.7     22.9       23.1     39.8
  100-119                4,005      32.5     34.4       34.9     32.7
  120 or more            1,504      12.2     14.2       14.6     23.6
    Total               12,311     100      100        100       37.9
Income relative to
state
  Less than 80           5,291      43.0     40.0       39.0     57.9
  80-99                  5,158      41.9     42.7       43.3     43.5
  100-119                1,493      12.1     13.8       14.1     33.1
  120 or more              369       3.0      3.5        3.6     23.1
    Total               12,311     100      100        100       46.9

NOTE. Data from the 2000 census are reported for census tracts and
metropolitan statistical areas as determined by 2004 definitions.
Data exclude census tracts in U.S.-affiliated areas.

(1.) Income standard is the median family income in the metropolitan
statistical area (MSA), nonmetropolitan portion of the state
(non-MSA), or state in which the census tract is located.

(2.) For calculations in this column, even when tracts are classified
by state standards, families are still classified by the income in
the MSA or non-MSA in which the family is located.

9. Characteristics of census tracts and the share of selected
areas and of banking institutions without lower-income tracts,
by location of tract, as of December 31, 2003

Percent except as noted

                                                   Urban

                  Item                     Center city   Suburban

Characteristic of census tract
Number of tracts                                26,278     26,298
Average land area per tract
  (square miles)                                   5.5       28.5
Average population density per tract
  (population per square mile)                 9,812.3    3,097.3
Average population per tract                   4,145.3    4,693.8
Percent of national population                    38.7       43.9

Share without lower-income census tracts
Area
County                                            12.0       31.9
Individual assessment area of large
  institution (1)                                  6.2       26.5
Aggregate assessment area
Large institution                                  5.8       14.3
Small institution (2)                            ...        ...

                                                Rural

                  Item                     Exurban   Remote

Characteristic of census tract
Number of tracts                             7,661     4,650
Average land area per tract
  (square miles)                             147.8     322.7
Average population density per tract
  (population per square mile)               494.5     423.4
Average population per tract               4,163.7   3,639.6
Percent of national population                11.3       6.0

Share without lower-income census tracts
Area
County                                        56.9      61.3
Individual assessment area of large
  institution (1)                             45.1      45.7
Aggregate assessment area
Large institution                             36.0      32.7
Small institution (2)                         55.0      56.9

                                                      Total

                  Item                      Urban     Rural      All

Characteristic of census tract
Number of tracts                            52,576    12,311    64,887
Average land area per tract
  (square miles)                              17.0     213.9      54.3
Average population density per tract
  (population per square mile)             6,454.4     467.7   5,318.4
Average population per tract               4,419.6   3,965.7   4,333.4
Percent of national population                82.6      17.4     100.0

Share without lower-income census tracts
Area
County                                        18.0      59.0      44.8
Individual assessment area of large
  institution (1)                             13.8      43.8      23.4
Aggregate assessment area
Large institution                              5.4      30.2      13.9
Small institution (2)                          7.9      53.9      28.3

NOTE. Data exclude U.S.-affiliated areas and tracts without income
information.

(1.) An assessment area consists of the area in which a banking
institution has its main office, branches, and deposit-taking
automated teller machines, as well as the surrounding areas in which
the institution has originated or purchased a substantial portion of
its loans. Assessment areas reported in the 2003 geographies, which
were determined from information supplied by the U.S. Census Bureau,
the U.S. Department of Agriculture, and the Office of Management and
Budget, have been mapped onto the 2004 tract definitions, which use
the Office of Management and Budget's 2004 designations of
metropolitan statistical areas. Large institutions report their
assessment areas each year and may have multiple assessment areas
corresponding to cities or states.

(2.) Aggregate assessment areas were approximated by the counties
in which small institutions had branches.

... Not applicable.

10. Number of retail loans, offices, families, and housing
structures in the assessment areas of large banking institutions
with both urban and rural branches, grouped by asset size of
institution and distributed by location of assessment area, as
of December 31, 2003

Percent

                                         Urban            Rural

  Asset size of institution
      and characteristic         Center city   Suburban   Exurban

$250 million to $500 million

Loans
  Home-purchase                     22.3         30.6      27.7
  Home-improvement                  10.5         32.6      38.3
  Small-business or small-farm      13.7         25.2      31.6
  Multifamily                       42.4         24.7      24.1
Offices (1)                         20.9         28.2      35.8
Families                            35.1         40.8      14.6
Housing structures
  Owner-occupied                    32.2         42.4      15.5
  Multifamily                       59.4         29.6       5.8

$500 million to $1 billion

Loans
  Home-purchase                     25.7         31.1      22.2
  Home-improvement                  21.5         30.6      28.4
  Small-business or small-farm      19.2         23.1      27.1
  Multifamily                       51.6         20.6      13.5
Offices (1)                         24.2         31.6      29.4
Families                            33.6         46.3      11.5
Housing structures
  Owner-occupied                    30.9         47.5      12.6
  Multifamily                       54.2         38.0       3.9

$1 billion to $5 billion

Loans
  Home-purchase                     34.6         38.4      15.6
  Home-improvement                  23.7         36.6      23.4
  Small-business or small-farm      31.4         28.2      20.0
  Multifamily                       55.9         24.5       9.4
Offices (1)                         32.7         34.2      22.2
Families                            39.0         48.7       7.4
Housing structures
  Owner-occupied                    36.2         50.8       7.9
  Multifamily                       57.8         37.1       2.6

More than $5 billion

Loans
  Home-purchase                     36.7         53.7       4.7
  Home-improvement                  29.9         54.6       8.9
  Small-business or small-farm      39.1         44.8       8.2
  Multifamily                       59.7         34.0       3.1
Offices (1)                         41.2         45.2       9.0
Families                            40.0         52.6       4.0
Housing structures
  Owner-occupied                    36.0         55.9       4.4
  Multifamily                       59.3         38.2       1.2

                                 Rural       Total

  Asset size of institution
      and characteristic         Remote   Urban   Rural

$250 million to $500 million

Loans
  Home-purchase                   19.5    52.9    47.1
  Home-improvement                18.6    43.0    57.0
  Small-business or small-farm    29.6    38.8    61.1
  Multifamily                     8.8     67.0    33.0
Offices (1)                       15.1    49.2    50.9
Families                          9.4     76.0    24.0
Housing structures
  Owner-occupied                  9.8     74.7    25.3
  Multifamily                     5.2     89.0    11.0

$500 million to $1 billion

Loans
  Home-purchase                   21.1    56.8    43.2
  Home-improvement                19.5    52.1    47.9
  Small-business or small-farm    30.6    42.3    57.7
  Multifamily                     14.3    72.2    27.8
Offices (1)                       14.8    55.8    44.2
Families                          8.5     79.9    20.1
Housing structures
  Owner-occupied                  9.1     78.3    21.7
  Multifamily                     3.8     92.7     7.3

$1 billion to $5 billion

Loans
  Home-purchase                   11.4    73.0    27.0
  Home-improvement                16.3    60.3    39.7
  Small-business or small-farm    20.4    59.5    40.5
  Multifamily                     10.2    80.3    19.7
Offices (1)                       10.9    66.9    33.1
Families                          4.9     87.7    12.3
Housing structures
  Owner-occupied                  5.2     87.0    13.0
  Multifamily                     2.5     94.9     5.1

More than $5 billion

Loans
  Home-purchase                   4.8     90.4     9.6
  Home-improvement                6.6     84.5    15.5
  Small-business or small-farm    8.0     83.9    16.1
  Multifamily                     3.1     93.8     6.2
Offices (1)                       4.5     86.4    13.5
Families                          3.3     92.7     7.3
Housing structures
  Owner-occupied                  3.7     91.9     8.1
  Multifamily                     1.3     97.5     2.5

NOTE. Analysis is restricted to institutions that were examined
in the past five years under the large-institution CRA exam, that
were in existence for at least one year, that received an
"outstanding" or "satisfactory" rating on the exam, and that had
assets of more than $250 million in 2003. Data exclude
strategic-plan, wholesale, and limited-purpose institutions (see
table 1, note 7) and institutions with headquarters in
U.S.-affiliated areas. Data also exclude census tracts in
U.S.-affiliated areas. For definition of assessment area, see
table 9, note 1.

(1.) Offices consist of headquarters and branches.

11. Retail loan amounts and deposits in the assessment
areas of large banking institutions with urban and rural
branches, grouped by asset size of institution and
distributed by location of assessment area, as of
December 31, 2003

Percent

                                           Urban

Asset size of institution and
  loan amounts and deposits      Center city   Suburban

$250 million to $500 million

Loans
  Home-purchase                     24.9         34.4
  Home-improvement                  12.1         40.6
  Small-business or small-farm      19.4         26.0
  Multifamily                       38.0         32.9
Deposits                            20.1         23.7

$500 million to $1 billion

Loans
  Home-purchase                     27.2         36.0
  Home-improvement                  20.7         31.6
  Small-business or small-farm      26.6         25.7
  Multifamily                       58.9         20.5
Deposits                            29.8         29.1

$1 billion to $5 billion

Loans
  Home-purchase                     34.9         43.3
  Home-improvement                  24.5         43.5
  Small-business or small-farm      40.0         32.2
  Multifamily                       57.8         29.4
Deposits                            43.3         29.3

More than $5 billion

Loans
  Home-purchase                     36.5         57.1
  Home-improvement                  27.1         61.1
  Small-business or small-farm      41.7         43.6
  Multifamily                       59.8         37.2
Deposits                            59.8         32.7

                                        Rural

Asset size of institution and
  loan amounts and deposits      Exurban   Remote

$250 million to $500 million

Loans
  Home-purchase                   24.8      16.0
  Home-improvement                33.0      14.2
  Small-business or small-farm    27.0      27.6
  Multifamily                     23.9       5.2
Deposits                          40.3      15.9

$500 million to $1 billion

Loans
  Home-purchase                   18.8      18.0
  Home-improvement                27.1      20.6
  Small-business or small-farm    23.7      23.9
  Multifamily                      9.8      10.7
Deposits                          27.5      13.6

$1 billion to $5 billion

Loans
  Home-purchase                   12.0       9.7
  Home-improvement                19.2      12.8
  Small-business or small-farm    14.3      13.6
  Multifamily                      6.0       6.8
Deposits                          18.8       8.5

More than $5 billion

Loans
  Home-purchase                    2.9      3.4
  Home-improvement                 6.2      5.5
  Small-business or small-farm     7.4      7.2
  Multifamily                      1.3      1.7
Deposits                           5.2      2.3

                                      Total

Asset size of institution and
  loan amounts and deposits      Urban   Rural

$250 million to $500 million

Loans
  Home-purchase                  59.2    40.8
  Home-improvement               52.7    47.3
  Small-business or small-farm   45.3    54.7
  Multifamily                    71.0    29.0
Deposits                         43.8    56.2

$500 million to $1 billion

Loans
  Home-purchase                  63.2    36.8
  Home-improvement               52.3    47.7
  Small-business or small-farm   52.3    47.6
  Multifamily                    79.4    20.5
Deposits                         58.9    41.1

$1 billion to $5 billion

Loans
  Home-purchase                  78.2    21.7
  Home-improvement               68.0    32.0
  Small-business or small-farm   72.2    27.9
  Multifamily                    87.2    12.8
Deposits                         72.6    27.3

More than $5 billion

Loans
  Home-purchase                  93.6     6.4
  Home-improvement               88.2    11.7
  Small-business or small-farm   85.3    14.6
  Multifamily                    97.0     3.0
Deposits                         92.5     7.5

NOTE: See general note to table 10.

12. Proportion of large banking institutions with both urban and
rural branches that overserve parts of their assessment areas in
terms of either number of loans or loan amount, by asset size of
institution, type of loan, and location of assessment area, as
of December 31, 2003

                                          Rural

       Loan measure and
          loan type              Exurban   Remote   Total   Urban

Number of loans, by asset size
of institution

$250 million to $500 million
  Home-purchase                   55.6      35.5    62.9    37.1
  Home-improvement                63.2      35.1    77.2    22.8
  Small-business or
    small-farm                    51.8      38.7    59.9    40.1
  Multifamily                     32.1      17.3    43.2    56.8
$500 million to $1 billion
  Home-purchase                   53.8      43.2    65.1    34.9
  Home-improvement                55.3      42.7    68.0    32.0
  Small-business or
    small-farm                    56.2      43.8    66.9    33.1
  Multifamily                     39.3      29.5    53.6    46.4
$1 billion to $5 billion
  Home-purchase                   65.8      44.7    71.1    28.9
  Home-improvement                67.6      47.9    76.8    23.2
  Small-business or
    small-farm                    60.8      45.8    67.3    32.7
  Multifamily                     40.6      32.3    53.4    46.6
More than $5 billion
  Home-purchase                   49.5      49.5    53.8    46.2
  Home-improvement                66.7      51.9    64.2    35.8
  Small-business or
    small-farm                    64.0      64.0    67.4    32.6
  Multifamily                     51.8      43.5    61.2    38.8

Loan amount, by asset size
of institution

$250 million to $500 million
  Home-purchase                   29.0      32.3    37.9    62.1
  Home-improvement                39.5      37.7    52.6    47.4
  Small-business or
    small-farm                    28.5      37.2    48.2    51.8
  Multifamily                     19.8      17.3    29.6    70.4
$500 million to $1 billion
  Home-purchase                   26.0      37.9    44.4    55.6
  Home-improvement                36.7      38.0    52.7    47.3
  Small-business or
    small-farm                    31.5      43.8    49.4    50.6
  Multifamily                     23.2      21.4    34.8    65.2
$1 billion to $5 billion
  Home-purchase                   30.3      46.1    41.4    58.6
  Home-improvement                50.0      49.3    62.0    38.0
  Small-business or
    small-farm                    26.8      47.1    41.8    58.2
  Multifamily                     15.0      22.6    26.3    73.7
More than $5 billion
  Home-purchase                   34.4      62.4    50.5    49.5
  Home-improvement                54.3      77.8    66.7    33.3
  Small-business or
    small-farm                    40.4      75.3    69.7    30.3
  Multifamily                     11.8      31.8    21.2    78.8

NOTE. See general note to table 10. Overserving by an institution
in part of its assessment areas is measured by the ratio of the
number of loans or the aggregate loan amount in that part to the
number of owner-occupied housing structures (in the case of
home-purchase and home-improvement loans), or to the number of
families (in the case of small-business or small-farm loans), or
to the number of multifamily housing structures (in the case of
multifamily loans) in that part. An institution overserves in part
of its assessment areas for a particular loan type if the ratio in
the part, either for number of loans or loan amount, exceeds the
average ratio for all the institution's assessment areas.

13. Share of large banking institutions that received an
"outstanding" rating on their most recent large-institution
CRA exam and the extent of community development lending among
large institutions, by asset size of institution and location
of headquarters, as of December 31, 2003

                                     Urban               Rural

   Characteristic and          Center
asset size of institution       city     Suburban   Exurban   Remote

"Outstanding" rating
$250 million to $500 million
  Number                           137       163        95        62
  Percent                          8.0       9.8       6.3      14.5
$500 million to $1 billion
  Number                           184       171        47        39
  Percent                         13.0      13.5      29.8      15.4
$1 billion to $5 billion
  Number                           217       113        32        12
  Percent                         29.0      28.3       3.1      16.2
More than $5 billion
  Number                           144        24         7         1
  Percent                         52.1      45.8        .0        .0

Made community development
  loans in 2003 (1)
$250 million to $500 million
  Percent                         63.0      53.8      56.3      53.8
  Average amount (thousands
    of dollars)                  3,833     3,164     2,074     1,346
$500 million to $1 billion
  Percent                         74.7      65.3      55.6      73.5
  Average amount (thousands
    of dollars)                  7,321     5,117     2,363     7,009
$1 billion to $5 billion
  Percent                         84.0      74.5      68.0      72.7
  Average amount (thousands
    of dollars)                 24,073    22,106     9,070    13,338
More than $5 billion
  Percent                         90.8      84.2     100.0     100.0
  Average amount (thousands
    of dollars)                291,814   188,318    19,945     6,716

NOTE. See general note to table 10.

(1.) Average amount of loans was among institutions with such lending.

14. Comparison of effects, on census tracts and on counties, of
options for defining census tracts as CRA-eligible, by location
of tract, as of December 31, 2003

Percent

                                                 Urban

                                           Center
                  Item                      city    Suburban

Current rule
Less than 80 percent of MSA or non-MSA
    median (1)
  CRA-eligible tracts                        43.5      18.6
    Population                               39.7      17.2
    Loans
      Small-business or small-farm           30.7      16.7
      Home-purchase                          20.3      13.8
      Multifamily                            43.2      23.0
    Branches                                 33.1      20.6
    Deposits                                 38.9      22.1
  Counties without CRA-eligible tracts       12.0      31.9
  Counties with only CRA-eligible tracts       .6       2.8

Options
Less than 90 percent of MSA or non-MSA
    median (1)
  CRA-eligible tracts                        53.5      31.5
    Population                               50.1      29.7
    Loans
      Small-business or small-farm           41.2      33.5
      Home-purchase                          30.2      30.4
      Multifamily                            54.7      39.6
    Branches                                 43.9      39.7
    Deposits                                 50.0      41.5
  Counties without CRA-eligible tracts        8.5      13.2
  Counties with only CRA-eligible tracts      1.9       9.7

Less than 100 percent of MSA or non-MSA
    median (1)
  CRA-eligible tracts                        63.0      47.0
    Population                               60.0      45.2
    Loans
      Small-business or small-farm           52.5      54.2
      Home-purchase                          42.4      49.8
      Multifamily                            65.9      58.9
    Branches                                 56.1      59.7
    Deposits                                 60.7      60.9
  Counties without CRA-eligible tracts        5.8       3.2
  Counties with only CRA-eligible tracts      4.1      20.4

Less than 80 percent of state median (2)
  CRA-eligible tracts                        41.6      17.5
    Population                               37.8      15.9
    Loans
      Small-business or small-farm           32.4      15.9
      Home-purchase                          22.0      13.8
      Multifamily                            44.1      21.4
    Branches                                 34.6      19.6
    Deposits                                 40.4      20.9
  Counties without CRA-eligible tracts       12.2      37.4
  Counties with only CRA-eligible tracts      1.2       3.2

Modified CDFI Fund criteria (3)
  Combined
    CRA-eligible tracts                      49.1      22.9
      Population                             45.1      21.5
      Loans
        Small-business or small-farm         36.4      22.5
        Home-purchase                        26.6      20.3
        Multifamily                          49.0      28.8
      Branches                               38.9      26.5
      Deposits                               44.2      27.9
    Counties without CRA-eligible tracts     11.2      29.6
    Counties with only CRA-eligible
      tracts                                  1.6       7.8
  Unemployment
    CRA-eligible tracts                      46.0      20.5
      Population                             42.1      19.1
      Loans
        Small-business or small-farm         33.2      20.0
        Home-purchase                        22.9      17.6
        Multifamily                          45.7      26.4
      Branches                               35.3      23.9
      Deposits                               40.8      25.3
  Counties without CRA-eligible tracts       11.6      30.6
  Counties with only CRA-eligible tracts      1.2       5.7

  Poverty
    CRA-eligible tracts                      45.5      19.1
      Population                             41.5      17.7
      Loans
        Small-business or small-farm         31.6      17.7
        Home-purchase                        21.2      15.1
        Multifamily                          44.3      23.6
      Branches                               33.7      21.7
      Deposits                               39.4      23.2
  Counties without CRA-eligible tracts       12.0      31.5
  Counties with only CRA-eligible tracts       .6       4.4

  Population loss
    CRA-eligible tracts                      47.8      21.0
      Population                             43.5      19.4
      Loans
        Small-business or small-farm         33.9      18.8
        Home-purchase                        24.1      16.0
        Multifamily                          46.2      25.1
      Branches                               36.7      22.7
      Deposits                               42.3      24.2
  Counties without CRA-eligible tracts       11.6      31.1
  Counties with only CRA-eligible tracts      1.0       3.9

MEMO
Number of tracts                           26,278    26,298

                                                Rural

                  Item                     Exurban   Remote

Current rule
Less than 80 percent of MSA or non-MSA
    median (1)
  CRA-eligible tracts                        12.6     17.9
    Population                               11.1     15.6
    Loans
      Small-business or small-farm           10.9     13.1
      Home-purchase                           9.1     12.7
      Multifamily                            14.9     13.3
    Branches                                 13.8     15.7
    Deposits                                 15.2     16.7
  Counties without CRA-eligible tracts       56.9     61.3
  Counties with only CRA-eligible tracts      1.7      4.8

Options
Less than 90 percent of MSA or non-MSA
    median (1)
  CRA-eligible tracts                        27.8     37.8
    Population                               25.4     34.3
    Loans
      Small-business or small-farm           27.4     36.4
      Home-purchase                          25.2     34.7
      Multifamily                            28.2     30.7
    Branches                                 32.3     40.8
    Deposits                                 33.7     41.6
  Counties without CRA-eligible tracts       29.5     31.3
  Counties with only CRA-eligible tracts      6.8     18.3

Less than 100 percent of MSA or non-MSA
    median (1)
  CRA-eligible tracts                        51.0     62.3
    Population                               48.1     57.7
    Loans
      Small-business or small-farm           52.8     65.8
      Home-purchase                          50.7     62.4
      Multifamily                            49.7     53.8
    Branches                                 58.3     69.4
    Deposits                                 59.7     69.8
  Counties without CRA-eligible tracts       10.0     10.3
  Counties with only CRA-eligible tracts     21.3     42.3

Less than 80 percent of state median (2)
  CRA-eligible tracts                        39.8     48.2
    Population                               37.5     44.6
    Loans
      Small-business or small-farm           40.0     46.9
      Home-purchase                          38.3     45.5
      Multifamily                            38.3     37.6
    Branches                                 46.2     51.8
    Deposits                                 47.3     52.4
  Counties without CRA-eligible tracts       18.7     24.4
  Counties with only CRA-eligible tracts     15.3     28.2

Modified CDFI Fund criteria (3)
  Combined
    CRA-eligible tracts                      29.6     38.1
      Population                             27.9     34.5
      Loans
        Small-business or small-farm         31.1     38.6
        Home-purchase                        30.3     36.1
        Multifamily                          29.1     31.3
      Branches                               33.9     41.3
      Deposits                               34.9     42.1
    Counties without CRA-eligible tracts     45.9     44.1
    Counties with only CRA-eligible
      tracts                                 24.3     33.7
  Unemployment
    CRA-eligible tracts                      23.7     26.8
      Population                             22.4     24.3
      Loans
        Small-business or small-farm         23.7     22.8
        Home-purchase                        22.2     23.4
        Multifamily                          24.5     21.9
      Branches                               26.4     25.1
      Deposits                               27.6     25.9
  Counties without CRA-eligible tracts       49.3     55.9
  Counties with only CRA-eligible tracts     14.7     16.3

  Poverty
    CRA-eligible tracts                      19.0     25.8
      Population                             17.3     23.7
      Loans
        Small-business or small-farm         18.0     22.4
        Home-purchase                        17.1     21.6
        Multifamily                          19.8     19.7
      Branches                               20.9     25.1
      Deposits                               22.1     25.9
  Counties without CRA-eligible tracts       54.9     57.0
  Counties with only CRA-eligible tracts     12.3     17.6

  Population loss
    CRA-eligible tracts                      15.2     25.7
      Population                             13.3     22.0
      Loans
        Small-business or small-farm         14.5     24.9
        Home-purchase                        12.9     22.3
        Multifamily                          16.5     20.5
      Branches                               17.8     28.3
      Deposits                               19.1     29.2
  Counties without CRA-eligible tracts       54.1     51.2
  Counties with only CRA-eligible tracts      5.7     17.8

MEMO
Number of tracts                            7,661    4,650

                                                    Total

                  Item                     Urban    Rural     All

Current rule
Less than 80 percent of MSA or non-MSA
    median (1)
  CRA-eligible tracts                        31.1     14.6     28.0
    Population                               27.7     12.6     25.1
    Loans
      Small-business or small-farm           22.5     11.9     15.8
      Home-purchase                          16.8     10.7     12.9
      Multifamily                            33.5     14.3     25.5
    Branches                                 25.8     14.7     18.6
    Deposits                                 30.0     15.9     20.9
  Counties without CRA-eligible tracts       18.0     59.0     44.8
  Counties with only CRA-eligible tracts      2.4      3.2      2.9

Options
Less than 90 percent of MSA or non-MSA
    median (1)
  CRA-eligible tracts                        42.5     31.6     40.4
    Population                               39.3     28.5     37.4
    Loans
      Small-business or small-farm           38.7     31.6     34.2
      Home-purchase                          33.0     29.5     30.7
      Multifamily                            48.0     29.1     40.1
    Branches                                 43.9     36.4     39.0
    Deposits                                 48.3     37.5     41.3
  Counties without CRA-eligible tracts        5.9     30.4     21.9
  Counties with only CRA-eligible tracts      8.5     12.4     11.1

Less than 100 percent of MSA or non-MSA
    median (1)
  CRA-eligible tracts                        55.0     55.3     55.1
    Population                               52.1     51.4     52.0
    Loans
      Small-business or small-farm           56.7     58.8     58.1
      Home-purchase                          50.2     56.1     54.0
      Multifamily                            63.9     51.2     58.6
    Branches                                 61.7     63.6     63.0
    Deposits                                 64.9     64.6     64.7
  Counties without CRA-eligible tracts        1.3     10.1      7.1
  Counties with only CRA-eligible tracts     17.1     31.5     26.5

Less than 80 percent of state median (2)
  CRA-eligible tracts                        29.5     43.0     32.1
    Population                               26.2     40.0     28.6
    Loans
      Small-business or small-farm           21.4     43.2     35.2
      Home-purchase                          16.4     41.6     32.8
      Multifamily                            32.3     38.0     34.7
    Branches                                 24.8     48.9     40.5
    Deposits                                 28.7     49.8     42.4
  Counties without CRA-eligible tracts       23.2     21.5     22.1
  Counties with only CRA-eligible tracts      2.8     21.6     15.0

Modified CDFI Fund criteria (3)
  Combined
    CRA-eligible tracts                      36.0     32.8     35.4
      Population                             32.6     30.2     32.1
      Loans
        Small-business or small-farm         28.0     34.6     32.2
        Home-purchase                        23.3     33.0     29.6
        Multifamily                          39.1     29.9     35.3
      Branches                               31.6     37.5     35.4
      Deposits                               35.4     38.3     37.3
    Counties without CRA-eligible tracts     16.9     45.1     35.3
    Counties with only CRA-eligible
      tracts                                  6.8     28.9     21.2
  Unemployment
    CRA-eligible tracts                      33.3     24.9     31.7
      Population                             29.9     23.1     28.7
      Loans
        Small-business or small-farm         25.5     23.3     24.1
        Home-purchase                        20.4     22.8     21.9
        Multifamily                          36.6     23.6     31.1
      Branches                               28.7     25.8     26.8
      Deposits                               32.7     26.8     28.9
  Counties without CRA-eligible tracts       17.3     52.5     40.3
  Counties with only CRA-eligible tracts      5.1     15.5     11.8

  Poverty
    CRA-eligible tracts                      32.3     21.6     30.3
      Population                             28.9     19.5     27.2
      Loans
        Small-business or small-farm         23.5     20.1     21.3
        Home-purchase                        18.1     19.2     18.8
        Multifamily                          34.3     19.8     28.2
      Branches                               26.9     22.9     24.3
      Deposits                               31.0     23.9     26.4
  Counties without CRA-eligible tracts       17.8     55.9     42.7
  Counties with only CRA-eligible tracts      3.9     14.9     11.1

  Population loss
    CRA-eligible tracts                      34.4     19.2     31.5
      Population                             30.7     16.3     28.2
      Loans
        Small-business or small-farm         24.7     19.4     21.3
        Home-purchase                        19.3     17.2     17.9
        Multifamily                          35.9     18.0     28.4
      Branches                               28.2     22.9     24.7
      Deposits                               32.3     24.0     26.9
  Counties without CRA-eligible tracts       17.8     52.7     40.6
  Counties with only CRA-eligible tracts      3.1     11.6      8.6

MEMO
Number of tracts                           52,576   12,311   64,887

NOTE. See general note to table 9, and for description of lending and
branch data reported in 2003 geographies, see related description for
assessment areas in table 9, note 1. Analysis is restricted to lending
done within assessment areas and excludes institutions not covered by
the CRA.

(1.) Median family income in census tract as a percentage of the
median family income in the metropolitan statistical area (MSA) or
nonmetropolitan portion of the state (non-MSA) in which the census
tract is located.

(2.) Median family income in census tract as a percentage of the
median family income in the state in which the census tract is
located.

(3.) For description of modification to CDFI Fund criteria, see
text. CDFI Fund Community Development Financial Institutions Fund.

15. Characteristics of CRA-eligible census tracts and counties and
of those that would be added under options for defining census
tracts as CRA-eligible, by location of tract, as of December 31, 2003

Percent except as noted

                                          Urban           Rural

                                    Center
              Item                   city     Suburban   Exurban

Current rule
Less than 80 percent of MSA or
    non-MSA median (1)
  Tract (average characteristics)
    Number added                     11,441      4,900       969
    Share of families with income
      Less than or equal to
        poverty level                  23.9       15.6      22.9
      Less than 50 percent of MSA
        or non-MSA median (1)          43.8       35.3      34.5
      Between 50 percent and 80
        percent of MSA or non-MSA
        median (1)                     22.7       24.8      22.2
    Median family income
        (dollars)                    30,067     36,343    27,741
      Median relative to MSA or
        non-MSA                        57.5       66.6      69.6
    Housing
      Median house value
        (dollars)                    98,973     94,019    53,528
      Median house age (years)         30.6       33.1      33.9
      Occupancy by owner               35.3       48.1      53.5
      Vacancy rate (2)                  9.8        9.2      15.6
    Population
      Over age 65                      10.7       12.7      14.1
      Minority (3)                     66.7       45.9      40.8
  County (average
      characteristics)
    Population change, 1990-2000        1.1        1.2       1.1
    Net migration rate,
      1995-99 (4)                      -1.3        2.4        .8
    Poverty rate, 2002                 14.1       11.4      17.5
    Unemployment rate, 2001             6.0        5.6       7.6

Options
Less than 90 percent of MSA or
    non-MSA median (1)
  Tract (average characteristics)
    Number added                      2,622      3,384     1,157
    Share of families with income
      Less than or equal to
        poverty level                  10.0        8.2      13.7
      Less than 50 percent of MSA
        or non-MSA median (1)          23.9       22.8      24.4
      Between 50 percent and 80
        percent of MSA or non-MSA
        median (1)                     21.9       22.9      21.4
    Median family income
        (dollars)                    44,231     45,911    34,668
      Median relative to MSA or
        non-MSA                        85.1       85.2      85.6
    Housing
      Median house value
        (dollars)                   121,304    103,360    64,487
      Median house age (years)         34.1       35.9      37.2
      Occupancy by owner               52.3       63.8      61.3
      Vacancy rate (2)                  6.4        8.1      15.4
    Population
      Over age 65                      13.0       14.0      15.8
      Minority (3)                     40.5       24.7      19.7
  County (average
      characteristics)
    Population change, 1990-2000        1.1        1.2       1.1
    Net migration rate,
      1995-99 (4)                        .1        3.0       2.4
    Poverty rate, 2002                 13.0       10.5      14.7
    Unemployment rate, 2001             5.8        5.4       6.6

Less than 100 percent of MSA or
    non-MSA median (1)
  Tract (average characteristics)
    Number added                      5,125      7,453     2,937
    Share of families with income
      Less than or equal to
        poverty level                   9.0        7.1      11.9
      Less than 50 percent of MSA
        or non-MSA median (1)          21.8       20.3      21.7
      Between 50 percent and 80
        percent of MSA or non-MSA
        median (1)                     20.9       21.6      20.4
    Median family income
        (dollars)                    46,727     48,883    36,962
      Median relative to MSA or
        non-MSA                        89.0       90.6      91.3
    Housing
      Median house value
        (dollars)                   125,857    110,561    68,236
      Median house age (years)         34.6       36.1      37.4
      Occupancy by owner               54.6       66.4      63.0
      Vacancy rate (2)                  6.1        7.5      15.0
    Population
      Over age 65                      13.2       13.7      15.7
      Minority (3)                     37.2       21.5      16.6
  County (average
      characteristics)
    Population change, 1990-2000        1.1        1.2       1.1
    Net migration rate,
      1995-99 (4)                        .2        2.8       2.3
    Poverty rate, 2002                 12.8       10.3      14.1
    Unemployment rate, 2001             5.8        5.3       6.4

Less than 80 percent of state
    median (5)
  Tract (average characteristics)
    Number added                        902      1,100     2,082
    Share of families with income
      Less than or equal to
        poverty level                  14.2       11.8      12.7
      Less than 50 percent of MSA
        or non-MSA median1             23.9       22.9      22.5
      Between 50 percent and 80
        percent of MSA or non-MSA
        median (1)                     20.6       21.2      20.5
    Median family income
        (dollars)                    37,626     39,441    35,969
      Median relative to MSA or
        non-MSA                        88.0       88.6      90.3
    Housing
      Median house value
        (dollars)                    98,707     92,916    66,651
      Median house age (years)         32.9       35.5      37.4
      Occupancy by owner               51.7       59.6      61.9
      Vacancy rate (2)                  7.6        9.9      15.7
    Population
      Over age 65                      12.8       14.1      15.9
      Minority (3)                     48.5       33.9      18.7
  County (average
      characteristics)
    Population change, 1990-2000        1.1        1.1       1.1
    Net migration rate,
      1995-99 (4)                       -.8         .6       2.5
    Poverty rate, 2002                 16.6       14.0      14.6
    Unemployment rate, 2001             7.2        6.9       6.5

Modified CDFI Fund criteria (6)
  Combined
    Tract (average
        characteristics)
      Number added                    1,466      1,133     1,296
      Share of families with
          income
        Less than or equal to
          poverty level                10.8        9.2      13.6
        Less than 50 percent of
          MSA or non-MSA               20.8       18.4      21.6
          median (1)
        Between 50 percent and 80
          percent of MSA or non-
          MSA median (1)               18.4       19.0      18.4
      Median family income
          (dollars)                  46,066     46,384    36,635
        Median relative to MSA or
          non-MSA                      97.2       99.5      96.7
      Housing
        Median house value
          (dollars)                 125,820    107,152    66,391
        Median house age (years)       34.8       35.2      36.9
        Occupancy by owner             53.8       65.2      64.4
        Vacancy rate (2)                6.6        8.2      15.2
      Population
        Over age 65                    13.7       13.0      14.7
        Minority (3)                   47.2       32.2      25.6
    County (average
        characteristics)
      Population change,
        1990-2000                       1.1        1.1       1.1
      Net migration rate,
        1995-99 (4)                    -4.7       -1.6       0.9
      Poverty rate, 2002               18.6       14.9      17.6
      Unemployment rate, 2001           7.8        8.1       8.6

  Unemployment
    Tract (average
        characteristics)
      Number added                      653        502       848
      Share of families with
          income
        Less than or equal to
          poverty level                13.3       11.4      12.4
        Less than 50 percent of
          MSA or non-MSA               21.8       19.8      21.0
          median (1)
        Between 50 percent and 80
          percent of MSA or non-
          MSA median (1)               17.5       19.0      18.7
      Median family income
          (dollars)                  44,263     43,430    37,586
        Median relative to MSA or
          non-MSA                      98.1       98.0      96.8
      Housing
        Median house value
          (dollars)                 157,326    110,582    71,274
        Median house age (years)       33.8       34.0      37.5
        Occupancy by owner             49.4       64.1      65.8
        Vacancy rate (2)                6.4       10.2      14.5
      Population
        Over age 65                    12.9       11.9      15.0
        Minority (3)                   53.2       33.9      21.4
    County (average
        characteristics)
      Population change,
        1990-2000                       1.1        1.2       1.1
      Net migration rate,
        1995-99 (4)                    -2.0        2.4       2.1
      Poverty rate, 2002               21.5       16.0      16.3
      Unemployment rate, 2001           9.9       10.8      10.0

Poverty
  Tract (average characteristics)
    Number added                        521        131       485
    Share of families with income
      Less than or equal to
        poverty level                  15.0       18.8      17.4
      Less than 50 percent of MSA
        or non-MSA median (1)          23.1       22.6      24.1
      Between 50 percent and 80
        percent of MSA or non-MSA
        median (1)                     17.1       18.3      17.6
    Median family income
        (dollars)                    42,919     34,671    33,820
      Median relative to MSA or
        non-MSA                        97.3       96.0        95
    Housing
      Median house value
        (dollars)                   165,803     74,434    56,220
      Median house age (years)         33.7       31.2      35.9
      Occupancy by owner               45.9       61.1      63.7
      Vacancy rate (2)                  6.7       11.9      16.0
    Population
      Over age 65                      12.8       11.1      14.3
      Minority (3)                     60.9       59.1      37.7
  County (average
      characteristics)
    Population change, 1990-2000        1.1        1.2       1.1
    Net migration rate,
      1995-99 (4)                      -4.0        1.3        .6
    Poverty rate, 2002                 24.7       24.0      22.3
    Unemployment rate, 2001             9.0       11.5       7.8

Population loss
  Tract (average characteristics)
    Number added                      1,124        613       194
    Share of families with income
      Less than or equal to
        poverty level                   9.3        7.1      14.1
      Less than 50 percent of MSA
        or non-MSA median (1)          20.8       17.0      21.1
      Between 50 percent and 80
        percent of MSA or non-MSA
        median (1)                     18.6       19.0      18.0
    Median family income
        (dollars)                    48,241     49,149    36,513
      Median relative to MSA or
        non-MSA                        96.9      101.1      98.7
    Housing
      Median house value
        (dollars)                   135,845    106,363    55,698
      Median house age (years)         35.8       36.3      35.6
      Occupancy by owner               52.5       65.9      58.8
      Vacancy rate (2)                  6.2        6.5      16.7
    Population
      Over age 65                      14.5       14.0      14.3
      Minority (3)                     44.6       30.3      28.1
  County (average
      characteristics)
    Population change, 1990-2000        1.0        1.0       1.0
    Net migration rate,
      1995-99 (4)                      -6.8       -6.1      -7.1
    Poverty rate, 2002                 18.2       13.8      17.0
    Unemployment rate, 2001             6.8        6.1       6.3

                                    Rural             Total

              Item                  Remote    Urban    Rural      All

Current rule
Less than 80 percent of MSA or
    non-MSA median (1)
  Tract (average characteristics)
    Number added                       833    16,341    1,802    18,143
    Share of families with income
      Less than or equal to
        poverty level                 23.1      21.4     23.0      21.6
      Less than 50 percent of MSA
        or non-MSA median (1)         34.7      41.3     34.6      40.6
      Between 50 percent and 80
        percent of MSA or non-MSA
        median (1)                    22.6      23.3     22.4      23.2
    Median family income
        (dollars)                   27,090    31,949   27,440    31,501
      Median relative to MSA or
        non-MSA                         69      60.2     69.3      61.1
    Housing
      Median house value
        (dollars)                   51,056    97,480   52,389    92,969
      Median house age (years)        35.1      31.4     34.5      31.7
      Occupancy by owner              53.3      39.2     53.5      40.6
      Vacancy rate (2)                18.4       9.7     16.9      10.4
    Population
      Over age 65                     14.6      11.3     14.4      11.6
      Minority (3)                    33.4      60.4     37.4      58.1
  County (average
      characteristics)
    Population change, 1990-2000         1       1.1      1.1       1.1
    Net migration rate,
      1995-99 (4)                      -.9       -.2        0       -.2
    Poverty rate, 2002                19.3      13.3     18.3      13.8
    Unemployment rate, 2001            7.2       5.9      7.4       6.1

Options
Less than 90 percent of MSA or
    non-MSA median (1)
  Tract (average characteristics)
    Number added                       927     6,006    2,084     8,090
    Share of families with income
      Less than or equal to
        poverty level                 13.3       9.0     13.5      10.2
      Less than 50 percent of MSA
        or non-MSA median (1)         24.1      23.3     24.3      23.5
      Between 50 percent and 80
        percent of MSA or non-MSA
        median (1)                    21.9      22.4     21.6      22.2
    Median family income
        (dollars)                   34,489    45,178   34,588    42,450
      Median relative to MSA or
        non-MSA                       85.3      85.2     85.5      85.3
    Housing
      Median house value
        (dollars)                   59,496   111,172   62,269    98,569
      Median house age (years)        38.2      35.1     37.6      35.7
      Occupancy by owner              59.3      58.8     60.4      59.2
      Vacancy rate (2)                18.6       7.3     16.8       9.8
    Population
      Over age 65                     16.7      13.6     16.2      14.2
      Minority (3)                    15.6      31.6     17.9      28.1
  County (average
      characteristics)
    Population change, 1990-2000       1.0       1.1      1.1       1.1
    Net migration rate,
      1995-99 (4)                       .3       1.8      1.5       1.7
    Poverty rate, 2002                15.4      11.6     15.0      12.5
    Unemployment rate, 2001            6.1       5.6      6.3       5.8

Less than 100 percent of MSA or
    non-MSA median (1)
  Tract (average characteristics)
    Number added                     2,063    12,578    5,000    17,578
    Share of families with income
      Less than or equal to
        poverty level                 11.7       7.9     11.8       9.0
      Less than 50 percent of MSA
        or non-MSA median (1)         21.8      20.9     21.7      21.1
      Between 50 percent and 80
        percent of MSA or non-MSA
        median (1)                    20.8      21.3     20.6      21.1
    Median family income
        (dollars)                   36,682    48,004   36,847    44,830
      Median relative to MSA or
        non-MSA                       90.6      90.3       91      90.5
    Housing
      Median house value
        (dollars)                   64,522   116,781   66,704   102,523
      Median house age (years)        38.5      35.5     37.8      36.2
      Occupancy by owner              61.1      61.6     62.2      61.8
      Vacancy rate (2)                17.5       7.0     16.1       9.6
    Population
      Over age 65                     16.8      13.5     16.1      14.3
      Minority (3)                    13.5      27.9     15.3      24.4
  County (average
      characteristics)
    Population change, 1990-2000       1.0       1.1      1.1       1.1
    Net migration rate,
      1995-99 (4)                       .5       1.7      1.5       1.7
    Poverty rate, 2002                14.4      11.3     14.2      12.1
    Unemployment rate, 2001            5.7       5.5      6.1       5.7

Less than 80 percent of state
    median (5)
  Tract (average characteristics)
    Number added                     1,407     2,002    3,489     5,491
    Share of families with income
      Less than or equal to
        poverty level                 12.4      12.9     12.6      12.7
      Less than 50 percent of MSA
        or non-MSA median1            22.6      23.4     22.6      22.9
      Between 50 percent and 80
        percent of MSA or non-MSA
        median (1)                    21.1      20.9     20.7      20.8
    Median family income
        (dollars)                   35,755    38,623   35,883    36,882
      Median relative to MSA or
        non-MSA                       89.0      88.3     89.7      89.2
    Housing
      Median house value
        (dollars)                   62,111    95,517   64,821    76,000
      Median house age (years)        38.2      34.3     37.7      36.5
      Occupancy by owner              60.8      56.1     61.4      59.5
      Vacancy rate (2)                18.1       8.8     16.6      13.8
    Population
      Over age 65                     16.7      13.5     16.2      15.3
      Minority (3)                    14.5      40.5     17.0      25.6
  County (average
      characteristics)
    Population change, 1990-2000       1.1       1.1      1.1       1.1
    Net migration rate,
      1995-99 (4)                       .9        .0      1.9       1.2
    Poverty rate, 2002                  15      15.2     14.7      14.9
    Unemployment rate, 2001            6.0       7.1      6.3       6.6

Modified CDFI Fund criteria (6)
  Combined
    Tract (average
        characteristics)
      Number added                     939     2,599    2,235     4,834
      Share of families with
          income
        Less than or equal to
          poverty level               13.3      10.1     13.5      11.7
        Less than 50 percent of
          MSA or non-MSA              22.0      19.8     21.8      20.7
          median (1)
        Between 50 percent and 80
          percent of MSA or non-
          MSA median (1)              19.2      18.6     18.8      18.7
      Median family income
          (dollars)                 36,788    46,204   36,699    41,810
        Median relative to MSA or
          non-MSA                     94.3      98.2     95.7      97.1
      Housing
        Median house value
          (dollars)                 62,328   117,679   64,686    93,151
        Median house age (years)      38.1      35.0     37.4      36.1
        Occupancy by owner            60.8      58.8     62.9      60.7
        Vacancy rate (2)              18.3       7.3     16.5      11.6
      Population
        Over age 65                   15.9      13.4     15.2      14.3
        Minority (3)                  19.6      40.6     23.1      32.5
    County (average
        characteristics)
      Population change,
        1990-2000                      1.0       1.1      1.1       1.1
      Net migration rate,
        1995-99 (4)                   -2.1      -3.4     -0.3      -2.0
      Poverty rate, 2002              17.0      17.0     17.3      17.2
      Unemployment rate, 2001          7.3       7.9      8.0       8.0

  Unemployment
    Tract (average
        characteristics)
      Number added                     412     1,155    1,260     2,415
      Share of families with
          income
        Less than or equal to
          poverty level               13.5      12.5     12.8      12.6
        Less than 50 percent of
          MSA or non-MSA              22.3      20.9     21.5      21.2
          median (1)
        Between 50 percent and 80
          percent of MSA or non-
          MSA median (1)              19.4      18.1     18.9      18.6
      Median family income
          (dollars)                 37,341    43,901   37,506    40,564
        Median relative to MSA or
          non-MSA                     93.6      98.0     95.7      96.8
      Housing
        Median house value
          (dollars)                 72,214   136,980   71,581   102,828
        Median house age (years)      38.3      33.9     37.8      35.9
        Occupancy by owner            60.3      55.8     64.0      60.1
        Vacancy rate (2)              20.5       8.0     16.5      12.4
      Population
        Over age 65                   15.5      12.4     15.2      13.9
        Minority (3)                  22.0      44.9     21.6      32.7
    County (average
        characteristics)
      Population change,
        1990-2000                      1.1       1.2      1.1       1.1
      Net migration rate,
        1995-99 (4)                     .8       -.1      1.7        .8
      Poverty rate, 2002              17.2      19.1     16.6      17.8
      Unemployment rate, 2001         10.2      10.3     10.0      10.2

Poverty
  Tract (average characteristics)
    Number added                       368       652      853     1,505
    Share of families with income
      Less than or equal to
        poverty level                 18.1      15.8     17.7      16.9
      Less than 50 percent of MSA
        or non-MSA median (1)         24.9      23.0     24.4      23.8
      Between 50 percent and 80
        percent of MSA or non-MSA
        median (1)                    18.4      17.3     17.9      17.7
    Median family income
        (dollars)                   32,122    41,262   33,087    36,629
      Median relative to MSA or
        non-MSA                       91.8      97.1     93.7      95.1
    Housing
      Median house value
        (dollars)                   51,471   147,360   54,172    94,491
      Median house age (years)        36.5      33.2     36.2      34.9
      Occupancy by owner              63.9      49.0     63.8      57.4
      Vacancy rate (2)                15.5       7.7     15.8      12.3
    Population
      Over age 65                     14.7      12.5     14.5      13.6
      Minority (3)                    28.7      60.6     33.8      45.4
  County (average
      characteristics)
    Population change, 1990-2000       1.0       1.1      1.1       1.1
    Net migration rate,
      1995-99 (4)                      -.6      -2.9       .1      -1.2
    Poverty rate, 2002                23.2      24.6     22.7      23.5
    Unemployment rate, 2001            7.7       9.5      7.8       8.5

Population loss
  Tract (average characteristics)
    Number added                       364     1,737      558     2,295
    Share of families with income
      Less than or equal to
        poverty level                 11.1       8.5     12.1       9.4
      Less than 50 percent of MSA
        or non-MSA median (1)         20.2      19.4     20.5      19.7
      Between 50 percent and 80
        percent of MSA or non-MSA
        median (1)                    19.3      18.7     18.9      18.7
    Median family income
        (dollars)                   39,179    48,561   38,252    46,055
      Median relative to MSA or
        non-MSA                       96.5      98.3     97.3      98.1
    Housing
      Median house value
        (dollars)                   59,298   125,444   58,046   109,067
      Median house age (years)        38.2      36.0     37.3      36.3
      Occupancy by owner              58.9      57.3     58.8      57.6
      Vacancy rate (2)                17.2       6.3     17.0       8.9
    Population
      Over age 65                     16.4      14.3     15.7      14.6
      Minority (3)                    17.6      39.5     21.2      35.1
  County (average
      characteristics)
    Population change, 1990-2000        .9       1.0      0.9       1.0
    Net migration rate,
      1995-99 (4)                     -7.6      -6.6     -7.4      -6.8
    Poverty rate, 2002                14.5      16.6     15.4      16.3
    Unemployment rate, 2001            5.1       6.6      5.5       6.3

NOTE. Data exclude tracts in U.S.-affiliated areas and tracts without
income information.

(1.) See table 14, note 1.

(2.) Vacant housing units as a percentage of total housing units.

(3.) Non-whites or people of Hispanic origin.

(4.) Difference between net migration in 1999 and net migration in 1995
as a percentage of the population in 1997.

(5.) See table 14, note 2.

(6.) For description of modification to CDFI Fund criteria, see text.

CDFI Fund Community Development Financial Institutions Fund.

16. Number and share of rural banking institutions whose number
of CRA-eligible census tracts in their assessment areas would
increase under options for defining census tracts as CRA-eligible,
as of December 31, 2003

                                       Income-based options

                                      Percent of
                                       non-MSA
                                       median

                                                     Less than
                                     Less    Less    80 percent
                                     than    than     of state
               Item                   90      100      median

Large institutions
Currently with no CRA-eligible
    tracts
  Number                               300     300        300
  Rural institutions with increase
      in CRA-eligible tracts
    Number                             155     231        200
    Percent                           51.7    77.0       66.7
  Average increase (percentage
    points)                           21.2    41.9       32.4

Currently with some CRA-eligible
    tracts
  Number                               695     695        695
  Rural institutions with increase
      in CRA-eligible tracts
    Number                             629     679        650
    Percent                           90.5    97.7       93.5
  Average increase (percentage
    points)                           16.4    37.2       27.0

Assessment areas currently with no
    CRA-eligible tracts
  Number                             1,268   1,268      1,268
  Rural assessment areas with
      increase in CRA-eligible
      tracts
    Number                             615     976        788
    Percent                           48.5    77.0       62.1
  Average increase (percentage
    points)                           31.2    51.7       43.7

Assessment areas currently with
    some CRA-eligible tracts
  Number                             1,629   1,629      1,629
  Rural assessment areas with
      increase in CRA-eligible
      tracts
    Number                           1,312   1,557      1,406
    Percent                           80.5    95.6       86.3
  Average increase (percentage
    points)                           18.5    36.7       28.3

Small institutions (1)
Currently with no CRA-eligible
    tracts
  Number                             2,141   2,141      2,141
  Rural institutions with increase
      in CRA-eligible tracts
    Number                           1,152   1,870      1,476
    Percent                           53.8    87.3       68.9
  Average increase (percentage
    points)                           32.7    57.4       47.9

Currently with some CRA-eligible
    tracts
  Number                             1,831   1,831      1,831
  Rural institutions with increase
      in CRA-eligible tracts
    Number                           1,409   1,731      1,551
    Percent                           77.0    94.5       84.7
  Average increase (percentage
    points)                           23.2    41.4       34.1

                                      Modified CDFI Fund criteria

                                                    Individual

                                                                Popu-
                                     Com-     Unem-      Po-    lation
               Item                  bined   ployment   verty    loss

Large institutions
Currently with no CRA-eligible
    tracts
  Number                               300       300      300      300
  Rural institutions with increase
      in CRA-eligible tracts
    Number                              49        27       11       18
    Percent                           16.3       9.0      3.7      6.0
  Average increase (percentage
    points)                           54.4      62.6     42.9     40.7

Currently with some CRA-eligible
    tracts
  Number                               695       695      695      695
  Rural institutions with increase
      in CRA-eligible tracts
    Number                             369       243      183      152
    Percent                           53.1      35.0     26.3     21.9
  Average increase (percentage
    points)                           27.5      23.7     20.0     14.8

Assessment areas currently with no
    CRA-eligible tracts
  Number                             1,268     1,268    1,268    1,268
  Rural assessment areas with
      increase in CRA-eligible
      tracts
    Number                             194       110       37       66
    Percent                           15.3       8.7      2.9      5.2
  Average increase (percentage
    points)                           77.1      77.0     82.0     72.4

Assessment areas currently with
    some CRA-eligible tracts
  Number                             1,629     1,629    1,629    1,629
  Rural assessment areas with
      increase in CRA-eligible
      tracts
    Number                             658       412      303      192
    Percent                           40.4      25.3     18.6     11.8
  Average increase (percentage
    points)                           40.3      37.7     35.0     26.1

Small institutions (1)
Currently with no CRA-eligible
    tracts
  Number                             2,141     2,141    2,141    2,141
  Rural institutions with increase
      in CRA-eligible tracts
    Number                             389       161       75      201
    Percent                           18.2       7.5      3.5      9.4
  Average increase (percentage
    points)                           83.8      85.7     88.4     80.7

Currently with some CRA-eligible
    tracts
  Number                             1,831     1,831    1,831    1,831
  Rural institutions with increase
      in CRA-eligible tracts
    Number                             681       306      421      192
    Percent                           37.2      16.7     23.0     10.5
  Average increase (percentage
    points)                           47.6      42.8     45.6     40.0

NOTE. See general note to table 9. A rural banking institution is an
institution whose assessment area contains at least one rural census
tract. For definition of relative tract income, see table 8, note 1.
For description of CDFI Fund criteria, see text discussion of table
14. For definition of large and small institutions, see table 1,
note 1. For description of assessment areas, see table 9, note 1.

(1.) Rural assessment areas were approximated by the rural counties
in which small institutions had branches. These approximations were
used to determine whether any of the census tracts served by small
institutions would become CRA-eligible.

CDFI Fund Community Development Financial Institutions Fund.

(1.) The agencies are the Board of Governors of the Federal Reserve System (Board), the Federal Deposit Insurance Corporation (FDIC), the Office of the Comptroller of the Currency (OCC), and the Office of Thrift Supervision (OTS).

(2.) For a more expansive overview of the history of the CRA, see Griffith L. Garwood and Dolores S. Smith (1993), "The Community Reinvestment Act: Evolution and Current Issues," Federal Reserve Bulletin, vol. 79 (April), pp. 251-67.

(3.) Office of the Comptroller of the Currency, Board of Governors of the Federal Reserve System, Federal Deposit Insurance Corporation, Office of Thrift Supervision (1995), "Community Reinvestment Act Regulations," Federal Register, vol. 60 (May 4), pp. 22156, 22178.

(4.) CRA ratings, the type of evaluation (for example, small-institution or large-institution), the date of the evaluation, and the name of the agency that conducted the evaluation are available from the Federal Financial Institutions Examination Council (FFIEC) at www.ffiec.gov. Comprehensive written evaluations, including "subratings," are available through links from the FFIEC's website to the websites of the supervisory agencies, which post the evaluations as PDF files. The sub-ratings are available in written form only; they are unavailable in a quantitative, easy-to-use format that would facilitate analysis.

(5.) Institutions that are large under the CRA and are covered by HMDA must report the census tracts of all properties for which loans have been extended or for which loan applications have been received unless the loan is made or the application is received in a county with a population of 30,000 or less, in which case reporting the census tract is optional. Small institutions covered by HMDA may, but need not, report the property locations (census tracts and counties) for their rural loans.

(6.) Some tract classifications adjust more frequently than once a decade because of changes in the boundaries of metropolitan areas.

(7.) OCC, Board, FDIC, OTS (1995), "'Community Reinvestment Act Regulations," pp. 22156, 22178.

(8.) Office of the Comptroller of the Currency, Board of Governors of the Federal Reserve System, Federal Deposit Insurance Corporation, Office of Thrift Supervision (2001), "Community Reinvestment Act Regulations," advance notice of proposed rulemaking, Federal Register, vol. 66 (July 19), p. 37602.

(9.) Office of the Comptroller of the Currency, Board of Governors of the Federal Reserve System, Federal Deposit Insurance Corporation, Office of Thrift Supervision (2004), "Community Reinvestment Act Regulations," Federal Register, vol. 69 (Feb. 6), p. 5729.

(10.) To be considered large, an institution must fail to meet the criteria for a small institution as of December 31 of both of the previous two calendar years.

(11.) The OTS implemented its increase in a final rule published on August 18, 2004. See Office of Thrift Supervision (2004), "Community Reinvestment Act Regulations," Federal Register, vol. 69 (Aug. 18), p. 51155.

(12.) See Board of Governors of the Federal Reserve System (2004), press release, July 16, www.federalreserve.gov/boarddocs/press/all/ 2004.

(13.) Federal Deposit Insurance Corporation (2004), "Community Reinvestment Act Regulations," Federal Register, vol. 69 (Aug. 20), p. 51611.

(14.) Office of Thrift Supervision (2004), "Community Reinvestment Act--Community Development, Assigned Ratings," Federal Register, vol. 69 (Nov. 24), p. 68257.

(15.) The proposal would leave unchanged the criteria for evaluating small institutions (those with less than $250 million in assets); as noted earlier, these criteria concentrate on retail lending (see box "The Small-Institution Evaluation"). The proposal would also leave unchanged the criteria for evaluating large institutions (those with more than $1 billion in assets), which would continue to be subject to a three-part evaluation (see box "The Large-Institution Evaluation").

(16.) Under the proposal, intermediate small institutions would not be subject to the large-institution service test. The service test evaluates, among other things, the geographic distribution of an institution's branches and its record of opening and closing branches, as well as its record of providing community development services--that is, financial services targeted to lower-income people. Under the proposal, the branching of intermediate small institutions would no longer be evaluated although, under the proposed community development test, the community development services of such institutions would be.

(17.) For convenience, our research ignored the changes that the OTS made to its regulations and assumed that the OTS regulations are the 1995 regulations.

(18.) We used 2003 as a test year because at press time it was the latest year for which public data on retail lending activities related to the CRA were available.

(19.) Fourteen of the 241 institutions were not part of a multibank holding company. They had exceeded the asset-size threshold for the large-institution examination as of the beginning of 2003, but their assets had fallen below $250 million as of the end of the year. Under the 1995 regulations, these institutions had reverted to the small-institution examination as of the beginning of 2004.

(20.) Nearly 380 of the status-changing institutions, although covered by the large-institution examination as of December 31, 2003, had last been evaluated under the CRA as small banking institutions (data shown under the "small-institution" subcategory); consequently, they had not yet been evaluated as "large."

(21.) In classifying rural counties, the U.S. Department of Agriculture makes the distinction between exurban and remote, among others.

(22.) Local banking markets are not necessarily equivalent to CRA assessment areas. Unlike CRA assessment areas, local banking markets are not drawn from the perspective of a particular institution.

(23.) The figures use counties, which approximate, rather than precisely match, banking markets.

(24.) "Call Report" is the informal name for the Report of Condition and Income, which commercial banking institutions must file each quarter with federal and state banking agencies. It is essentially equivalent to the Thrift Financial Report, which savings institutions must file each quarter with the Office of Thrift Supervision.

(25.) Every institution in the analysis had at least one comparable institution on the other side of the threshold in the same state, of the same institution type, and in the same area type.

(26.) We refrained from conducting the comparison for any of the measures that use lower-income classifications because 2001 classifications were based on the 1990 census and 2003 classifications were based on the 2000 census. The change in classifications makes a comparison of lower-income activity in 2001 with lower-income activity in 2003 problematic.

(27.) Some empirical support exists for the substitutability explanation. We examined the CRA performance evaluation reports (PEs) for the twenty-three institutions in our sample that had assets between $250 million and $500 million and that received "outstanding" CRA ratings (column 1, row 1, of table 7). There is a mild negative correlation (-.2) between the dollar volume of community development lending and the investments reported in the PEs. However, some evidence also suggests that the substitutability explanation applies only to smaller institutions. An examination of the dollar volume of community development lending and the investments reported in the PEs of the fifty institutions in our sample that had assets between $500 million and $1 billion and that received "outstanding" CRA ratings (column 2, row 1) shows a significant positive correlation of .5.

(28.) We also conducted a similar analysis that restricted the comparisons to retail loans, retail loan dollars, offices, families, housing structures, and deposits in lower-income tracts. The results for this comparison are substantially the same as those for the comparison based on the full set of census tracts.

(29.) In two size classes (the largest and the smallest), remote areas received fewer multifamily loans as measured against the distribution of families than did other areas. However, in both cases, remote areas received more multifamily loans as measured against multifamily housing structures, arguably a better measure of comparison.

(30.) The calculations of retail loan dollars per deposit dollars tend to show higher lending to the urban part than do the calculations of retail loan numbers per population because retail loans in urban areas are generally larger than in rural areas, a reflection of higher property values.

(31.) Community Development Financial Institutions Fund, U.S. Department of the Treasury (2004), "Community Development Financial Institutions Program," Federal Register, vol. 69 (May 11), p. 26259. The fund's definition of an investment area contains an additional criterion, which states that the area has "significant unmet needs for loans, equity investments, or financial services." We disregarded this criterion because the fund refrained from defining it in objective, quantitative terms.

(32.) The two population criteria that we use in our adaptation of the fund's criteria are based on 2000 census data.

(33.) Institutions that filed 2003 HMDA and CRA small-business data used census tract definitions based on the 2000 census. Metropolitan area boundaries based on the 2000 census were not implemented for filings related to HMDA and the CRA until 2004. In constructing the numbers we report here, we use the 2004 definitions of metropolitan statistical areas.

(34.) Rhode Island is the only state in which the nonmetropolitan area median income is higher than the overall state median income.

(35.) See David A. McGranahan and Calvin L. Beale (2002), "Understanding Rural Population Loss," Rural America, vol.17 (Winter). The article is available on the website of the Economic Research Service, U.S. Department of Agriculture (www.ers.usda.gov). The authors found that the rural areas with population loss are distinct from those with high poverty.

(36.) Assessment areas of small institutions are approximated by the counties in which they have branches.

RELATED ARTICLE: The large-institution evaluation.

The regulations that implement the CRA establish three tests by which the performance of most large retail banking institutions is evaluated: a lending test, an investment test, and a service test.

The lending test measures lending activity for many types of loan, including home mortgage, small-business, and small-farm loans. The assessment criteria are the proportion of an institution's loans in its assessment areas, the distribution of lending across borrowers of different incomes, the distribution of lending across census tracts of different incomes, the extent of community development lending, and the use of innovative or flexible lending practices to address the credit needs of lower-income individuals or areas.

The investment test considers a banking institution's qualified investments that benefit its assessment area or a broader statewide or regional area that includes its assessment area. A qualified investment is a lawful investment, deposit, membership share, or grant that has community development as its primary purpose.

The service test considers the availability of an institution's system for delivering retail banking services and judges the extent of its community development services and their innovativeness and responsiveness. Among the assessment criteria for retail banking services are the geographic distribution of an institution's branches and the availability and effectiveness of alternative systems for delivering retail banking services, such as automated teller machines, in lower-income areas and to lower-income persons.

RELATED ARTICLE: The small-institution evaluation.

Small institutions are eligible for streamlined CRA evaluations and are exempt from CRA data reporting obligations. The performance of a small institution is measured by its efforts to help meet the credit needs of its assessment area. These efforts are evaluated according to the following criteria:

* the institution's overall ratio of loan dollars to deposits

* the percentage of loans or, as appropriate, other lending-related activities in the assessment area

* the institution's record of lending to borrowers of different income levels and to businesses and farms of different sizes

* the geographic distribution of the institution's loans

* the institution's record of responding to written complaints about its performance in helping to meet credit needs in assessment areas

Robert B. Avery, Glenn B. Canner, and Shannon C. Mok, of the Division of Research and Statistics, and Dan S. Sokolov, of the Division of Consumer and Community Affairs, prepared this article. Onka L. Tenkean provided research assistance.