Tim Critchfield, Tyler Davis, Lee Davison, Heather Gratton, George Hanc, and Katherine Samolyk^*

The U.S. banking system has long had a multitude of small institutions. This characteristic of the industry has been shaped by a number of factors. The dual banking system—that is, the coexistence (since the end of the Civil War) of both federal and state chartering—has fostered the creation of small banks, and this effect was reinforced by chartering regulations at both the national and state levels that were frequently permissive. In addition, the fear of concentration, as well as efforts to keep local markets free of outside competition, led many states to impose longstanding limits on branching, and this legacy of unit banking helped swell the numbers of small banks, particularly in the Midwest. The lack, until fairly recently, of the technology necessary for creating very large banking organizations was another factor contributing to the multiplicity of small banks. Of course, during the last quarter of the twentieth century the requisite technological advances occurred at the same time that legal impediments to branching were being gradually removed. Thus, for the last decade of the century in particular, the industry saw a great deal of consolidation, much of it involving community banks, whose numbers fell significantly.¹ (See table 1.) Moreover, community banks' shares of deposits, assets, and offices have fallen steadily and significantly since 1985. (See table 2.) Given these trends and the oft-cited notion that such small banks are destined to disappear, victims of their inability to compete with larger institutions, one might ask why the future of community banks is of interest.

Table 1 - Number of FDIC-Insured Community Banks, 1985-2003

Table 2 - Shares of Banking Industry Assets, Deposits, and Offices, 1985-2003

One reason is that although the number of community banks (those with less than $1 billion in assets, a definition explained on the next page) has decreased, thousands of such banks remain: at year-end 2003 community banks constituted 94 percent of all banks in the nation. Thus, by this criterion, what happens to these banks is not insignificant. Another reason is that from an economic viewpoint, these institutions remain very important in specific business and economic sectors, notably small-business and agricultural lending. Small businesses play a critical role in the U.S. economy as a whole and in economic growth in particular, so their ability to find credit—and where they find it—is of consequence. Some observers have expressed concern that a continued banking industry consolidation that significantly diminished the number of community banks serving small-business and agricultural lending could leave the credit needs of such businesses unmet (although evidence as to the validity of this concern is mixed).

The future of community banks is worth examining from a third viewpoint as well—that of deposit insurance. Community banks' prospects are of significant interest to the FDIC because small-bank failures have represented a disproportionate share of FDIC losses in recent years; between 1998 and 2002, for example, community banks with 63 percent of failed-bank deposits accounted for approximately 72 percent of the FDIC's failure costs.² Many of these failed small banks experienced at least some period of very high growth within five years before failure, and some of the failed community banks, whether through new ownership or a change in business plan, had adopted rapid-growth, high-risk policies, which resulted in high resolution costs when the institutions failed. Such a rapid transformation of a bank's risk profile is rarer in the case of a large bank.

A community bank can be defined in different ways, but size is usually the determining factor. These banks are generally thought of as relatively small institutions that do most of their business within a fairly circumscribed geographic area. For the purposes of this article, community banks are defined as banking organizations (bank and thrift holding companies, independent banks, and independent thrifts) with aggregate bank or thrift assets of less than $1 billion; in addition, bank asset-sizes are calculated using assets measured in 2002 dollars.³ Some studies may not include thrifts, but if thrifts and banks can be viewed as competitors, it is logical to include both kinds of financial organizations. However, it must be noted that some analyses in this article, particularly those examining earnings and performance, require the exclusion of de novo banks (defined here as banks less than five years old) because during the early years of a bank's existence, earnings and growth are atypical. In addition, because of historical differences between banks and savings institutions, certain analyses of performance and balance sheets treat commercial banks separately from savings institutions.

This article first explores some of the more significant characteristics of community banking, examining the importance of community banks in small-business lending in terms of their ability to handle "soft" data, their tendency to rely on retail deposits for funding, and their emphasis on personal service. The tremendous consolidation that community banks have experienced has already been mentioned, and the second section of the paper investigates the decline in community bank numbers from 1985 to 2003, analyzing both the nature (failures, mergers, and new banks) and the geographic distribution of the decline. Was consolidation more pronounced in formerly unit-bank states than in other areas? How did consolidation differ between large metropolitan, small metropolitan, and rural areas and between growing and declining markets? This section also examines changes in the presence and the importance of community banking in different types of local deposit markets. Having examined changes in community bank presence, we turn our focus to these banks' balance sheets, business lines, and performance. Where has community banks' share as lenders suffered, and where have these banks held their own? Have the characteristics of community bank funding changed? How have community banks performed, both compared with larger banks and within their own ranks? How has community bank performance been affected by growth in the markets in which community banks are present? The article ends with some discussion of the prospects for community banks in light of their competitive strengths and the challenges facing them.

The Economic Role of Community Banks

Although the number of community banks has declined over the past 20 years, the performance of these banks and the fact that their numbers remain high confound predictions of their virtual demise—predictions made just a few years ago. For example, in 1997 one bank analyst predicted that the industry would consolidate at a rate of 300 banks per quarter, with a total of less than 1,000 banks remaining. A 1996 prediction held that consolidation would mean the United States would have "well under 5,000" banks just four years later; much of this decline would obviously have involved community banks.⁴ Such prognostications are, of course, often inaccurate. It should be noted that this view was not universally shared. As early as 1991 former FDIC Chairman William Isaac believed that consolidation did not pose a danger to well-run community banks; in 1996 Alan Greenspan was quoted as stating that those who were predicting the end of the community bank were "just plain wrong"; and by 1997, others were predicting (rightly) that the decline in small-bank numbers was slowing dramatically.⁵

Since community banks have not vanished, it appears that many of them must be doing something right; moreover, the formation of significant numbers of new community banks since 1992 (to be discussed in greater detail below) demonstrates that these banks are perceived to be viable. Researchers have therefore sought to determine just what the "something right" is and whether it will continue to be important. That "something" is strongly related to community banks' economic role, and three areas of that role will be discussed here: community banks' success in providing credit to certain business sectors, their ability to attract retail deposits, and their capacity to build on the provision of personal services to their customers.

One of the more significant elements of community banks' economic role is their function as providers of credit: they serve important segments of the business-loan and farm-loan markets. Although overall their share of small-business loans (loans of less than $1 million at origination) has declined during the past decade, they still provide almost a third of all small commercial and industrial loans and more than 40 percent of small commercial real estate loans. They are even more important as farm lenders, providing 65 percent of all farm real estate loans, 61 percent of all farm operating loans, and roughly 75 percent of small farm loans (loans of less than $500,000 at origination) reported on bank balance sheets. A detailed examination of community bank lending is presented below in the section "Community Bank Industry Shares, Portfolios, and Performance."

Much recent literature has identified the strength of community banks in these areas as stemming from their ability either to successfully lend to what have been variously described as "informationally opaque" borrowers—borrowers without long credit histories suitable for credit-scoring or other model-based lending practiced by large banks—or to engage in relation- or reputation-based lending or lending in low-volume markets. As a recent article notes, "large hierarchical firms are at a comparative disadvantage when information about individual investment projects is innately soft."⁶ Soft data include a borrower's character or ability to manage, and this information is generally gleaned through a local presence and personal interactions with borrowers; also thought to be helpful is a favorable organizational structure (close proximity of lending officers to management).⁷ In contrast, large banks prefer hard data (e.g., credit history, income, debts, and other data available from financial statements and credit reports) and are less willing to lend to "informationally difficult credits."⁸ With the ability to process the soft data, community banks are thought to have certain comparative advantages in lending to informationally opaque borrowers, and these advantages are helpful in underwriting and monitoring loans to small businesses and farmers. Empirical support for this view is provided by a recent study that found that small banks earn higher risk-adjusted returns on business loans than large banks; the study concluded that small banks make "better choices" in lending to businesses.⁹

Community banks have also been defined by their tendency to rely more on retail and insured deposits for their funding than large banks have done. A recent study notes that at year-end 2002, community banks "held 24 percent of deposits [as a percentage of deposits at all banks] in accounts of $100,000 or less, but only 15 percent of deposits in accounts over that amount."¹⁰ Given this emphasis, it is not surprising that community banks usually charge lower fees for deposit services.¹¹ In 2002, the Federal Reserve Board found that, on average, small institutions charged lower fees than large banks. For example, the average annual fees charged by large banks for simple passbook accounts were 72 percent higher than those charged by the smallest banks, and the average stop-payment fee was 38 percent higher at large banks than at the smallest banks.¹² It should be noted, however, that the fee advantage held by smaller institutions, though still present, has been declining; the decline may indicate that small banks are seeking to exploit fee income somewhat more than they have in the past.¹³ Community banks, because they rely on retail deposits and need to attract them, also appear to pay higher rates on retail deposits than large banks competing in multiple markets.¹⁴ Paying the higher rates has been feasible because surviving small banks have been able (until very recently) to earn a higher rate of return on their assets, maintaining profitability even while growing more rapidly than large banks during nearly the past two decades.¹⁵

A third significant element in community banks' economic role is the manner in which they interact with customers. Although advances in information technology, such as the Internet, have enabled many customers to transact banking business without having recourse to a bank's premises, there apparently remain customers who prefer face-to-face contact. Community banks have typically seen personal service as their most important competitive advantage, and they market personal service and local connections to prospective customers. Many community banks seek to demonstrate this service by being active in their communities. For example, a significant percentage of community bankers responding to a recent survey noted that they participated in civic groups, worked with local chambers of commerce, supported local schools, assisted local relief efforts, and offered special help to low-income segments of the community.¹⁶ Recent research has shown that the formation of new banks is strongly correlated with mergers that shift "ownership away from small organizations or toward distant organizations"; one explanation for this correlation is that large organizations tend not to adequately serve "small, relationship-based" customers. The new institutions may be finding a market in providing for the needs of customers to whom the business methods of larger banks are unsatisfactory.¹⁷

Anecdotal evidence supports the view that small banks can attract such customers. In a recent Federal Reserve System survey of community bankers, respondents commonly noted that because of their local knowledge and personal service, they were able to draw business away from larger institutions. They also reported that some community banks experienced significant asset growth in the wake of recent acquisitions of other community banks by large institutions.¹⁸ Another indication of the "personal-service" phenomenon is large banks' efforts to emphasize personal service even though their comparative advantage would seem to be in mass-market lending based on hard data (credit history and other objective indicators of risk).¹⁹ Whether face-to-face contact will continue to be as important is a subject dealt with below.

Consolidation and the Geography of Community Banking

There is some concern that the economic role played by community banks has diminished. Their presence has clearly declined as the banking industry has been transformed into one composed of fewer, larger institutions. Changes in community bank presence can be measured in a number of ways. Two approaches are used here. One is to examine the components of change (mergers, failures, and new banks) between 1985 and 2003 in different types of markets (rural, small metropolitan, and large metropolitan [and, within the last, urban and suburban] as well as in markets experiencing population growth and population decline) and under different past restrictions on branching. The other approach used here is to analyze changes in community banks' shares of deposits and deposit-taking offices across the same types of markets also between 1985 and 2003. Both approaches allow us to see if there were kinds of markets or states where, in the face of consolidation and competition, community banks fared better or worse than they did in other markets or states.

Changes in the Number of Community Banks
Between 1985 and 2003, the total number of community banks declined by just under half (table 1). The greatest decrease was among small community banks (those with assets below $100 million in 2002 dollars), but it should be noted that a significant portion of the overall decline came about because small institutions outgrew the community bank size class. The number of community banks having inflation-adjusted assets of less than $100 million declined by 64 percent.²⁰ These small banks accounted for 92 percent of the decline in the total number of community banks. The decline in the number of larger community banks (those having assets of between $100 million and $1 billion in 2002 dollars) was much smaller—this group experienced only a 13 percent drop in number.

Before exploring the consolidation that led to the decrease in community bank numbers, we examine the positive side of the ledger—the formation of new banks—because trends in their establishment have implications for the future of community banks. New-bank formation fell into three periods: the first, from 1985 to 1990, corresponded with a relatively permissive chartering environment and saw considerable numbers of new banks formed (though formations dwindled as the period drew to an end);²¹ the second period, from 1991 to 1995—from the last part of banking crisis through the beginning of the industry's recovery—had few new banks; and the third period, from 1996 to 2003—as the industry thrived and consolidation created new opportunities-once again saw significant numbers of new banks. (See table 1.)

The substantial number of new banks confirms that many investors believe the community bank model remains viable, at least where local economies are growing. Since 1992 there have been approximately 1,250 new community banks, of which about 150 have been merged and about 1,100 still exist as independent organizations.²² This market test is impressive testimony on behalf of viability, even though some of these de novos developed substantial risk factors as they matured. Young banks, because they have tended to locate in rapidly growing markets and because they have concentrated more heavily on real estate lending, are substantially more vulnerable to serious real estate problems than their established counterparts.²³

In the 1980s new institutions did not fare well, but institutions formed in the 1990s can be expected to do better. First, newly chartered banks now face more stringent supervision.²⁴ Second, in 1991 the FDIC obtained separate statutory authority to approve deposit insurance for national banks;²⁵ previously approval had been automatic. Third, new banks in the 1990s might have been able to tap more experienced management than new banks in the 1980s because in the 1990s many de novo banks were formed in the same geographic areas where there had been merger activity. Thus, the supply of locally available bank management personnel would have increased.²⁶ Fourth and most important, serious regional recessions comparable to those of the 1980s have been absent.²⁷ Only 4 of the approximately 1,250 new community banks established between 1992 and 2003 have failed.

Although new-bank formation has been significant, the effect of consolidation on the community bank population far outweighs it. There were essentially two components to the decline: mergers and failures. Throughout the entire period 1985–2003, mergers accounted for most of the decrease in the number of community banks; failures were significant as well, but (not surprisingly) were almost completely confined to the years of industry problems—the years before 1993.²⁸ This consolidation, coupled with the lifting of structural restrictions and the frequent expression of belief that community banks were doomed to vanish, might have seemed to augur great geographical disparities in the landscape of community banking (even aside from the sheer decrease in numbers).

As we began our investigation, it seemed likely that further examination of the decrease in community bank numbers might help explain why the number of community banks had been halved. We explored two logical explanations for the decline: (1) that the removal of interstate branching restrictions likely played a significant role, and (2) that community banks located in markets having differing economic prospects would have experienced differential declines in numbers. (We analyzed four market segments: rural markets, small metropolitan markets, and suburban and urban parts of large metropolitan markets.) Both analyses, however, yielded surprising results and demonstrated that neither of these explanations was persuasive; the declines were, in fact, proportionally similar no matter how the pie was sliced. (See figure 1.)

D

To examine the hypothesis that, with the removal of branching restrictions, formerly unit-bank states would have witnessed a disproportionate decline in community bank numbers, we compared 12 such states with the rest of the country.²⁹ We found that community bank numbers declined by 53 percent in the unit-bank states and by 46 percent in the non-unit bank states. The unit-bank states contained 42 percent of community banks in 1985 but still had 39 percent of them in 2003. This decline in share stemmed largely from less new-bank activity and proportionally more failures in the unit-bank states.³⁰ These relatively small differences fail to suggest that unit-banking laws had artificially maintained high numbers of community banks, and it is hard to argue, at least from experience, that by virtue of their previous banking statutes these states will see greater consolidation in the future. However, since many of the unit-bank states are predominantly rural and since banks in rural areas have been comparatively less attractive as merger targets, it may be that not enough time has passed for consolidation to occur.

An examination of community bank presence in different types of markets yielded a similar picture-one of proportional stability in the community bank population.³¹ There were significant declines of 46 to 52 percent in all types of markets (figure 1), but the differences in the percentage changes of community banks in each type of market between 1985 and 2003 were small. However, there have been significant differences in the dynamics underlying the declines in the number of community banks. Rural areas saw proportionally fewer mergers and very little de novo entry in comparison with both small metro and large metro areas, and the largest amount of merger and de novo activity took place within large metro areas.³² When we extended our analysis to community banks in both growing and declining markets, we saw similar patterns. Predictably, the overall drop in the number of community banks was less in growing markets than in declining markets. Also predictably, particularly in all three types of metropolitan markets, mergers and new banks were far more numerous in areas of growing population than in areas of declining population.³³

The other article in this issue of the Banking Review notes that many banks are located in rural areas with declining populations, and that long-standing trends in farm depopulation and consolidation have led to economic decline in many of these areas-most notably in the Great Plains states. Despite depopulation and its attendant economic effects, however, reductions in the number of banks even in areas experiencing the most profound depopulation mirrored the reductions in rural areas across the country. The long-term effects of depopulation, coupled with a lack of succession plans at closely held community banks, may eventually lead to problems with the survival of community banks in those states.³⁴ Thus far, however, these banks have not performed badly, and predicting with confidence how quickly consolidation will occur in these areas as a result of such long-term processes—and how much of it there will be—is difficult.

It is, however, worth noting that community banks in the Great Plains represent only about 13 percent of all U.S. community banks. And not all rural areas are declining; some are growing because of high birth rates and high immigration. During the 1990s, the rural West grew by 20 percent—twice the national average. The overall population in the 343 rural counties in the Western Census Region increased by about 27 percent from 1985 to 2001; only just over a quarter of those counties experienced population decline, whereas the remainder saw their populations expand.³⁵ Furthermore, many rural areas are no longer dominated by agriculture. Indeed, the U.S. Department of Agriculture finds that in seven out of eight rural counties the economy is now dominated by manufacturing, services, and other employment not related to farming. Even within agricultural areas, future job growth is more likely to come from industries related to farming than from farming itself.³⁶ The performance of, and prospects for, banks located in rural areas that are not experiencing depopulation are likely to mirror those of similar-size banks in urban areas.

Community Bank Presence in Local Deposit Markets
Although the distribution of community banks across different types of markets has remained remarkably stable, the distribution of community bank deposits across local banking markets has shown more variation. In this section we look at change in the deposit-taking presence of different-size institutions in various types of local banking markets and at the implications of these changes for where community banks tend to operate. We find that changes in the composition of local deposit markets reflect the increasing geographic reach of larger (noncommunity) banks into new markets—a spread made possible by branching deregulation and changing banking technologies that have reduced the costs associated with distance. Changes in local deposit markets also reflect a consolidation of community banks themselves into fewer, larger institutions. But, as with trends in the number of community banks, deposit market patterns suggest that community banks have not been left to wither in areas with declining economic prospects. Rather, community banks continue to play an important role, albeit a smaller one than before, in all types of local banking markets.

Community banks' share of the deposits held in all types of local markets certainly declined between 1985 and 2003. The largest decline was in the urban parts of large metro markets, where the community bank share was halved; in other types of markets, decreases in community bank market shares were proportionally smaller (see table 3). Moreover, these changes in community bank market shares understate the extent to which surviving community banks have actually maintained their competitive position in the face of local consolidation activity. In other words, given the shift that has taken place toward fewer, larger banks, one would have expected to see even greater declines in community bank market shares than have actually occurred. Indeed, when we net out changes in deposit share that are due to the reclassification of banks into larger-size categories (because of subsequent mergers, acquisitions, or asset growth), we find that the remaining community banks have been increasing their deposit shares; this is particularly true in small MSAs and suburban areas of large MSAs.

Table 3 - Changes in the Distribution of Domestic Deposits by Type of Geographic Area, 1985-2003

It is also instructive to look at changes in the extent to which the different size categories of banks have any deposit-taking presence (i.e., the extent to which they report any deposit-taking offices) in local banking markets. The 25 largest banking organizations were those best positioned to expand their geographic reach. In 1985 they reported having offices in roughly half of large urban areas, in 40 percent of small MSAs, but in only 11 percent of rural counties. Midsize banks had deposit-taking offices in practically all metropolitan markets but in fewer than half of all rural counties. Hence, as recently as the mid-1980s, a significant number of rural banking markets were served entirely by community banks. By mid-2003, the 25 largest banks had increased their deposit-taking reach to more than 45 percent of rural markets and almost all urban markets. Of course, the widening reach of the very largest banks is not surprising, for they tended to be the banks most constrained by the branching restrictions that were lifted during the period.

Like measures of deposit market share, the relationship between bank consolidation and the geographic scope of banking offices yields information about the nature of consolidation activity at the local-market level. Between 1985 and 2003, the number of rural markets where community banks reported having any deposit-taking offices declined; but given the consolidation that has taken place, one would have expected this decrease to have been much more pronounced. Conversely, the very largest banks have increased the number of rural markets where they have deposit-taking branches; but in many of the rural markets where they acquired a branching presence they have not maintained it. Rather, the data suggest that other community banks entered markets where a community bank presence had been lost because of merger activity.

These patterns in local deposit markets indicate that changes observed in community bank presence understate the extent to which surviving community banks are actually prospering. Adjusting for reclassifications in size category due to acquisition activity or asset growth, we find that despite experiencing market share declines, community banks—here measured in terms of their local deposit taking—were actually growing. In other words, activity by existing (and new) community banks has offset what would have been larger declines in market share due to bank consolidation.

Other studies have found similar patterns for community bank assets, deposits, and small-business lending. A study of the performance of smaller community banks shows that, after adjustments for mergers, the growth of assets has been "significantly faster" at small banks than at large banks in every year from 1985 to 2000.³⁷ Deposit growth—both total deposits and uninsured deposits—followed the same pattern. Along the same lines, a study of small-business lending by community banks found that, adjusting "for size category reclassifications due to consolidation or asset growth and for local market conditions," community bank small-business-loan market shares increased from 1994 to 2000.³⁸ Together these findings indicate that the relative growth of surviving (and new) community banks (measured in terms of assets, deposits, and small-business lending) has been such that one would have underestimated community banking's continuing presence by looking only at the pace of merger activity.

Changes in the presence of larger (noncommunity) banks in local deposit markets have affected where community banks tend to operate. Not surprisingly given the spreading reach of the top 25 banks, the concentration of the largest banks' deposits in large urban centers has declined as they have diversified into smaller markets. Perhaps also not surprisingly, the concentration of community banks' deposits in rural markets has risen, a trend suggesting that community banks' comparative advantage has shifted even more toward serving small, less densely populated markets. However, the declines in deposits held in metro areas are smaller than the declines implied by reclassifications due to acquisition activity or community bank growth during the period; indeed, the share of community banks' deposits located in suburban markets actually increased between 1985 and 2003. This increase is consistent with the notion that there is a niche for service-oriented community banks in suburban markets.

To understand the geographic deposit patterns in relation to longer-term economic prospects in local markets, we conducted a parallel analysis of deposit trends for growing versus declining markets, defined in terms of positive and negative population growth.³⁹ Community banks do not appear to have been relegated to providing services in markets where the economic base is dwindling. Community banks have seen their deposit market shares decline in all types of markets, but those declines are no more pronounced in growing markets than in declining ones.⁴⁰

Community Bank Industry Shares, Portfolios, and Performance

As noted above, the many observers who argue that the community banking segment of the industry remains viable often base their claims on the importance of community banks in certain types of loan markets—specifically, in lending to small businesses and farms. A significant amount of research holds that community banks' strength as lenders stems from their ability to form the relationships necessary to lend to informationally opaque borrowers (an advantage widely viewed as important in small-business and small-farm lending), and studies have documented the importance of smaller banks in such lending.

This section describes the evolving role of community banks in the banking industry. It examines the ways in which community banks as intermediaries are different from larger banks in terms of their industry shares, portfolio composition, and performance. The analysis of community banks' performance includes a comparison between community banks that remained community banks and surviving community banks that outgrew the community bank size classification or were acquired out of it. We also relate the performance of community banks to the longer-term growth of the local markets where they were located.

Unless otherwise indicated, all analyses of "banks" and "banking" include both commercial banks and savings institutions. Here we do, however, also pre-sent some trends for commercial banks and savings institutions separately to highlight relevant differences between these two types of institution. Despite their increasing similarity, these segments of the industry have evolved from very different places and continue to exhibit differences (particularly on the asset side of their balance sheets) that are important to consider when one is assessing community banking's prospects.

Industry Shares of Assets and Liabilities
Between 1985 and 2003 community banks' share of total banking industry assets declined by nearly half, from 27 percent to less than 14 percent. (See table 4.) This overall decline reflects large relative declines in the shares of consumer credit and home mortgages funded by community banks. But despite having lost out in some credit markets, smaller banks appear to be holding their own in others—notably real estate lending to businesses and farms. Although community banks control less than 14 percent of banking-sector assets, they fund almost 29 percent of the industry's commercial real estate lending and more than 65 percent of farm real estate loans. And in terms of small commercial and small farm loans, community banks are even more important: as of mid-2003, community banks held 37 percent of small loans to businesses (real estate and commercial & industrial loans) reported by banks and almost three quarters of outstanding small farm loans (real estate and operating loans).

Table 4 - Share of Banking-Sector Assets and Funding, 1985, 1994, and 2003

Trends in the shares of industry assets held by community banks are consistent with the view that larger banks have a growing advantage in the increasingly standardized consumer credit and home mortgage markets. Meanwhile, community banks remain important for less-standardized types of lending, such as small-business loans and loans collateralized by business real estate. Moreover, as discussed above, community banks that survived the consolidation trend have actually increased their market share, offsetting some of the effect of community bank mergers.

Looking only at industry shares of assets held by all community banks, however, does obscure some important differences between commercial banks and saving institutions as separate segments of the industry.⁴¹ In 1994, the largest organizations controlled only 8 percent of savings institution assets but had already come to dominate the commercial banking sector. Although some large organizations (such as Citigroup) have increased their presence, midsize organizations continue to dominate the savings institution industry. Meanwhile, the shares of consumer credit and home mortgages held by community savings institutions have declined less (in relative terms) than the shares held by community commercial banks. However, in both segments of the industry, community banks appear to be holding their own as business lenders, particularly in funding small loans to businesses and farms.

Turning to the liability side of the banking industry's balance sheet, we see in the bottom panel of table 5 the changes in the distribution of bank liabilities and equity across bank size groups between 1985 and 2003. Community banks continue to hold higher shares of deposits (compared with their share of banking sector assets) and rely less on other types of borrowing than larger organizations. However, community banks' shares of the industry's deposits have generally moved lower with their overall share of industry assets. Recently concerns have been expressed about whether Federal Home Loan Bank (FHLB) advances are propping up small institutions, and we note that the share of total FHLB advances owed by community banks appears also to have tracked their declining share of the industry.

Table 5 - Banking-Sector Balance-Sheet Ratios: 1985, 1994, and 2003

There are, however, some differences between commercial banks and savings institutions in how liabilities are distributed across the bank size classes.⁴² First, the share of commercial bank equity held by community commercial banks has declined more than these banks' share of commercial banking assets; the opposite has been true for community savings institutions. These contrasting patterns reflect differences in the types of institutions that needed to be recapitalized after the banking sector problems of the 1980s and early 1990s. In the commercial banking industry, it was the larger institutions that needed greater recapitalization, whereas in the savings institution industry, recapitalization was more pronounced among smaller institutions. Second, in the commercial banking sector, community banks account for a disproportionate share of total FHLB advances, but among savings institutions, the opposite is true: community banking's share of total FHLB advances to saving institutions has been declining as borrowing among institutions controlled by the very largest organizations has expanded dramatically.

Portfolio Ratios
To understand what trends in the distribution of banking industry assets and liabilities imply for the portfolio composition of community banks vis-à-vis their larger counterparts, we constructed parallel data that measure portfolio ratios for community banks and for their larger counterparts. Again, we first discuss trends evident for all community banking institutions and then highlight key differences between community commercial banks and community savings institutions. Table 5 reports portfolio ratios for each size class of banks (community banks, midsize banks, and the top 25 banking organizations).⁴³

Given trends in industry shares on the asset side of the balance sheet, it is not surprising that community banks have increased their business real estate lending—including commercial real estate loans, farm real estate loans, and construction & land development loans—as a share of their assets. In contrast, the largest banking organizations have not exhibited comparable shifts. Instead, consumer credit and home mortgage lending now account for greater shares of the total assets controlled by the 25 largest banking organizations.

It is important that increases in business real estate lending by community banks are not merely substituting for other types of lending (such as C&I loans or consumer credit). After moving lower during the late 1980s and early 1990s, the loan-to-asset ratio for all community banks rose from 57 percent in 1994 to more than 63 percent in 2003. To some extent this increase undoubtedly reflects lending opportunities associated with the economic expansion of the 1990s. These portfolio trends, however, also reflect community banks' need to generate sufficient earnings to maintain profitability.

Turning to the composition of community bank liabilities, one finds (as mentioned above) that anecdotes about the reliance of community banks on retail deposit funding are borne out by the data. Although deposits as a share of total liabilities for community banks are lower than a decade ago, this share still exceeds 90 percent, and these deposits are almost all domestic deposits. Portfolio ratios also indicate that FHLB advances have become a more important funding source for community banks; but this is also true for larger banking organizations (table 5, lower panel).

There are significant differences in portfolio composition between community commercial banks and community savings institutions, particularly on the asset side of the balance sheet.⁴⁴ As discussed below, these differences have important implications for the relative performance of, and prospects for, these two types of community banks. The increase in community banks' loan-to-asset ratio reflects greater lending (as a share of assets) by commercial community banks, and specifically more real estate lending of all types. In contrast, community savings institutions (which historically have had higher loan-to-asset ratios than community commercial banks) remain primarily home mortgage lenders.⁴⁵ In mid-2003, 38 percent of community savings institution assets were home mortgage loans, and another 13 percent were mortgage-backed securities; the next-largest loan component was commercial real estate lending, which accounted for 9 percent of savings institution assets.

Community commercial banks and community savings institutions differ as well in the composition of their liabilities.⁴⁶ The former rely more on deposits and less on other borrowing—mainly FHLB advances—than do the latter. And although both community commercial banks and community savings institutions have increased their reliance on FHLB borrowing as a source of funds, large savings institutions rely more on FHLB advances as a source of funding. On the other hand, the recapitalization of savings institutions has reduced this sector's overall riskiness in terms of their leverage measured relative to their buffer stock of capital.

Performance
Despite or perhaps because of their differences from larger banking organizations, community banks have been able to compete with the larger organizations in terms of performance during the past decade. Aggregate performance patterns of institutions in different size classes suggest that community banks have been able to earn more as lenders than larger organizations have, but community banks also face rising relative operating costs. Here we analyze aggregate performance trends for community banks and larger banking organizations, highlighting the differences between community savings institutions and community commercial banks that reflect the composition of their portfolios—particularly on the asset side of the balance sheet. As we discuss below, these differences suggest that community banks that engage primarily in home mortgage lending (i.e., community savings institutions) do not generally have the same competitive advantages as either their larger counterparts or community banks that are primarily commercial lenders (i.e., community commercial banks).

Table 6 reports aggregate performance ratios for all banking organizations (by size category) from 1985 through 2003. Since 1993, community banks have tended to earn a healthy return on assets (ROA), exceeding 1 percent. And until very recently, the ROA for the community banking sector was very comparable to that earned by the 25 largest banking organizations (although the ROAs measured for midsize banks exceeded those mea-sured for both of these groups). However, because smaller institutions have tended to have higher capital ratios than larger institutions, a given level of earnings has translated into a lower return on equity (ROE) for the smaller institutions.⁴⁷ Thus ROE measured for community banks is below that for larger banks, and the ROEs earned by small community banks have tended to be lower than those for larger community banks.

Table 6 - Banking-Sector Performance Ratios, 1985-2003

Among commercial banks, earning differentials across the three size groups do not reflect poorer interest margins for community banks.⁴⁸ To the contrary: their profitability reflects higher net interest margins earned by these smaller banks. Even among community banks in the commercial banking sector, the smaller ones have tended to have higher net interest margins than the larger ones. However, the size-related differentials in net interest margins among all but the very largest banks have narrowed in recent years. At the same time, smaller banks have increasingly faced higher relative costs, here measured by the ratio of noninterest expenses to the sum of net interest and noninterest income. In terms of this "cost ratio," the gap has been growing between community banks and their larger commercial bank counterparts; the gap has also been growing among community banks in the different size categories. These patterns suggest an "economies-of-scale" interpretation of performance differentials across the bank size groups during the past decade. Although smaller institutions earned more on their assets, these earnings did not translate into higher ROAs because smaller institutions also had higher costs. Moreover (as noted), the need to hold more capital translated into lower equity returns among community banks.

Performance differentials evident among saving institutions in the different size classes appear to reflect the continuing role of these institutions as mortgage lenders.⁴⁹ ROA measured for large savings institutions have been rising relative to the ROA measured for community savings institutions. And net interest margins for community savings institutions have moved closer to those earned by their large counterparts, while cost ratios for community saving institutions have been rising. Therefore, the lower profitability evident for community savings institutions appears to reflect the higher costs facing these banks, without the higher net interest margins to cover them. Overall, these patterns suggest that community savings institutions face greater competitive disadvantages than their commercial banking counterparts, which are more focused on business lending. In addition, these patterns are consistent with the evolution in mortgage lending toward standardized transactions in a national market.

Performance and Community Bank Migration

To better understand the declining population of community banks, it is useful to compare the relative performance of institutions that remained community banks with the relative performance of institutions that outgrew the size classification or were acquired by larger banks. In particular, was it the better performers that became part of the population of larger banks? To examine this question, we tracked the performance of all institutions (other than de novos) that had originally been classified as community banks to see if there were differences in performance between those that were still classified as community banks in subsequent years and those that had either grown out of the community bank classification or been acquired out of it. Because industry conditions in the 1980s and early 1990s were starkly different from conditions in the later 1990s, we conducted separate analyses of the two nine-year subsets of the 1985–2003 period. We also analyzed commercial banks and savings institutions separately. For each year (and both segments of the banking industry) we first measured the performance of institutions that had been classified as community banks at the beginning of the eight-year period and were still community banks as of the year in question; we also measured the performance of institutions that had been community banks at the beginning of the period but had outgrown the classification or been acquired by larger banks in a given year.⁵⁰

Certain general patterns emerged from this analysis. Not surprisingly, patterns evident for the banks that outgrew the community bank size classification are consistent with some of the size-related performance differentials discussed above. However, a comparison of community banks that were acquired with those that remained community banks fails to suggest that those continuing as community banks were generally poorer performers.⁵¹ Moreover, differences in performance between banks that remain community banks and those that outgrow the classification are likely to reflect differences in the economic conditions in the markets where they are located.

Performance and Local Market Conditions

Because of community banks' small size, their portfolios and performance have an inherently local dimension. In analyzing their performance, therefore, we examined the extent to which community bank performance has been related to longer-term local-market demographic and economic prospects in the markets where these institutions are located. Some recent studies have looked at the prospects for community banks in rural markets where population has been declining.⁵² But it is also useful to look at the prospects in high-growth areas, where community banks may play an important role in meeting the strong small-business loan demand attendant on local growth.

In this analysis of the relationship between community bank performance and longer-term conditions in local banking markets, we classified longer-term local market conditions in terms of population growth between 1985 and 2003.⁵³ Each market (defined in terms of metropolitan statistical areas [MSAs] and non-MSA counties) was placed in one of three population-growth classes: (1) low growth if population growth was negative, (2) moderate growth if average annual population growth was between zero and 2 percent, and (3) high growth if population growth averaged more than 2 percent per year during the 1985-2003 period.

During this period, urban markets had higher growth on average than rural markets, although 12 percent of MSAs had negative population growth.⁵⁴ These negative-growth metro markets tended to be in the northeastern United States, whereas high-growth metropolitan areas tended to be in the South and West. Not surprisingly, rural markets tended to have lower population growth (and lower real personal income growth) than urban markets: 40 percent of rural counties experienced negative population growth between 1985 and 2003, 49 percent had moderate population growth, and only 10 percent had high population growth.⁵⁵ As in the analysis above of changes in the number of community banks, we examined the link between local population growth and community bank performance in particular market segments: rural markets, small metro markets, and suburban and urban parts of large MSAs.⁵⁶

We calculated five performance measures for community banks headquartered in markets that experienced negative, moderate, or high population growth.⁵⁷ The results indicate that community banks located in markets exhibiting higher growth during our study period tended to have greater earnings growth and, for the past decade, somewhat higher ROAs and larger net interest margins. At the same time, cost ratios also exhibited some relation to local market conditions, with community banks in higher-growth markets also tending to have higher expenses relative to their income.⁵⁸ In recent years, however, cost ratios have tended to converge across markets. Higher net interest margins suggest that community banks in robust regions have benefited from local lending opportunities to a greater extent than community banks in lower-growth markets.

On the other hand, even community banks in low-growth (by our definition, negative-growth) markets seem to have been buoyed up by the economic expansion of the 1990s. Although community banks in higher-growth markets have higher loan-to-asset ratios, community bank lending (relative to assets) has increased most in low-growth markets. And it is noteworthy that even in regions with low growth, community bank performance has been solid during the past decade, as has the performance of the banking industry generally.

It should also be noted that there are some qualitative differences in the relationship between local growth and bank performance ratios in different types of markets.⁵⁹ For example, we find the greatest variation in cost ratios for community banks in urban parts of large metro markets, where local rents and other costs are likely to be more sensitive to local conditions. But the general patterns, particularly for profitability and net interest margins, are evident in the different segments of local markets studied here.

To further test the robustness of the relationship between local growth and community bank performance, we constructed comparable sets of performance measures that tracked the performance of institutions that had been community banks in 1985.⁶⁰ Specifically, we constructed measures to reflect the performance of all community banks in 1985 that had neither failed nor been merged into a bank not classified as a community bank.⁶¹ These measures were intended to explore the notion that observed community bank performance could understate true community bank performance because top performers outgrow the classification.⁶² Somewhat surprisingly, there was very little difference in the cohort-level performance mea-sures for this broader group when compared with the cohort-level performance measures for community banking organizations classified in terms of their current asset size. Hence, at least for this period, banks exiting the community banking population (through internal growth or acquisition activity) do not appear to be markedly better performers than community banks that remained community banks.⁶³ In sum, although there is a clear link between the local environment and the performance of community banks, community banks seem quite able to survive in a variety of environments.

Competition Faced by Community Banks

Community banks face many competitors. A useful way to assess these is with an examination of just what kinds of financial institutions community bankers themselves have identified in surveys as their most significant competitors. (See tables 7 and 8.) Community bankers view other community banks as their prime competitors, and also see credit unions as significant competitors. When community banks are broken down by asset size, the same pattern emerges, but community bankers running the largest community banks view regional or megabanks as competitors much more than do those running the smallest community banks. The idea that other community banks are prime competitors is borne out when one looks at core business lines: other community banks are regarded as the main competitors in short- and medium-term loans to businesses and farmers, unsecured loans to consumers, and consumer and business deposits. Other types of firms are prime competitors in markets that have a nationwide reach—for example, finance companies for auto loans, and mortgage companies for first mortgages. (See table 9.) Perhaps surprisingly, regional banks and megabanks are not identified as the most important competitors except in a few business lines, such as business and personal trust operations and home equity loans.

Table 7 - Perceived Competitors of Community Banks, Survey Results, 1994-2003 (Percentage of Respondents)

Table 8 - Perceived Competitors of Community Banks, by Community Bank Size, 2004 (Percentage of Respondents)

Table 9 - Top Two Competitors of Community Banks by Business Lines, 2003

The attitude of community bankers (particularly those associated with smaller institutions) to their large-bank competitors may reflect their belief that their business model is effective in its emphasis on reputational lending and personal service and that they have an advantage in their presumed ability to attract customers dissatisfied by the more impersonal approach of large banks. Credit unions would seem to be a natural competitor to community banks and might well have been viewed as even more significant competition in the surveys, were it not for the effects of both their size and their location. Among credit unions, as of year-end 2003, 88 percent held under $100 million in assets, whereas only 50 percent of community banks were in that size category. And credit unions are located mainly in urban areas in the central and eastern states. Eighty percent of credit unions are located within MSAs, whereas only 53 percent of community bank offices are (as of midyear 2003). These differences in size and geography suggest that some community banks face formidable credit union competition, while others do not.

For several reasons (credit unions' perceived importance as competition, their tax-exempt status, and their exemption from provisions of the Community Reinvestment Act, as well as legislation that has allowed credit union membership to expand significantly), it is useful to look briefly at the trends in deposit share of credit unions and community banks. Community banks' share of deposits, as noted above, has decreased steadily, from 25.5 percent in 1994 to 17.7 percent in 2003; during the same period, credit unions with over $100 million in assets have seen their deposit share increase, while smaller credit unions have lost deposit share. Deposit share growth was greatest for the largest credit unions (those with more than $1 billion in assets—by our definition larger than a community bank). Overall, credit union deposit share has increased, but not dramatically (from 7.7 percent to about 9.5 percent).⁶⁴ By far the largest gains in deposit share have been made by the 25 largest banks, with their share rising from 28.4 percent in 1994 to 44.2 percent in 2003.⁶⁵ (See table 10.) An examination of deposit share in rural areas and in large and small metropolitan areas finds similar general patterns: to varying degrees, in each of these areas community banks have lost deposit share, credit unions with more than $100 million in assets have had gains in deposit share (and again, credit unions with more than $1 billion in assets saw the greatest percentage growth in deposit share—particularly in all types of metropolitan areas), and the largest banks have experienced gains (for the largest banks, the gains were especially strong in urban and suburban areas within large MSAs).⁶⁶ In terms of credit union deposit share, overall it has been the large credit unions that have experienced the fastest growth; most credit unions, however, remain small in comparison even with small community banks. Whether credit unions will capture a significantly greater share of the market in the future remains an open question.

Table 10 - Share of Deposits and Offices
Community, Midsize, and Top 25 Banks vs. Large and Small Credit Unions, 1994-2003

The extent to which these trends in deposit share are causally related is hard to determine. Because the geographic overlap between community banks and credit unions is limited, changes in the aggregate deposit shares of these groups may not reflect increased competition between them. To address the question of competition, we performed a separate analysis using markets with both a community bank presence and a relatively large and growing credit union presence. In these markets, community banks experienced a decline in deposit share that was not, on average, significantly greater than the decline they experienced in all areas. For example, in the 20 MSAs in which small credit unions both increased their deposit share from 1994 to 2003 and held 8 percent or more of the deposits at the end of this period, community banks experienced a mean loss in deposit share of about 7 percent. In comparison, the mean decline in deposit share of community banks in all MSAs during this period was approximately 6 percent, and the median declines were almost identical. An analysis of rural counties yielded similar results.

It is useful to examine the industry more closely by looking at credit unions according to their membership types.⁶⁷ Perhaps most credit unions, especially the smaller ones that make up much of the industry, have a local component and so could be seen as competitors with community banks. Federally chartered community credit unions, which are defined as those whose members are within a well-defined local community, neighborhood, or rural district, might be perceived as particularly competitive with community banks, especially as recent NCUA rules have allowed for broad interpretations of "local."⁶⁸ Our analysis will therefore emphasize trends observed in these institutions.

The credit union industry, like banks, has undergone significant consolidation in recent years. When we look at the number of institutions within this context, the only type of credit union that has seen significant growth since 1997 is the community credit union, though at about 1,000 institutions these still make up only about 11 percent of the industry. Since 1997, community credit unions' industry asset share has also increased, from just over 3 percent to more than 12 percent (state-chartered credit unions' asset share has also increased, while both single- and multiple-common-bond credit unions' shares have decreased). Community credit unions' average asset size has also seen the fastest growth, more than doubling to over $76 million dollars in 2003, although this size means that they remain small in comparison with all but the smallest community banks.⁶⁹

Some industry observers have asserted that credit unions are acting more like commercial banks, and point to their increasing entry into business loans as evidence for this trend. Although some individual credit unions are engaging more heavily in this activity and the industry ratio of member business loans to assets has nearly doubled since 1997, this measure, until now, has grown only to just under 1.5 percent. Community credit unions' ratio of business loans to assets dropped between 1997 and 2000 but has risen since then, from just over 1 percent to about 1.8 percent. So although business loan activity has been increasing in recent years, on an industry-wide basis it remains a relatively small part of credit union lending. In terms of performance, credit unions overall have done reasonably well since 1997, though not as well as community banks. Community credit unions' ROAs have generally been the lowest of the four types since 1997, dipping as low as 80 basis points in 2001, though in the past two years they have performed more in line with state-chartered and multiple-common-bond institutions, with ROAs above 90 basis points (single-common-bond credit unions have performed best). Credit union ROEs have been lower than those of community banks, and since 1997, community credit unions' ROEs have been either in line with or somewhat lower than other types', starting the period at about 9 percent, dipping to about 7.5 percent in 2001, but recovering to about 8.7 percent in 2003.⁷⁰

Prospects for Community Banks

Our examination of community banks' future must take into account what may happen to their numbers, as well as these banks' competitive strengths and challenges.

Decline in Community Bank Numbers
Merger activity has slowed in recent years; coupled with the continued creation of new banks, this has meant a significant reduction in the consolidation of community banks. Furthermore, the pattern of community banks' numerical decline does not suggest that any one type of area or market is particularly likely to face accelerated consolidation in the near future. Additional declines may nevertheless be expected. Low returns on equity (resulting partly from higher capital ratios) may lead to consolidation of some institutions, as stockholders seek higher returns through increased leverage at merged institutions. This presumed causal relationship is less relevant, however, to owner-operated banks that do not rely on uninsured or unprotected sources of funds. These owners' returns may be increased by compensation received as bank officers, and there may be nonpecuniary benefits to playing a leading role in the local community. In addition, there may be no outside shareholders to challenge the owners' decisions to remain independent. But financial considerations may not be the only reason for consolidation among community banks. Indeed, in view of apparent lags in the response of individual community banks to market developments, the depth and timing of future consolidation among community banks remain uncertain. These lags may reflect not only a lack of interest on the part of outside banks in acquiring banks located in slow-growth markets but also, as just mentioned, the ability of banks in these markets to perform at levels satisfactory to their owners.

Competitive Strength and Challenges
We have seen that community banks that avoided acquisition by larger banks had relatively strong growth rates and sustained profitability. These growth rates and profitability have been partly attributed to the institutions' ability to underwrite and monitor loans to small businesses that might have been ignored by large, distant institutions. Some observers suggest that relationship lending is likely to become less important as more data become available on small businesses' performance and as further technological advances make it possible to disseminate to lenders more hard data on small businesses.⁷¹ Similarly, personal service based on local presence and direct contact with customers may become somewhat less important as younger customers, accustomed to transactions on the Internet, grow older and become financially dominant. Still, it would be a mistake to discount the future importance to many bank customers of direct contact with bank employees—or at least with tangible bank premises. How else to explain that in the face of advancing technology and increasing computer literacy, the number of bank offices has held relatively steady and the number of ATMs has continued to grow beyond earlier expectations? The number of bank offices, despite the tremendous consolidation that has occurred in the industry, has hovered within a fairly narrow range since 1985, and more recently has been increasing (see table 11); the number of ATMs increased 241 percent between 1990 and 2001.⁷² It seems that for the foreseeable future, the ability to offer personal service economically will be a competitive advantage for many small banks. Another competitive strength is the strong market position of community banks (including de novos) in economically healthy rural, suburban, and small metropolitan markets. With sizable market shares in such areas come customer recognition and awareness, which are likely to be advantageous in the future.

Table 11 - Bank Offices by Size Category, 1985–2003

Community banks' competitive strengths, however, must be matched against the competitive challenges they face in a number of respects. The need to attract and hold qualified personnel will remain a significant issue. Community banks may face the need to find a more diverse set of both funding options and sources of income. Possibilities for moving beyond reliance on core deposits exist, as does the potential for generating more fee-based income, but difficulties may accompany such courses. Community banks, having at first lagged behind large banks in adopting and using technology, are now rapidly making up their deficit and must navigate the best path for its use. Finally, community banks face the problem of the fixed costs incurred in complying with banking regulations.

Surveys of community bank executives indicate that attracting and retaining qualified personnel is perhaps these executives' most important concern. Retaining key employees was identified as a factor critical to success by 93 percent of respondents in the Grant Thornton survey published in 2003, including those from both large and small community banks.⁷³ Although only 13 percent of respondents in an American Bankers Association (ABA) survey published in 2003 found retention a significant problem (a situation probably reinforced by a soft economy), the survey nevertheless demonstrated that finding qualified candidates for important positions was often hard (and sometimes very hard). (See tables 12 and 13.) The difficulty filling a particular type of position sometimes depended on the community bank's size or location or both. There was, however, some general concern about the unavailability of qualified employees and a belief that potential employees were moving into nonbanking jobs, as well as concern that large banks were better able to attract personnel.⁷⁴ (See table 14.) Overall, community banks located in declining or slow-growth economies are likely to experience the most difficulty with employment.

Table 12 - Problems Attracting and Retaining Employees, 2000-2003 (Percentage of Respondents Reporting Significant Problems)

Table 13 - Level of Difficulty in Finding Qualified Candidates, 2003 (Percentage of Respondents)

Table 14 - Reasons for Problems Attracting and Retaining Employees, 2002 (Percentage of Respondents)

Community banks largely depend on core deposits, and for many community banks, this may not pose any concern.⁷⁵ In recent years, however, core deposits in many areas have lagged behind total deposits and behind loan demand. Responses to ABA surveys from 1997 to 2001 suggest that deposit growth lagged behind loan demand at a significant number of community banks. More recently, a return flow from the stock market has eased funding problems, but perhaps only temporarily.⁷⁶ Raising rates to increase core deposits in the local community may be costly because of cost increases for existing accounts, or because of a limited supply of local funds in slow-growth areas. Funds attracted from the outside through brokers or the Internet may be volatile.

At this point, Federal Home Loan Bank (FHLB) advances are the main supplement to core deposits. There has been some use of other "nontraditional" sources, such as fed funds, sales of participations, and repurchase agreements.⁷⁷ (See table 15.) Measured as a percentage of community bank liabilities, FHLB advances clearly have been used increasingly since 1993; these advances moved from just over 2 percent of liabilities in 1993 to almost 6.8 percent in 2003. Not surprisingly, larger community banks have made greater use of advances during the period, but even among the smallest community banks the use of advances has increased substantially. (See figure 2.) Although narrower funding options may be a handicap, community banks have been able to offset higher interest costs with higher loan rates charged to idiosyncratic borrowers who have limited access to large-bank funding.

Table 15 - Nontraditional Funding Methods of Community Banks
(Percentage of respondents)

D

In principle, small banks have fewer prospects for diversifying their sources of income; mergers may allow them to expand their opportunities to do so. Diversification does not, however, lead only to benefits. A recent study notes that diversification can carry risks because community banks may "move beyond areas of comparative advantage and enter businesses where they lack the necessary expertise, technology, or scale to compete successfully," and concludes that community banks "do better when they stay focused on major activities but gain by diversifying within that area of expertise."⁷⁸

Some observers have suggested that community banks should rely more on fee income. Fee income of small banks is largely from deposit services, and (as noted above) fees on deposit services tend to be lower at small banks than at large banks. Because small banks generally offer a narrower service base, their total fee income is proportionally lower than that of large banks. Broadening fee income would make community banks less dependent on core deposits and less vulnerable to reductions in interest margins in periods of rising interest rates. Raising fees associated with deposit services, however, might conflict with community banks' efforts to attract retail deposits and with their personal service image. As noted above, the gap between large- and small-bank fees appears to be closing, so some community banks may be making greater use of fees to generate income.⁷⁹ In a survey recently conducted by the Federal Reserve System (admittedly with a very small sample—only ten bankers), half the participants noted that they intended to increase revenues through fee income, and those interviewed realized that "chronically delinquent or overdraft customers are profitable."⁸⁰ Broadening fee income from other sources generally entails broadening the service base. However, some of the sources of fee income of large banks, such as investment banking, securitization, and back-up lines of credit, require a large base of transactions and therefore are not feasible for small banks.

Community banks will face a strategic choice between trying to perform relatively narrow functions more efficiently and trying to broaden into new activities that may involve greater risk and greater cost for small-scale operations.⁸¹ It is not clear which course will prevail or whether all banks will make the same choice. The important point seems to be that numerous community banks have found it possible to grow and prosper not by trying to emulate large banks with many business lines but by performing largely traditional functions more efficiently.⁸²

Although automation of back-office operations is essentially universal, community banks have lagged behind large banks in adopting other technology—specifically, Internet banking. However, delayed entry after some of the initial problems have been resolved has not necessarily been a major disadvantage.⁸³ Moreover, community banks are adopting Internet services fairly rapidly. One measure of the Internet presence of community banks is found in their reporting a Web address on their Call Reports. As of September 30, 2003, there were 7,374 community organizations containing 5,663 institutions that reported Web addresses. A survey conducted by the Independent Community Bankers of America indicates that 77 percent of community banks have Internet sites and that 75 percent of these community banks have transactional Web sites;⁸⁴ an earlier Grant Thornton survey suggests similar magnitudes.⁸⁵ Small banks are less able to make large investments in technology individually but have generally been able to meet their needs by outsourcing or purchasing widely available systems for in-house use. However, when a community bank is just one client of a service provider, the bank may face varying levels of loss of flexibility.⁸⁶ The great uncertainty, however, concerns future technological change. De Young and Hunter have laid out two possibilities: First, new technology may enable large banks to personalize their services while maintaining the advantages of large-scale operations. Second, small banks may be able to retain the advantages of their personalized approach while overcoming the disadvantages of small-scale operations.⁸⁷ It is difficult to know which of the two is more probable.

It should also be noted that the provision of a bank safety net and the existence of regulatory agencies to enforce compliance have led to substantial reporting and other regulatory burdens. These requirements normally involve fixed costs that tend to be proportionally heavier for small banks; thus, regulatory burden is likely to have some effect on these banks' long-term prosperity.⁸⁸ Conceivably such regulatory requirements will contribute to further consolidation.

Conclusion

Community bank numbers shrank by almost half between 1985 and 2001; during this period their market share dropped considerably, with small community banks affected most. Taken at face value, such observations might suggest that community banks face considerable difficulties. However, more detailed examination presents a rather different picture. Community banks made up 95 percent of the industry in 1985, yet despite the tremendous consolidation that has occurred, they still constituted 94 percent of U.S. banks in 2003. Moreover, closer inspection of the geography of consolidation reveals some surprising results: by many different measures, community bank declines were proportionally similar. The number of community banks decreased only slightly faster in formerly unit-bank states than in non-unit-bank states. Community banks did not disappear more rapidly in declining markets than in growing ones. Declines were similar across rural, small metropolitan, and large metropolitan areas, and within the last category, across both urban and suburban areas, although the factors that contributed to community bank numbers (failures, mergers, and new banks) differed subtly with the type of area. In particular, both mergers and de novos were concentrated in large metropolitan areas. Community banks have been able, however, to maintain their presence in all areas.

Community banks experienced a significant loss of deposit share between 1985 and 2003, a period during which large banks were by far the largest gainers in deposit share, and credit unions saw modest gains. Community banks also had significant losses in asset share; in lending this was most pronounced in sectors (such as consumer credit) that have been commoditized. They are, however, holding their own in real estate lending to businesses. It should be noted that community savings institutions remain primarily mortgage lenders. As regards earnings, community banks are at a minimum performing satisfactorily, and the performance of community commercial banks reflects higher net interest margins sufficient to offset higher costs. Community banks did perform better in high-growth than in low-growth markets, but even in the latter their performance may be regarded as acceptable, with returns on assets of 100 basis points or better for the last decade of the period studied.

Overall it is impressive that community banks, while facing intensified competition due not only to the removal of branch restrictions (which had protected many from competition) but also to the growth of nonbank competitors, have been able to achieve both respectable earnings and growth in recent years. Community banking, therefore, appears to be a viable business model. Research suggests that these banks possess certain advantages as lenders to small businesses, small farmers, and other informationally opaque borrowers through their ability to assess the risks of borrowers who lack long credit histories, to process soft data such as borrower reputations, or to operate effectively in situations where the proximity of decision making to customers is important. The proposition that community banks have informational advantages in lending to small business is supported by research suggesting that small banks have higher risk-adjusted returns on business loans than large banks. The willingness of private investors to risk their own money to create new banks is a powerful market test of the viability of small banks, at least in areas of high population density. Moreover, a concentration of new banks in areas where large and distant banks have taken over local institutions also suggests that many customers may prefer the more personal approach of community banks. Consumer attitudes may change and larger banks may seek to emulate the personal service approach of smaller institutions. However, community banks should continue to occupy an important position in the banking industry for the foreseeable future if policymakers can maintain economic stability and moderate the impact of regulatory burden.

^*The authors are all with the Division of Insurance and Research of the Federal Deposit Insurance Corporation: Tim Critchfield is a senior financial analyst; Tyler Davis, a research assistant; Lee Davison, a historian; Heather Gratton, a senior financial analyst; George Hanc, a former associate director; and Katherine Samolyk, a senior financial economist. The authors thank Robert DeYoung for his helpful comments.

¹ Consolidation in the 1990s mostly involved mergers between two community banks, and merger targets were usually community banks (DeYoung, Hunter, and Udell [2003], 14; their paper also provides a useful history of the relaxation of legal impediments to branching during the past 30 years).

² These figures count First National Bank of Keystone as a community bank. Although it had slightly more than $1 billion in assets the year before it failed, it had grown very quickly for the previous five years and so was well below $1 billion in assets during most of the period when it engaged in the high-risk policies that ultimately led to its failure.

³ Under this definition, a bank or thrift that has less than $1 billion in assets but is within a holding company with more than $1 billion in assets is therefore not a community bank. (In this paper, the terms thrift and savings institution are synonymous.)

⁴ Spiegel, Gart, and Gart (1996), 18-19; Kline (1997).

⁵ Isaac (1991); De Senerpont Domis (1996); Kline (1997).

⁶ Stein (2002), 1912.

⁷ See, for example, Nakamura (1994); Berger and Udell (2002); DeYoung, Hunter, and Udell (2003); Brickley, Smith, and Linck (2001); and Berger and Udell (2003).

⁸ Berger et al. (2002).

⁹ Carter, McNulty, and Verbrugge (2004).

¹⁰ Keeton, Harvey, and Willis (2003), 28.

¹¹ Timothy Hannan, cited in Keeton, Harvey, and Willis (2003), 28.

¹² For simple passbook accounts, the dollar amounts were $36.96 versus $21.48; for stop-payment orders, $23.54 versus $17.00 (Federal Reserve Board [2003], appendix B). The Federal Reserve Board defines small banks as institutions with less than $100 million in assets; medium-size banks, assets between $100 million and $1 billion; and large banks, more than $1 billion in assets. In 2002 medium-size banks' fees were usually somewhere between the fees of small and large banks.

¹³ Federal Reserve Board (1999); Kimmelman (1999).

¹⁴ Timothy Hannan and Robin A. Prager, cited in Keeton, Harvey, and Willis (2003), 28.

¹⁵ Bassett and Brady (2002).

¹⁶ Grant Thornton (2002). Grant Thornton mailed surveys to the chief executives of 5,393 community banks and savings institutions in November 2001. The response rate was 8 percent.

¹⁷ Keeton (2000). See also Berger et al. (1999); Seelig and Critchfield (2003).

¹⁸ DeYoung and Duffy (2002), 9.

¹⁹ For example, it is not unusual for large banks to advertise "relationship banking accounts," and many large banks seek to be customer-friendly by turning their branches into "stores."

²⁰ Because this comparison has been adjusted for inflation, it compares the number of banks in 2003 that had less than $100 million in assets with the number of banks in 1985 that had less than $66 million in assets. See table A.2.

For a discussion of chartering policies in the 1980s, see FDIC (1997), 106ff.

²² Twenty-one have disappeared: 17 were voluntarily liquidated, and only 4 failed.

²³ Yom (2003).

²⁴ DeYoung (2000), 5. DeYoung notes that the payment of dividends by these banks is also restricted.

²⁵ The Federal Deposit Insurance Corporation Improvement Act of 1991.

²⁶ Seelig and Critchfield (2003).

²⁷ DeYoung (2000) notes that in his analysis, banks chartered closest to the peak of the "banking recession" failed at relatively high rates. For a discussion of the effect of the recessions of the 1980s and early 1990s on banking, see FDIC (1997).

²⁸ Interestingly, banks growing large enough to leave the ranks of community banks made up a steady trickle of the decrease during the entire period, except briefly when capital standards were being increased in response to the banking crisis and few or no banks managed to grow out of the community bank classification.

²⁹ The states described as having "prevalent unit banking" (a categorization determined by the Conference of State Bank Supervisors "based on the type of banking seemingly prevalent in each state" ) as of year-end 1977 were Colorado, Illinois, Kansas, Minnesota, Missouri, Montana, Nebraska, North Dakota, Oklahoma, Texas, West Virginia, and Wyoming (Conference of State Bank Supervisors [1978], 95).

³⁰ See tables A.1 and A.2.

³¹ We measured community bank presence by looking at the location of a bank's headquarters. It is important to go beyond simple comparisons between MSAs (metropolitan statistical areas) and non-MSAs and to identify suburban areas. By our definition, only large MSAs (those with populations over 500,000) can have suburbs. Initially, the urban area within the MSA was defined by the Census's central city, so all counties within those central cities were identified as "central counties"—hence, urban counties. However, significant numbers of central cities in large MSAs spanned multiple counties, so the "central-city" measure was less useful. Therefore, in large MSAs that had more than two central (or urban) counties, an adjustment was made: if population density in the MSA exceeded 1,000 per square mile, all counties that exceeded this density were designated as urban; all other counties in those MSAs were designated as suburban. In large MSAs where the population density was less than 1,000 per square mile, any county that exceeded the median population density of that MSA's central counties was classified as urban; those below the median were classified as suburban. It should be noted that we used 2003 census classifications of counties and projected them back to 1985. Therefore, if a county became part of an MSA at any time during the period, that county would always have been classified as part of an MSA.

³² See table A.3.

³³ See tables A.5 and A.6.

³⁴ Walser and Anderlik (2004).

³⁵ This region includes the states of Arizona, California, Colorado, Idaho, Montana, New Mexico, Nevada, Oregon, Utah, Washington, and Wyoming. The 93 counties that experienced decline saw a decrease of about 11 percent from a population of 910,000. The 250 counties that experienced population growth had an increase of 32 percent from a population of 5.8 million.

³⁶ Whitener and McGranahan (2003). In the Department of Agriculture's study, rural counties were those outside of MSAs.

³⁷ Bassett and Brady (2001), 722. It should be noted that these authors' definition of "small" banks does not conform to our definition of community banks. Bassett and Brady defined small banks as insured commercial banks with an asset size below that of the largest 1,000 banks (in other words, with assets below $331 million in 2000). They defined large banks as the 100 largest institutions (assets of at least $6.94 billion in 2000); institutions between these two size groups were defined as medium-size. Medium-size banks experienced greater "merger-adjusted" asset growth than large banks but less than small banks.

³⁸ Avery and Samolyk (2004), 320. This study looks at small-business lending by community banks in local banking markets, and it defines community banks as we do here.

³⁹ See table A.7.

⁴⁰ It should be noted that in both growing and declining markets, the larger market-share declines associated with bank consolidation activity have been offset by market-share increases for the remaining community bank population. In addition, the share of community banks' deposits held in low-growth markets actually declined during the 1985-2003 period.

⁴¹ See table A.8.

⁴² See table A.8, bottom panel.

⁴³ Comparable data for the two subsets, commercial banks and saving institutions, classified by the size of the banking organizations that control them, are presented in table A.9.

⁴⁴ See table A.9.

⁴⁵ This is true of larger savings institutions as well.

⁴⁶ See table A.9.

⁴⁷ With respect to earnings performance, pretax ROAs of community banks tend to suggest that profitability has been lower for smaller institutions than for larger banks in recent years (reported on table 6). However, the gap between community-bank ROAs and larger-bank ROAs is narrowed after corporate taxes are taken into account. Community banks hold a larger percentage of their assets in lower-yield, nontaxable municipal bonds. In addition, with the passage of the Small Business Job Protection Act of 1996 (effective January 1, 1997), banks that meet certain conditions have been able to convert to Subchapter S-corporation status. Such corporations are exempt from income taxation at the corporate level. Income is allocated to shareholders on a pro rata basis before taxation and is then taxed at the individual-shareholder level. Currently, approximately 1,800 community banks are S corporations.

⁴⁸ See table A.10.

⁴⁹ See table A.11.

⁵⁰ See table A.12. Here we are able to track only the performance of institutions that were originally classified as community banks and that still file Call Reports. We cannot track the performance of community banks that failed or were absorbed into a noncommunity-bank charter through a merger. Of course, many community banks were merged into institutions that remained community banks.

⁵¹ During the more troubled 1985-1994 period, however, bank health did appear to have been related to whether an institution outgrew the community bank classification, particularly for savings institutions. The relatively small number of savings institutions that moved out of the community bank size class tended to be those that were better capitalized and had fewer asset-quality problems. Among commercial banks, those that outgrew the community bank classification tended to have lower nonperforming asset ratios despite having significantly higher loan-to-asset ratios. During the 1994-2003 period, however, performance differences between banks that remained small and those that became larger were attributable to differences in size.

⁵² For example, a recent study of small-bank performance in the Kansas City Federal Reserve District assesses bank performance in counties with low per capita income growth. That study, however, focuses on the performance of banks in the 25 percent of counties in the district where per capita income growth was lowest. See Myers and Spong (2003).

⁵³ We compared levels in 2003 with levels in 1985 in each market. To quantify the changes in terms of annual averages, we divided the net change over the 18-year study period (a growth rate) by 18. We also looked at real personal income growth, which measures the growth of local economic activity and reflects both population growth and the growth of per capita income. The results for both types of growth classifications were similar; thus, only the results for the population-growth classification are discussed here.

⁵⁴ For the period 1985-2003, 62 percent of MSA markets had average annual population growth of between 0 and 2 percent, and 26 percent of MSA markets had average annual population growth of more than 2 percent.

⁵⁵ This is another way in which our analysis differs from that of Myers and Spong (2003). They look at the distribution of counties (both urban and rural) in the Kansas City Federal Reserve District in terms of growth, and simply classify the bottom quartile as low growth.

⁵⁶ See note 31 for an explanation of these market segments.

⁵⁷ See figures A.1-A.5.

⁵⁸ The cost ratio, also called the efficiency ratio, is generated in the following manner: noninterest expense less amortization of intangible assets as a percent of net interest income plus noninterest income. The ratio measures the proportion of net operating revenues that are absorbed by overhead expenses, so that a lower value indicates greater efficiency.

⁵⁹ See table A.13.

⁶⁰ For brevity, we do not report these results.

⁶¹ Thus we identified all banks that were part of a community banking organization in 1985 and included them in cohort-level performance measures even if they had outgrown the community bank size classification or been acquired out of it. We can include community banks that became affiliates of noncommunity banking organizations, since they still reported separate Call Report data. But we cannot include community banks that were merged into a noncommunity banking organization, since they no longer reported separate Call Report data.

⁶² These measures differ from measures reported for existing community banks in that they include institutions that have outgrown the community-bank size classification or institutions that have been acquired by large organizations.

⁶³ We also examined whether the community bank performance patterns in local market were affected if de novo institutions were included, and, again, we found patterns related to local growth that were similar to those reported here.

⁶⁴ For another recent examination of credit unions as competitors to small banks, see Gunther and Moore (2004), 10-11.

⁶⁵ It has been argued that a different picture of growth in credit union deposit share would be found if deposits held by the 25 largest banks were excluded. With the use of this (somewhat artificial) method, credit union deposit share has grown substantially, from 10.8 to 15.4 percent, whereas the deposit shares of community and midsize banks have decreased.

⁶⁶ See tables A.14 and A.15. As of this writing, branch-level deposit data for credit unions had not been collected by the National Credit Union Administration (NCUA). However, as of year-end 2003, 71 percent of all credit unions had less than $30 million in total assets, and the NCUA noted that most credit unions were small, localized operations. Therefore, our study assumes that the county in which a credit union is headquartered is an adequate proxy for the location of its deposits and is comparable to the Summary of Deposits data collected on banks. This assumption is particularly compelling for credit unions with under $100 million in total assets, which we refer to as "small" credit unions.

⁶⁷ The data allow us to break up the industry into four groups: single common bond, multiple common bond, community, and state-chartered credit unions.

⁶⁸ U.S. General Accounting Office (2003), 32-33.

⁶⁹ See table A.16.

⁷⁰ See table A.16.

⁷¹ See Petersen and Rajan (2003), 2535; also Berger and Udell (2003), 219ff.

⁷² Although some industry observers had predicted that the ATM would be obsolete by 2000, the growth rate of ATMs between 1996_and 2001 was significantly greater than it had been between 1983 and 1996. See Cobas, Mote, and Wilcox (2003), 51ff.

⁷³ Grant Thornton, Tenth Annual Survey of Community Bank Executives (2003). This survey was sent to 5,014 CEOs and senior officers of community banks and savings institutions and had a response rate of 10.2 percent.

⁷⁴ Cocheo (2002, 2003). The survey published in 2003 was sent to 5,474 top management subscribers to the ABA Banking Journal, mostly from banks under $1 billion in assets, and had a response rate of 14 percent. The survey published in 2002 was sent to 6,492 subscribers and had a response rate of 14.2 percent.

⁷⁵ Core deposits are domestic deposits less time deposits above $100,000.

⁷⁶ The percentages of respondents describing deposit growth as lagging behind loan demand from 1997 to 2001 were, respectively, 57, 59, 39, 48, and 66; in 2003, the percentage dropped to 33, with 40 percent of respondents stating that deposit growth exceeded loan demand (Cocheo [2000, 2001, 2003]).

⁷⁷ Surveys from 1999 to 2001 show increasing use of nontraditional methods of funding.

⁷⁸ Stiroh (2004), 137.

⁷⁹ Large banks still almost always charge higher retail deposit fees than do smaller banks. In many cases, however, although fees at both large and small banks have generally been increasing, the rate of increase at small banks has been more rapid. For example, between 1999 and 2002 overdraft fees at small banks went up 25.2 percent; at large banks, 16.9 percent; in 1999, large banks charged 43.7 percent more for overdraft fees than did small banks; in 2002, the comparable figure dropped to 34.2 percent. During the same period, fees for stop-payment orders at small banks went up 24 percent; at large banks, 15 percent; in 1999, large banks charged 49.3 percent more for stop-payment orders than did small banks; in 2002, the comparable figure dropped to 38.4 percent. There is a similar pattern for insufficient-funds charges. Between 1999 and 2002, monthly NOW account fees for single-fee accounts with a low balance dropped 3 percent at large banks while rising 3.4 percent at small banks. It should be noted that for some fees, such as monthly passbook account fees, large banks have increased charges more quickly than small banks, and that the dollar amount differences between all fee levels at large and small banks have not changed substantially since 1999 (Federal Reserve Board [2001 and 2003], appendix B).

⁸⁰ Federal Reserve System (2002), 3, 15.

⁸¹ See DeYoung and Duffy (2002); DeYoung and Hunter (2003).

⁸² See DeYoung, Hunter, and Udell (2003).

⁸³ Keeton, Harvey, and Willis (2003), 38.

⁸⁴ See Golter and Solt (2004).

⁸⁵ See table A.17.

⁸⁶ At least to an extent, these problems can be mitigated through user groups. See Golter and Solt (2004).

⁸⁷ De Young and Hunter (2003), 196-97.

⁸⁸ For a review of the issue of the cost of bank regulation, see Elliehausen (1998).

REFERENCES

Avery, Robert B., and Katherine A. Samolyk. 2004. Bank Consolidation and Small Business Lending: The Role of Community Banks. Journal of Financial Services Research 25, nos. 2-3:291-325.

Bassett, William F., and Thomas F. Brady. 2001. The Economic Performance of Small Banks, 1985-2000. Federal Reserve Bulletin (November): 719-28.

____________. 2002. What Drives the Persistent Competitiveness of Small Banks? Finance and Economics Discussion Working Paper Series, no. 2002-28. Federal Reserve Board.

Berger, Allen N., Seth D. Bonime, Lawrence G. Goldberg, and Lawrence J. White. 1999. The Dynamics of Market Entry: The Effects of Mergers and Acquisitions on De Novo Entry and Small Business Lending in the Banking Industry. Finance and Economics Discussion Working Paper Series, no. 1999-41. Federal Reserve Board.

Berger, Allen N., Nathan H. Miller, Mitchell A. Petersen, Raghuram G. Rajan, and Jeremy C. Stein. 2002. Does Form Follow Function? Evidence from the Lending Practices of Large and Small Banks. NBER Working Paper Series, Working Paper 8752. National Bureau of Economic Research.

Berger, Allen N., Richard J. Rosen, and Gregory F. Udell. 2001. The Effect of Market Size Structure on Competition: The Case of Small Business Lending. FRB Chicago Working Paper No. 2001-10. Federal Reserve Bank of Chicago.

Berger, Allen N., and G. F. Udell. 2002. Small Business Credit Availability and Relationship Lending: The Importance of Bank Organizational Structure. Economic Journal 112:32-53.

Berger, Allen N., and Gregory F. Udell. 2003. The Future of Relationship Lending. In The Future of Banking, edited by Benton E. Gup, 203-27. Quorum Books.

Brickley, James A., Clifford W. Smith, Jr., and James S. Linck. 2001. Boundaries of the Firm: Evidence from the Banking Industry. Financial Research and Policy Working Paper No. FR 00-01. Bradley Policy Research Center.

Carter, David A., James E. McNulty, and James A. Verbrugge. 2004. Do Small Banks Have an Advantage in Lending? An Examination of Risk-Adjusted Yields on Business Loans at Large and Small Banks. Journal of Financial Services Research 25, nos. 2-3:233-52.

Cobas, Maria Gloria, Larry R. Mote, and James A. Wilcox. 2003. A History of the Future of Banking: Predictions and Outcomes. In The Future of Banking, edited by Benton E. Gup, 49-76. Quorum Books.

Cocheo, Steve. 1997. Community Bank Competitiveness Survey: Community Banks Face Their Future. ABA Banking Journal 89, no. 4:55-59, 77 and 79.

____________. 1998. Second Annual Community Bank Competitiveness Survey: "Adaptability" is Community Banking's Middle Name. ABA Banking Journal 90, no. 2:47-64.

____________. 1999. Third Annual Community Bank Competitiveness Survey: Puzzling Out a Survival Plan. Supplement. ABA Banking Journal 91, no. 2.

____________. 2000. Fourth Annual Community Bank Competitiveness Survey. Supplement. ABA Banking Journal 92, no. 2.

____________. 2002. Sixth Annual Community Bank Competitiveness Survey: Getting Down to Business. Supplement. ABA Banking Journal 94, no. 2.

____________. 2003. Seventh Annual Community Bank Competiveness Survey: Regaining Your Perspective. Supplement. ABA Banking Journal 95, no. 2.

Conference of State Bank Supervisors. 1978. A Profile of State-Chartered Banking. Conference of State Bank Supervisors.

De Senerpont Domis, Olaf. 1996. Community Banks Are Here to Stay, Greenspan Declares. American Banker (March 6).

DeYoung, Robert. 2000. How Long Are Newly Chartered Banks Financially Fragile? FRB Chicago Working Paper 2000-09. Federal Reserve Bank of Chicago.

DeYoung, Robert, and Denise Duffy. 2002. The Challenges Facing Community Banks: In Their Own Words. Federal Reserve Bank of Chicago Economic Perspectives 26, no. 4:2-17.

DeYoung, Robert, and William C. Hunter. 2003. Deregulation, the Internet, and the Competitive Viability of Large Banks and Community Banks. In The Future of Banking, edited by Benton E. Gup, 173-201. Quorum Books.

DeYoung, Robert, William C. Hunter, and Gregory F. Udell. 2003. The Past, Present, and Probable Future for Community Banks. FRB Chicago Working Paper No. 2003-14. Federal Reserve Bank of Chicago.

Elliehausen, Gregory. 1998. The Cost of Bank Regulation: A Review of the Evidence. Staff Study No. 171. Board of Governors of the Federal Reserve System.

Federal Deposit Insurance Corporation (FDIC). 1997. History of the Eighties—Lessons for the Future. Vol. 1, An Examination of the Banking Crises of the 1980s and Early 1990s. FDIC.

Federal Reserve Board (FRB). 1999, 2001, 2003. Annual Report to the Congress on Retail Fees and Services of Depository Institutions. FRB.

Federal Reserve System. 2002. Community Bank Study. Federal Reserve System.

Grant Thornton. 1994. Survey of Senior Community Bank Executives. Grant Thornton.

____________. 1995. Annual Survey of Senior Community Bank Executives. Grant Thornton.

____________. 1996. Annual Survey of Community Bank Executives. Grant Thornton.

____________. 1997. Fourth Annual Survey of Community Bank Executives. Grant Thornton.

____________. 1998. Fifth Annual Survey of Community Bank Executives. Grant Thornton.

____________. 1999. Sixth Annual Survey of Community Bank Executives. Grant Thornton.

____________. 2000. Seventh Annual Survey of Community Bank Executives. Grant Thornton.

____________. 2001. Eighth Annual Survey of Community Bank Executives. Grant Thornton.

____________. 2002. Ninth Annual Survey of Community Bank Executives. Grant Thornton.

____________. 2003. Tenth Annual Survey of Community Bank Executives. Grant Thornton.

Golter, Jay, and Martha Solt. 2004. Bank Technology: Past Forces, Current Conditions, and Future Trends. FDIC Staff Paper.

Gunther, Jeffery W., and Robert R. Moore. 2004. Small Banks' Competitors Loom Large. Federal Reserve Bank of Dallas Southwest Economy (January/February): 9-13.

Isaac, William. 1991. Is There a Future for Midsize Banks? American Banker (September 19).

Keeton, William R. 2000. Are Mergers Responsible for the Surge in New Bank Charters? Federal Reserve Bank of Kansas City Economic Review 85, no. 1:21-41.

Keeton, William, James Harvey, and Paul Willis. 2003. The Role of Community Banks in the U.S. Economy. Federal Reserve Bank of Kansas City Economic Review (Q2):15-43.

Kimmelman, John. 1999. Small Banks Neglecting an Edge: Lower Fees. American Banker (November 2).

Kline, Alan. 1997. Slide in Small-Bank Population Seen Leveling Off. American Banker (September 19).

Myers, Forest, and Kenneth Spong. 2003. Community Bank Performance in Slower Growing Markets: Finding Sound Strategies for Success. Federal Reserve Bank of Kansas City Financial Industry Perspectives, 15-30.

Seelig, Steven A., and Tim Critchfield. 2003. Merger Activity as a Determinant of De Novo Entry into Urban Banking Markets. FDIC Working Paper 2003-01. FDIC.

Spiegel, John, Alan Gart, and Steven Gart. 1996. Banking Redefined: How Superregional Powerhouses Are Reshaping Financial Services. Irwin Professional Publishing.

Stein, Jeremy C. 2002. Information Production and Capital Allocation: Decentralized versus Hierarchical Firms. Journal of Finance 57, no. 5:1891-1921.

Stiroh, Kevin. H. 2004. Do Community Banks Benefit from Diversification? Journal of Financial Services Research 25, nos. 2-3:135-60.

U.S. General Accounting Office. 2003. Credit Unions: Financial Condition Has Improved, but Opportunities Exist to Enhance Oversight and Share Insurance Management. GAO-04-91.

Walser, Jeffrey W., and John M. Anderlik. 2004. Rural Depopulation: What Does It Mean for the Future Economic Health of Rural Areas and the Community Banks That Support Them? FDIC Banking Review 16, no. 3:57-95.

Whalen, Gary. 2001. The Impact of the Growth of Large, Multistate Banking Organizations on Community Bank Profitability. Economic and Policy Analysis Working Paper 2001-5. Office of the Comptroller of the Currency.

Whitener, Leslie A., and David A. McGranahan. 2003. Rural America: Opportunities and Challenges. Amber Waves 1, no 1: 15-21.

Yom, Chiwon. 2003. Severity of Young Banks' Exposure to Real Estate Markets. FDIC Staff Paper.

Appendix

Figures A.1–A.5 present aggregate performance measures for community banks classified by the population growth of the MSA or non-MSA county where each bank was headquartered. High-population-growth markets include all MSAs and non-MSA counties that experienced an average annual growth rate of more than 2 percent during the 1985–2003 period. Medium-population-growth markets include all MSAs and non-MSA counties that experienced an average annual growth rate between zero and 2 percent during the 1985–2003 period. Low-population-growth markets include MSAs and non-MSA counties that experienced negative population growth over the 1985–2003 period. Community banks are defined as banking organizations (bank and thrift holding companies, independent banks, and independent thrifts) that controlled bank assets or thrift assets of less than $1 billion (in 2002 dollars).

D

D

D

D

D

Table A.1 - FDIC-Insured Community Banks Unit-Bank States vs. Non-Unit-Bank States, 1985 and 2003

Table A.2 - FDIC-Insured Small Community Banks Unit-Bank States vs. Non-Unit-Bank States, 1985 and 2003

Table A.3 - Number of FDIC-Insured Community Banks by Type of Geographic Area, 1985 and 2003

Table A.4 - Number of FDIC-Insured Small Community Banks by Type of Geographic Area, 1985 and 2003

Table A.5 - Number of FDIC-Insured Community Banks in Growing Markets, 1985 and 2003

Table A.6 - Number of FDIC-Insured Community Banks in Declining Markets, 1985 and 2003

Table A.7 - Changes in the Distribution of Domestic Deposits by Type of Geographic Area, Declining Markets, and Growing Markets, 1985-2003

Table A.8 - Share of Banking-Sector Assets and Funding, Commercial Banks vs. Savings Institutions 1985, 1994, and 2003

Table A.9 - Banking-Sector Balance-Sheet Ratios, Assets, and Liabilities, Commercial Banks vs. Savings Institutions

Table A.10 - Commercial Bank Performance Ratios, 1985-2003

Table A.11 - Savings Institution Performance Ratios, 1985-2003

Table A.12 - Performance Measures for Community Banks Based on Their Subsequent Size-Group Classification

Table A.13 - Community Bank Performance and Local Growth by Type of Geographic Area, 1985-2003

Table A.14 - Mean Share of Deposits and Offices by Type of Geographic Area, Banks vs. Credit Unions by Size Category, June 1994

Table A.15 - Mean Share of Deposits and Offices by Type of Geographic Area, Banks vs. Credit Unions by Size Category, June 2003

Table A.16 - Federally Insured Credit Union Summary Data by Type of Membership: Year-end 1997-2003 (Dollars in Thousands)

Table A.17 - Percentage of Community Banks Offering Various Internet Services in 2002 and Expectations for those Services Three Years Later