Content uploaded by Arthur B. Kennickell
Author content
All content in this area was uploaded by Arthur B. Kennickell on Apr 08, 2016
Content may be subject to copyright.
Finance and Economics Discussion Series
Divisions of Research & Statistics and Monetary Affairs
Federal Reserve Board, Washington, D.C.
Banking Market Definition: Evidence from the Survey of
Consumer Finances
Dean F. Amel, Arthur B. Kennickell, and Kevin B. Moore
2008-35
NOTE: Staff working papers in the Finance and Economics Discussion Series (FEDS) are preliminary
materials circulated to stimulate discussion and critical comment. The analysis and conclusions set forth
are those of the authors and do not indicate concurrence by other members of the research staff or the
Board of Governors. References in publications to the Finance and Economics Discussion Series (other than
acknowledgement) should be cleared with the author(s) to protect the tentative character of these papers.
Banking Market Definition: Evidence from the Survey of Consumer Finances
Dean F. Amel, Arthur B. Kennickell and Kevin B. Moore*
July 7, 2008
Abstract
This paper uses data from the triennial waves of the Survey of Consumer Finances from
1992 to 2004 toexamine changes in the use of financial services with implications for the
definition of banking markets. Despite powerful technological and regulatory shifts over
this period, households' banking markets overall remained largely local--the median
distance to a provider of financial services remained under four miles.However, there
has been rapid growth in the use of non-depository financial institutions over the period,
particularly non-local ones. This increase occurred across a wide variety of demographic
and other household classifications. The evidence on the clustering of financial services is
mixed. Households showed a slightly greater tendency to buy multiple banking services
from their primary provider of such services in 2004 thanin 1992, while they also became
much more likely to procure services from firms that were not their primary provider.
* Division of Research and Statistics, Board of Governors of the Federal Reserve System, Washington,
D.C. 20551. The views expressed here are those of the authors and do not necessarily reflect those of the
Board of Governors or its staff. We thank Bob Adams, Lamont Black, Ken Brevoort, Beth Kiser, and
Robin Prager for helpful comments. Any remaining errors are the responsibility of the authors.
1
1. Introduction
Any analysis of competition under U.S. antitrust statutes begins with the
definition of the relevant market in both geographic and product space. The established
legal standard for analysis of competition in the U.S. banking industry in recent decades
has been that banking markets are geographically local – encompassing an area roughly
equivalent to a metropolitan area or one or two rural counties – and that the sole product
market is the cluster of financial products and services typically supplied by commercial
banks. This paper examines recent evidence from the Survey of Consumer Finances on
these two aspects of banking market definition in the context of earlier results, taking into
account measurement issues introduced by changes in market structure over time.
The use of local geographic banking markets has generated some controversy in
recent years. The expansion of bank branch networks (due partly to relaxation of legal
constraints on the geographic expansion of banks), the increased centralization of deposit
and loan rate setting by some banking organizations, the emergence of nationwide
networks of automated teller machines, and the growth of Internet banking have led some
to argue that banking markets are now statewide or larger in scope. Geographic
expansion by financial institutions also blurs the precision of measures of the distinction
between local and non-local institutions; for example, if a customer had an account at a
non-local institution that subsequently opened a branch in the customer’s local market,
the account could be inferred to be a local one simply as a result of a change in market
structure independent of the account owner. Similarly, if someone mainly accesses the
services of an institution remotely via mail, telephone, fax or the Internet, but the
institution has a local physical presence that is used periodically, the institution might
reasonably be classified as either local or non-local; even if nearly all use were remote,
the possibility of using a local office for problem resolution might still argue for treating
the institution as a local one.
The use of the cluster of banking services as a product market has generated less
commentary, despite evidence that bank customers increasingly utilize multiple providers
of financial services, including non-bank firms, for many of their financial needs. The
use of the cluster is convenient both for bank regulators and potential bank acquirers,
2
because it simplifies antitrust analyses that could become quite costly and lengthy if
every bank product and service were considered to be in a separate market.
Since the passage of the Gramm-Leach-Bliley Act (or Financial Services
Modernization Act) of 1999, however, the appropriate definition and measurement of the
cluster have become somewhat more complicated. In response to this law, many
financial institutions moved to provide a broader array of services to consumers, thus
enlarging the possibility of clustering of services. At the same time, it has become more
common for large financial services companies to consist of many sub-units that may
separately provide traditional deposit services; specialized loan services, such as
mortgages, lines of credit, credit cards, vehicle-related and other installment loans,
higher-risk short term loans, etc.; brokerage and trust services; and a variety of other
types of investment services. Some such sub-units may have their own common brand
names distinct from that of the parent company. As a result, it is not clear that customers
always recognize the extent to which they may be buying services from different arms of
one company. To capture clustering in the presence of such differentiation requires a
broader definition of the financial entity examined.1
This paper reviews trends in the use of financial services by U.S. households,
using data from the triennial waves of the Survey of Consumer Finances (SCF) from
1992 to 2004. The core of the work presented here follows analysis by Amel and Starr-
McCluer (2002) and Kwast, Starr-McCluer and Wolken (1997), utilizing data from two
surveys not available when the previous papers were written. In addition, the paper
analyzes differences among households with different demographic and financial
characteristics in the geographic patterns of their procurement of financial services.
Finally, the paper relates geographic patterns to the structure of local banking markets to
determine if household behavior is influenced by the local competitive environment in
the banking industry.
The following section reviews the legal basis for current banking market
definitions as well as previous research on such definitions. Section 3 describes the
relevant portions of the SCF. Section 4 reviews evidence on the geographic dispersion of
1 As an illustration of the extent to which large financial institutions with many sub-units have expanded in
recent years, the percentage of the banks and thrift institutions operating a branch in a market but
headquartered in a different state has increased from 1 percent in 1994 to 18 percent in 2004.
3
suppliers of household financial services. Section 5 examines at the extent to which
purchases are clustered at financial institutions. Section 6 analyzes the effects of
demographic variables and market structure on the geographic characteristics of the
demand for banking services. The paper concludes with a summary and discussion of
policy implications.
2. Legal and Economic Bases for Banking Market Definitions
The legal standard for defining a market for purposes of antitrust analysis has
developed over the decades since passage of the Sherman Act in 1890. According to the
current Horizontal Merger Guidelines of the Department of Justice and Federal Trade
Commission, a market is a “product or group of products and a geographic area such that
a hypothetical profit-maximizing firm, not subject to price regulation, that was the only
present and future producer or seller of those products in that area likely would impose at
least a ‘small but significant and nontransitory’ increase in price, assuming the terms of
sale of all other products are held constant.”2
The applicability of antitrust standards to banking was not clear until 1963, when
the Supreme Court held that the antitrust laws, and in particular section 7 of the Clayton
Act (1914), applied to banking.3 The Court's opinion regarding the applicability of the
Clayton Act to banking was reaffirmed by the Congress in 1966, when it amended both
the Bank Merger Act (1960) and the Bank Holding Company Act (1956), and in 1978,
when it passed the Change in Bank Control Act.
In addition to clarifying the applicability of the antitrust laws to banking, the
Philadelphia National Bank case laid the basis for the geographic and product market
definitions still used in banking today. The Court ruled that, because important classes of
bank customers are locally limited, the local geographic area is the relevant geographic
market for an analysis of competition in banking.4 In addition, the Court found “that the
cluster of products (various kinds of credit) and services (such as checking accounts and
trust administration) denoted by the term ‘commercial banking’ composes a distinct line
2 U.S. Department of Justice and the Federal Trade Commission, Horizontal Merger Guidelines,
www.usdoj.gov/atr/public/guidelines/horiz_book/hmg1.html. The current guidelines were first
published in 1997.
3 United States v. Philadelphia National Bank, 374 U.S. 321, 372 (1963).
4
of commerce.”5 The rationale given for the cluster was that some commercial banking
products or services are so distinctive that they are entirely free of effective competition
from products or services of other financial institutions, while others enjoy either cost
advantages or settled consumer preferences that insulate them within a broad range from
substitutes furnished by other institutions.
In the years since Philadelphia National Bank, the courts have indicated a
willingness to consider the argument that, as a result of changes in the financial sector,
commercial banking services as a whole are not a single product line and that competition
may extend beyond local markets. However, the courts have consistently held that
arguments for changing the basic market definition in banking must be based on
persuasive economic evidence.
Empirical evidence on the product cluster of financial services comes from
previous research on the SCF and from the Surveys of Small Business Finances (SSBF)
conducted in 1993, 1998, and 2003. Data from the SSBF indicate that small businesses in
the United States obtain, on average, two financial services from their primary financial
institution and from local depository institutions in general, most of which are
commercial banks.6 In contrast, small businesses typically obtain only one financial
service from non-depository and non-local suppliers. Analysis using the waves of the
SCF from 1989 to 1998 found some weakening over time in the extent to which
households cluster their purchases of financial services.7 While the average number of
financial services purchased from a household’s primary institution had increased over
time, the percentage of a household’s total services obtained from the primary institution
had declined.
Empirical evidence on the geographic scope of banking markets suggests that
banking markets are still local for most households and small businesses. The SSBFs
4 Ibid., p. 360.
5 Ibid., p. 356.
6 Previous research on the SSBF can be found in Myron L. Kwast, Martha Starr-McCluer and John D.
Wolken, “Market Definition and the Analysis of Antitrust in Banking,” The Antitrust Bulletin, 44 (1997),
pp. 973-995, and Marianne P. Bitler, Alicia M. Robb and John D. Wolken, “Financial Services Used by
Small Businesses: Evidence from the 1998 Survey of Small Business Finances,” Federal Reserve Bulletin
87 (2001), pp 183-205. Unpublished results from the 2003 survey are consistent with those from previous
surveys.
7 See Dean F. Amel and Martha Starr-McCluer, “Market Definition in Banking: Recent Evidence,” The
Antitrust Bulletin 47 (2002), pp. 63-89.
5
from 1993 to 2003 indicate that the majority of small businesses obtain most of their
financing from local financial institutions, primarily commercial banks. For example, as
of year-end 1998, 82 percent of all financial services, 94 percent of checking and savings
services, and more than 75 percent of financial management services used by small
businesses were obtained from local sources.8 In addition, a survey conducted for the
National Federation of Independent Businesses finds that small businesses perceive their
banking markets to be very locally limited.9
Past research on the SCF has found that households rely almost exclusively on
local financial institutions for products and services such as transactions accounts,
certificates of deposit and lines of credit, but look increasingly to non-local providers for
other services. Other empirical studies that find differences in deposit and loan rates
across metropolitan areas and rural counties also suggest that banking markets are local.
Research conducted in the early 1990s, confirmed by more recent empirical work, finds
significantly higher loan interest rates and significantly lower deposit interest rates in
more concentrated local banking markets.10
3. The Survey of Consumer Finances
This paper uses data from the 1992, 1995, 1998, 2001, and 2004 SCFs. The
survey gathers a great deal of information on the financial services used by households.
This paper focuses on services typically supplied by commercial banks, including
checking accounts, savings accounts, money market deposit accounts, money market
mutual funds, certificates of deposit, IRA and Keogh accounts, brokerage accounts, trust
and other managed assets, home mortgages, motor vehicle loans, home equity and other
lines of credit, and other consumer loans.
The SCF is conducted every three years under the sponsorship of the Federal
Reserve Board and in cooperation with the Statistics of Income Division of the Internal
8 For 1993 small business results, see Kwast, Starr-McCluer and Wolken, op. cit. The statistics for 1998
and 2003 are based on unpublished tabulations from the relevant business surveys provided by John
Wolken.
9 Dean F. Amel and Kenneth P. Brevoort, “The Perceived Size of Small Business Banking Markets,”
Journal of Competition Law and Economics 1 (2005), pp. 771-784.
10 See Erik A. Heitfield, “What Do Interest Rates Say about the Geography of Retail Banking Markets?”
The Antitrust Bulletin 44 (1999), pp. 333-47; and Erik Heitfield and Robin A. Prager, “The Geographic
Scope of Retail Banking Markets,” Journal of Financial Services Research 25 (2004), pp. 37-55.
6
Revenue Service. The survey collects detailed information on households’ assets,
liabilities, use of financial services and other financial characteristics. Since 1989, the
survey has included questions about the types of financial institutions used and their
distance from the home or office of the household member who uses them most often.
Information is collected on up to six (seven in 2004) different financial
institutions used by the household, along with information linking the financial services
used by households to those institutions.11 However, the type of an institution may be
difficult for some respondents to identify, despite provision of definitions by
interviewers. Some respondents may be unable to distinguish commercial banks and
savings banks, for example, and, as discussed earlier in this paper, the proliferation of
entities within broad financial services companies inherently complicates even the
conceptual problem of identifying the type of an institution where a household uses
services that cut across such entities.12 Interviewers are instructed to ask respondents for
the approximate distance of each institution (or its ATM, if that is more commonly used
than an office) from the home or office of the person who uses it most frequently, but in
some cases the distance may be unclear. The household may use a variety of offices, and
some of those may be local offices while others may be more remote; in such cases, the
respondent would be asked for the distance associated with the office used most
frequently. Distance may be particularly unclear when the institution is used only by
telephone or over the Internet; when the respondent can only say that the institution is
used by either of these means or by mail to a remote location, the distance is assumed in
this analysis to be more than 50 miles. Because the SCF offers interviewers a code
“more than 50 miles” to use in cases where the respondent is uncertain of just how far
away an institution at least this distant might be, the data provide comparably useful
11 The institutions are ones reported by the survey respondent as institutions where they have deposits or
other such accounts or where they have loans. In the case of mortgages, the institution may not be the
originating lender, but rather the current servicer of the loan. For first-lien mortgages on a primary
residence, the SCF contains an indication of whether the current servicer is the same as the originating
institution, and from 2001 forward it contains information on the distance to the originating institution.
12 It is also open to question whether a household that only uses services of one subsidiary of a large
financial services company will be more likely to report the type of the subsidiary or the parent firm. In
some cases, respondents may identify the type of the firm with the historical code of the institution—for
example, as might be the case for an insurance company that expanded to provide banking services through
subsidiaries.
7
information only up to this limit. For consistency in analysis in this paper, all distances
reported as a value greater than 50 miles are truncated at 50 miles.
For purposes of this paper, we divide financial institutions into depository and
non-depository institutions, with the former subdivided into commercial banks, thrift
institutions (including savings and loans and savings banks) and credit unions.13 Non-
depository institutions are divided into finance companies, brokerage firms, mortgage
finance companies, other nontrivial institutions, and entities described as an individual or
small set of individuals.14 All financial institutions are also divided into local and non-
local firms, with the boundary between the two set somewhat arbitrarily at 30 miles from
the home or office of the household member most involved in managing the account.15
In addition, financial firms are divided between those identified by the survey respondent
as the household’s primary financial institution and all other providers of financial
services.
4. Geographic Banking Markets
Data from the 2004 SCF show that the distances between households and the
financial institutions at which they get their financial services remain quite short. Table 1
gives a summary of the distributions of distance from a household to its provider for a
number of different financial services for both 1992 and 2004. As noted earlier, the
distance data used in this paper have been truncated above at 50 miles.
The median distance to a provider has increased for all financial products except
checking, savings, and money market accounts and certificates of deposit; because these
are common types of accounts, the median distance for all accounts stayed constant over
the twelve-year period. Among the instances where the 90th percentile of the distance to
a provider was already at 50 miles in 1992, the distribution of distance can be seen to
have shifted further outward for all products except money market accounts.
13 The classification of financial institutions as depository or non-depository suffers from the same
problems as the classification of the overall institution type. A given household may use both depository
and non-depository subsidiaries of an institution. Here institutions classified in the SCF as broad financial
services companies are treated as commercial banks.
14 See the notes to table 4 for a definition of the institution classification used in this paper.
15 Using a 40-mile cut-off to define the local market does not yield any substantive differences.
8
Nonetheless, the median distance remains quite short—under 4 miles—in 2004 for the
deposit accounts and for lines of credit.
Given that nearly half of households in 2004 households used the Internet for
financial services, tools or information (see Bucks et al. [2006]), it might be expected that
this means of access was a driver in increasing the use of non-local institutions. For
households with relationships with non-local financial institutions in 2004, 19 percent
of accounts or loan services were accessed via the Internet, while the comparable figure
for local institutions was 18 percent. Both estimates increased from a level in 1992 that
was so slight or indistinct as to be undetectable in the SCF. Thus, the Internet has become
an important means of accessing financial services regardless of the location of the
institution; at the same time, it may also have facilitated the discovery of non-local
institutions or the opening of non-local accounts.
Although the shares of services used at local institutions remained high in 2004,
the surveys from 1992 to 2004 show a steady decline in the shares of accounts and loans
at local institutions (table 2). Although trends within the period are mixed for individual
types of service, the share of services at local institutions was lower in 2004 than in 1992
in all cases except money market accounts. While the great majority of checking and
savings accounts, certificates of deposit, and money market accounts remain at a local
institution, the shares of vehicle loans and other loans fell below half and other types also
declined substantially.
If the focus is restricted to local depository institutions, the patterns are similar,
but some of the declines from 1992 to 2004 are much sharper—particularly for loans. In
2004, the majority of loans other than lines of credit were obtained from an institution
other than a local depository. For deposits, there were also large declines in the shares of
IRAs and Keogh accounts and trust accounts at local depositories, while brokerage
accounts declined further from a level that was already very low in 1992.
Examination of the dollar shares of holdings of accounts or loan balances
provides an alternative picture of the market share of local institutions. Comparing table
2 to table 3, the overall dollar share of accounts held with local institutions was lower
than the share of accounts held with local institutions in all the years shown. The
difference was most notable for checking, savings, and money market accounts; results
9
for local depositories alone are similar. The dollar share of accounts declined over 1992
to 2004, but the shift was already apparent in 1995; IRA/Keogh accounts, brokerage
accounts, and trusts experienced the largest declines. For loans, shares of accounts and
dollar shares were similar in all years and both measures declined over the period.
Table 4 approaches the local/non-local breakdown from another perspective. For
households that actively use at least one institution for deposit or loan services, the table
shows the share of such households that procure at least one financial product or service
from local or non-local financial institutions of various types. Looking first at all
financial institutions, the table shows that there has been little change in the use of local
institutions, while the use of non-local institutions has increased more than 50 percent: In
2004, 57 percent of households used at least one non-local provider of financial services,
up from 35 percent in 1992.
These overall figures mask substantial differences over time in the use of
depository institutions and non-depositories. Almost all households continue to use a
local depository institution, but the composition shifted from 1992 to 2004. The share of
households using local credit unions rose slightly, but the share using local commercial
banks and thrifts declined more---much more in the case of thrifts, reflecting the decline
in the numbers of thrifts through failure or acquisition by commercial banks. Taken with
the overall steady level of use, these shifts show that fewer households are using multiple
types of depository. At the same time, use of non-local depositories rose overall, while
the use of non-local commercial banks and credit unions rose and use of non-local thrifts
declined as in local markets.
In 1992, local and non-local non-depositories overall were used at virtually the
same rate. By 2004, there had been a 50 percent increase in the use of local non-
depository financial institutions overall, while the use of non-local non-depositories had
more than doubled. For all four categories of non-depository financial institution, it was
more common for households to use a non-local firm than to use a local firm in 2004; in
1992, this was true only for mortgage finance companies. The growth in the use of non-
local finance companies was particularly pronounced over the twelve years analyzed.
5. Banking Product Markets
10
One indication that the tendency of households to cluster the purchases of
financial services at one provider may have weakened over time is the increasing
percentage of households that utilize non-depository institutions. Table 5 shows that
from 1992 to 2004 almost all households continued to have a relationship with at least
one depository institution. The decline in the use of thrift institutions over time indicates
either that households are reducing the number of depository institutions with which they
have accounts, or they have switched their thrift institution relationships to relationships
with multiple commercial banks or multiple credit unions.
Over the same period, there was a near doubling in the percent of households with
accounts at one or more non-depository institutions, and there were substantial increases
for all four categories of such firms. The percentage of households with an account at a
finance company tripled from 1992 to 2004, the percentage with some sort of account
with a brokerage increased by three-quarters, the percentage with a relationship with a
mortgage finance firm almost doubled, and the use of other non-depository institutions
more than tripled. This suggests an increased willingness to look beyond depository
institutions for at least some financial products and services.
Clustering of the use of financial services might most readily be examined by
looking at the services selected at the institution deemed by the respondents to be the one
where the household does the most business.16 In 2004, 87 percent of households with
any accounts reported that the “primary” institution was one where they had their main
checking account; the figure in 1992 was 84 percent. About a third of those not reporting
such a connection in 2004 did not have a checking account.
16 In the initial enumeration of financial institutions in the SCF, the respondent is asked to begin with “the
one where you do the most business”. The phrase “the most business” is not defined for the respondent, so
the basis of this classification may vary across cases; some may consider the intensity of business in a
single account, others may consider the number of distinct number of services used, the amount of money
deposited or owed, or other distinctions.
11
Table 6 shows that the use of non-depository institutions as non-primary
providers of financial services nearly doubled from 1992 to 2004. By construction, all
households in this analysis have a primary institution. While 95 percent of all households
had a depository institution as their primary provider in 2004, this was a slight decline
from the level in 1992. Somewhat surprisingly, given the drop in use of thrift institutions
found in previous tables, commercial banks lost the most primary relationships of any
category of firm, while thrifts suffered more modest (though proportionately greater)
losses and credit unions increased their share of primary relationships by 50 percent.
Among non-depository institutions, brokerages nearly doubled their share of primary
relationships from 1992 to 2004, but still held only 2 percent of all primary relationships
in 2004.
The percent of households with any non-primary relationships rose overall, but
the percent of such households with secondary relationships with a depository institution
dropped about 10 percent from 1992 to 2004, with only commercial banks maintaining
about the same percentage of households having such non-primary accounts. All
categories of non-depository institution saw large increases in the share of households
with which they had a non-primary relationship. Finance companies more than tripled
their penetration in this category to 23 percent of all households, mortgage finance
companies doubled the share of households with which they dealt to 23 percent, and
brokerage firms increased their penetration by 80 percent, to 30 percent of all households.
The percentage of households with such a relationship with other non-depositories more
than tripled to 11 percent.
Table 7 looks at primary financial relationships by type of account rather than by
type of institution. While the overall percentage of services acquired from a primary
institution first moved up then turned down in a range of 53 to 56 percent over the five
SCFs from 1992 to 2004, this relatively small change masks variation in patterns across
product lines. Primary institutions gained shares of all savings accounts, money market
accounts and lines of credit. The share of all checking accounts held by primary
institutions held fairly steady---at about 72 percent---as did the shares of certificates of
deposit, brokerage accounts and trust accounts. Primary institutions lost considerable
12
shares of the markets for IRAs and Keogh accounts, mortgages, vehicle loans and other
loans.
The memo lines at the bottom of table 7 show a decline in the plurality of
households that get only one service from their primary institution, from 44 percent in
1992 to 39 percent in 2004. About 30 percent of households relied on the primary
institution for two services, a 10 percent increase over twelve years, and households
using four or more services increased from 13 to 15 percent. In contrast to the
implications of table 6, these data suggest a modest increase in the popularity of
clustering services in recent years.
6. The Effects of Demographic Factors and Market Structure on Market Size
The data show clearly that over the 1992-2004 period, people in general became
more likely to use financial services at institutions outside of their local market area, but
are there types of households that are more or less intensive users of such services?
Table 8 addresses this question by presenting evidence on the use of non-local
institutions for accounts or loans, broken down by demographic, geographic and market
characteristics. The most striking finding is the increase across all the groups in the
likelihood of using a non-local institution at all or for account or loan services.17
Moreover, among the groups shown, the shift is small only for the group with less than a
high-school education and only in the case of account services. Overall, proportional
increases were greatest among two groups with relatively low levels of use in 1992 and
2004: account-holding households headed by persons 65 or older and those in the lowest
40 percent of the income distribution.
The use of non-local providers for both deposit accounts and loans is substantially
higher among age groups younger than 65 years than for the two older cohorts. This
suggests that, over time, the use of non-local institutions will continue to grow relative to
the use of local firms. In time, this could weaken the case for local geographic markets in
banking.
17 The increase in the use of non-local institutions is largely driven by an increase in the use of non-local
non-depositories. Changes in the use of non-local depositories over 1992-2004 were minimal for all
groups.
13
Substantial differences remained in the levels of use across other groups as well.
Households in higher net worth or education groups, white non-Hispanics and
homeowners had higher rates of use than their complements in both 1992 and 2004.
Such households tend to use more services, and thus have more chances to choose a non-
local institution.
The last factor in the table, the Herfindahl index for commercial banks, thrifts and
credit unions in the MSA or other geographic area where a household is located, is a key
factor in defining the competitiveness of local banking markets. Markets with a value of
the index of greater than 1800 are generally considered to be highly concentrated
markets. In 1992, there was little difference between households in highly concentrated
markets and households in less concentrated markets in the use of non-local institutions
overall and separately for accounts and loans. However, between 1992 and 2004 growth
in this use was much stronger among households in the more competitively structured
markets. Thus, both types of market became effectively more competitive, but the most
competitive ones became relatively more competitive. These results give no indication
that households in more concentrated markets look outside the local market for services.
Future research will address the degree to which this differential change reflects
differences in the penetration of electronic banking.
To get an indication of the degree of independence of the factors in the table in
explaining the use of non-local services, probit models were run to estimate the
propensity to have a non-local provider overall and for either account or loan services.
These models also controlled for key aspects of the survey sample design and for the
number of services used. The regression results are similar to the findings in table 8 and
are not reported for the sake of brevity.
7. Summary and Policy Implications
In recent years, there have been changes in technology and regulation with the
potential to reshape the definitions of geographic and product markets for banking
services. The availability of financial services firms via the Internet opens a national
market to consumers who are willing to use this medium. Changes in the scope and
structure of financial service firms have potentially more complicated effects. Many
14
firms have taken advantage of the demise of restrictions on interstate banking to open
offices across many states or across the country; this factor raises the question of whether
branches of such firms are “local” in a meaningful sense. Additionally, with the passage
of the Gramm-Leach-Bliley Act in 1999, many banks have moved to the provision of a
broader spectrum of financial services, increasing the possibility that consumers might
choose to cluster their financial services with a smaller set of institutions. Moreover, this
possibility might be reinforced by firms’ permitted use of consumers’ confidential
financial information within the corporate family to tailor the marketing of services. At
the same time, the growth of specialized financial institutions could reduce the tendency
of households to cluster their purchases of financial products from one provider. In this
paper, we use the Survey of Consumer Finances from 1992 to 2004 to examine the net
effect of interaction of these supply-side changes with consumers’ observed behavior.
Over the 1992 to 2004 period, the median distance to a provider of financial
services remained under four miles. This stability largely reflects the continued use of at
least nominally local institutions for checking, savings and money market accounts and
certificates of deposit. But markets for many other banking services appear to have
become notably less local over this time. Additionally, there has been rapid growth in the
use of non-depository financial institutions over the period. This shift is true for local
non-depositories, but it is especially marked for non-local non-depository institutions.
The evidence on the clustering of financial services is mixed. Households
showed a slightly greater tendency to buy multiple banking services from their primary
provider of such services in 2004 than in 1992. However, over the same period they
became much more likely to procure services from firms that were not their primary
provider. Underlying these changes was an increase in the number of products used,
which allowed for both increases.
The increase in the use of non-local financial firms occurred across a wide variety
of demographic and other household classifications. While, as expected, the young,
wealthy and highly educated became relatively more likely to use non-local providers of
services than others, all ages, income levels and education levels increased their use of
non-local providers over 1992-2004, often by quite dramatic amounts. The general trend
toward the use of non-local firms was also visible in both urban and rural markets and in
15
both highly concentrated and less concentrated banking markets, as traditionally
measured. However, the data do not suggest that households in highly concentrated
banking markets were relatively more likely to turn to non-local providers of financial
services.
The rapid increase in the scale and scope of the use of the Internet and other
nontraditional methods of providing services will make continued survey research on
banking markets crucial to financial regulators. In addition, the 2007 SCF contains
questions designed to determine whether the provider of a financial service is based
locally or headquartered in a different state than the household. While allowing a more
complete examination of issues surrounding the local presence of providers of financial
services, these questions also will allow analysis of the potential segmentation of
financial markets between local and multi-state organizations, as suggested by recent
work by Adams, Brevoort and Kiser (2007).
16
Bibliography
Adams, Robert, Kenneth Brevoort and Elizabeth Kiser (2007). ''Who Competes with
Whom? The Case of Depository Institutions,” Journal of Industrial Economics, 55,
pp. 141-167.
Amel, Dean F. and Kenneth P. Brevoort (2005), “The Perceived Size of Small Business
Banking Markets,” Journal of Competition Law and Economics 1, pp. 771-84.
Amel, Dean F. and Martha Starr-McCluer (2002), “Market Definition in Banking: Recent
Evidence,” The Antitrust Bulletin 47, pp. 63-89.
Bitler, Marianne P., Alicia M. Robb and John D. Wolken (2001), “Financial Services
Used by Small Businesses: Evidence from the 1998 Survey of Small Business Finances,”
Federal Reserve Bulletin 87, pp. 183-205.
Bucks, Brian K., Arthur B. Kennickell and Kevin B. Moore (2006), “Recent Changes in
U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances,”
Federal Reserve Bulletin, http://www.federalreserve.gov/pubs/bulletin/2006/06index.htm.
Heitfield, Erik A. (1999), “What Do Interest Rates Say about the Geography of Retail
Banking Markets?” The Antitrust Bulletin 44, pp. 333-47.
Heitfield, Erik A. and Robin A. Prager (2004), “The Geographic Scope of Retail Banking
Markets,” Journal of Financial Services Research 25, pp. 37-55
Kwast, Myron L., Martha Starr-McCluer and John D. Wolken (1997), “Market Definition
and Analysis of Antitrust in Banking,” The Antitrust Bulletin 44, pp. 973-995.
U.S. Department of Justice and the Federal Trade Commission, Horizontal Merger
Guidelines, www.usdoj.gov/atr/public/guidelines/horiz_book/hmg1.html.
16
Table 1
Distance from institutions used, by type of service
1992 2004
Distance in miles, by percentile Distance in miles, by percentile
25th Median 75th 90th 25th Median 75th 90th
All accounts 0 3 8 50 1 3 15 50
Checking 0 2 5 12 0 2 5 18
Savings 0 3 8 30 0 3 10 50
Money market 0 3 10 50 0 3 10 50
CDs 0 3 6 20 0 2 9 50
IRA/Keogh 2 5 25 50 3 15 50 50
Brokerage 4 10 50 50 5 25 50 50
Trust 3 15 50 50 4 30 50 50
All loans 2 7 50 50 3 22 50 50
Mortgages 3 9 50 50 3 25 50 50
Vehicles 2 7 22 50 4 40 50 50
Lines of credit 0 3 10 50 1 4 18 50
Other loans 2 10 50 50 5 50 50 50
17
Table 2
Shares of services acquired from local institutions and local depositories, by type of service, 1992-2004
Share of services acquired from:
Local institution Local depository institution
1992 1995 1998 2001 2004 1992 1995 1998 2001 2004
All accounts 88.7 86.3 85.3 84.3 82.5 80.4 77.3 75.7 74.0 72.0
Checking 95.1 94.6 94.2 94.3 93.9 94.4 93.8 93.3 93.2 93.1
Savings 90.0 89.2 90. 91.7 86.4 88.5 88.0 89.7 90.9 81.8
Money market 85.1 82.2 79.5 81.4 86.3 72.1 66.4 63.5 65.1 76.5
CDs 91.8 90.1 91.1 91.2 87.6 88.9 87.9 88.0 87.0 85.1
IRA/Keogh 77.2 68.2 65.2 61.8 58.6 53.4 41.8 36.8 31.2 26.8
Brokerage 71.6 64.1 59.8 56.6 53.2 8.7 6.3 6.3 7.0 6.0
Trust 59.0 56.2 56.6 51.4 52.0 30.1 21.2 21.1 23.5 17.3
All loans 71.3 61.3 57.8 56.2 52.7 61.1 51.9 43.2 44.9 39.0
Mortgages 66.7 55.9 53.2 52.2 51.9 56.3 48.0 41.7 41.2 39.7
Vehicles 77.9 64.2 61.5 58.0 48.7 69.5 56.2 49.8 49.0 37.4
Lines of credit 86.4 83.1 82.3 84.5 78.2 79.7 77.7 72.8 798 73.0
Other loans 62.0 57.1 50.1 46.1 44.8 46.8 40.6 23.4 25.6 19.3
18
Table 3
Percentages of dollars held in or owed to:
Local institution Local depository institution
1992 1995 1998 2001 2004 1992 1995 1998 2001 2004
All accounts 83.0 74.4 70.6 66.6 68.1 63.9 52.9 42.5 36.1 40.0
Checking 92.7 91.4 85.2 90.3 87.6 88.8 87.9 83.7 82.2 82.3
Savings 90.3 87.0 91.9 90.2 82.3 87.6 82.1 88.4 87.6 70.3
Money market 80.8 74.6 73.3 73.0 78.4 53.6 46.7 45.2 41.4 58.6
CDs 94.2 87.9 89.6 91.4 87.7 91.3 84.7 85.3 83.9 82.8
IRA/Keogh 74.5 63.9 63.8 59.0 58.7 40.0 28.6 21.8 18.7 18.3
Brokerage 72.0 64.6 59.7 57.0 52.9 9.1 6.8 6.7 7.5 6.2
Trust 68.3 59.7 57.1 55.4 54.1 42.2 28.1 24.7 22.2 13.4
All loans 65.5 51.6 48.8 50.6 49.0 65.9 44.5 37.7 39.6 37.5
Mortgages 65.0 50.5 48.0 50.4 49.2 55.5 43.5 37.3 39.4 38.0
Vehicles 73.5 61.5 57.3 52.8 45.9 66.1 56.0 47.8 45.0 35.5
Lines of credit 86.4 83.1 82.3 84.5 78.0 79.6 77.7 72.8 79.8 73.0
Other loans 62.4 54.8 46.9 49.8 47.5 44.5 42.9 26.8 30.4 29.2
19
1. Includes households that used at least one financial service; 86 percent of households in 1992, 87 percent in 1995, 91 percent in 1995, 91 percent in 1998 and
92 percent in 2004.
Classification of institutions:
Commercial banks: taken here to include institutions described specifically commercial banks, broad financial services companies,
various credit card issuers and miscellaneous Internet-based bill-paying services.
Thrift institutions: taken here to include savings and loan institutions and savings banks.
Credit unions: taken here to include only credit unions.
Finance companies: general finance companies, automobile finance companies, other types of store or dealer and collection agencies.
Brokerage firms: taken here to include institutions specifically described as brokerages, insurance companies, money market mutual
funds and private bankers.
Mortgage finance companies: taken here to include institutions specifically described as mortgage banks, mortgage brokers or real
estate investment company.
Other non-depositories: taken here to include all remaining entities that were reported as holding accounts or loans.
Table 4
Percentage of households using local and non-local financial institutions, 1992-2004
1
Percentage of households using:
Local institution Non-local institution
1992 1995 1998 2001 2004 1992 1995 1998 2001 2004
All institutions 98.4 98.4 98.2 98.5 97.9 35.3 44.2 51.3 51.8 56.8
Depositories 97.6 97.7 96.3 97.3 96.5 20.4 23.5 22.2 21.4 23.0
Commercial banks 80.6 80.5 78.0 74.7 76.1 11.5 15.4 13.8 14.4 15.2
Thrift institutions 26.4 19.7 19.2 18.7 17.1 4.3 4.3 4.1 3.2 2.9
Credit unions 26.8 27.4 27.8 30.1 28.3 6.6 6.9 6.7 6.3 7.0
Non-depositories 19.6 22.1 29.6 26.7 30.2 19.9 29.6 39.8 41.3 46.8
Finance companies 4.4 5.0 7.5 5.6 7.5 3.9 9.1 13.3 13.1 17.9
Brokerage firms 11.0 13.2 15.7 16.2 16.3 8.1 10.9 15.8 16.7 18.0
Mortgage finance companies 3.7 3.1 5.5 5.3 6.1 9.2 12.5 16.0 18.5 19.1
Other non-depositories 1.6 2.0 3.9 1.6 4.4 1.6 2.2 4.4 3.4 7.2
20
Table 5
Percentage of households using institutions of different types, 1992-2004
(must have asset or loan associated with institution)
1992 1995 1998 2001 2004
All institutions 100.0 100.0 100.0 100.0 100.0
Depositories 99.0 99.3 98.0 99.0 98.6
Commercial banks 83.4 83.9 81.5 78.7 80.4
Thrift institutions 29.4 23.2 22.2 21.2 19.4
Credit unions 31.8 33.0 32.7 34.7 33.3
Non-depositories 34.6 44.4 57.5 56.3 63.0
Finance companies 7.8 13.4 19.7 17.9 23.9
Brokerage firms 17.2 21.7 28.4 28.5 30.6
Mortgage finance companies 12.8 15.4 21.0 23.2 24.7
Other non-depositories 3.1 4.1 8.0 4.8 11.2
21
Table 6
Percentage of households using primary and non-primary institutions, 1992-2004
Percentage of households using:
Primary institution Non-primary institutions
1992 1995 1998 2001 2004 1992 1995 1998 2001 2004
All institutions 100.0 100.0 100.0 100.0 100.0 71.8 75.1 78.2 76.6 79.7
Depositories 96.5 96.7 94.7 95.8 94.8 63.0 60.7 59.1 57.7 57.7
Commercial banks 71.3 70.8 67.1 65.2 65.9 41.4 43.3 41.3 40.7 41.7
Thrift institutions 13.3 11.3 12.1 12.2 11.2 19.7 14.6 12.7 11.3 10.5
Credit unions 11.8 14.6 15.5 18.3 17.7 22.2 20.8 20.4 19.9 18.6
Non-depositories 3.5 3.3 5.3 4.2 5.2 32.3 42.5 54.6 54.0 60.4
Finance companies 0.8 0.9 1.5 1.0 1.2 7.1 12.7 18.6 17.1 23.0
Brokerage firms 1.0 1.5 2.4 2.4 2.0 16.6 20.8 26.8 27.1 29.5
Mortgage finance companies 1.2 0.8 0.9 0.6 1.5 11.7 14.7 20.2 22.6 23.3
Other non-depositories 0.5 0.2 0.5 0.2 0.5 2.7 3.9 7.5 4.6 10.7
22
Table 7
Percentages of services acquired from primary institutions, 1992-2004
1992 1995 1998 2001 2004
All accounts 53.3 55.1 55.5 56.1 53.7
Checking 71.7 73.0 72.9 73.2 72.3
Savings 51.1 54.4 60.4 64.7 57.4
Money market 49.9 49.1 51.6 52.4 60.1
CDs 45.1 48.4 50.7 53.7 46.0
IRA/Keogh 30.4 27.7 25.9 23.8 20.9
Brokerage 11.6 11.8 12.5 13.2 12.6
Trust 15.3 17.7 17.2 19.4 13.6
All loans 33.6 28.2 25.3 27.2 23.6
Mortgages 31.7 24.0 22.7 24.0 23.5
Vehicles 33.0 28.2 26.9 28.9 19.1
Lines of credit 48.8 50.4 52.7 51.2 53.7
Other loans 27.9 25.5 15.5 18.1 12.6
Memo items:
Only one service used 43.6 43.5 36.3 32.5 38.9
Only two services used 27.4 30.4 30.4 33.5 30.6
Only three services used 16.1 13.6 18.0 16.8 15.8
Four or more services used 13.0 12.5 15.3 17.2 14.7
23
Table 8
Percent of households using at least one non-local institution, by demographic or market characteristic and type of service, 1992 and 2004.
Any non-local
service Any account Any loan
1992 2004 1992 2004 1992 2004
All households 35.3 56.8 19.1 40.9 23.0 43.4
Age
34 or younger 39.5 60.1 16.7 42.4 29.4 50.7
35-44 42.0 64.2 21.6 41.3 29.4 55.7
45-54 40.2 63.3 21.8 47.5 27.7 50.2
55-64 35.9 62.1 24.4 45.3 19.9 45.1
65-74 22.7 42.2 14.0 32.8 11.1 22.5
75 or older 16.0 31.2 14.0 26.5 2.9 11.8
Income percentile group
0-19.9 14.6 28.4 8.7 16.9 8.1 20.2
20-39.9 22.6 43.6 12.4 26.5 12.4 31.9
40-59.9 35.9 60.2 16.4 38.2 25.4 48.2
60-79.9 42.8 67.1 23.2 49.4 27.6 51.5
80-90 48.3 79.8 25.1 63.3 36.6 63.1
90-100 56.3 77.1 35.7 72.1 34.3 57.3
Net worth percentile group
0-24.9 27.4 39.8 10.3 20.1 21.1 35.2
25-49.9 31.9 57.2 14.4 34.4 23.2 48.2
50-74.9 35.7 60.1 18.0 43.7 24.8 48.1
75-89.9 39.1 66.1 25.2 58.3 21.8 42.5
90-100 50.0 67.7 38.2 64.9 23.1 38.2
24
Table 8 (cont)
Percentage of households using at least one non-local institution, by demographic or market characteristic and type of service, 1992 and 2004.
Any non-local
service Any account Any loan
1992 2004 1992 2004 1992 2004
Education
Less than high school 16.9 29.4 8.2 10.8 10.2 24.1
High school diploma 26.9 48.7 12.6 29.6 18.7 38.1
Some college 37.9 61.2 22.1 43.9 23.1 47.9
College degree 49.5 69.4 28.1 57.5 32.5 51.3
Race
White non-Hispanic 37.0 61.0 20.7 45.8 23.7 45.4
Non-white or Hispanic 28.3 44.2 12.7 26.2 19.8 37.5
Homeownership status
Renter 26.7 40.6 16.4 29.4 15.1 29.6
Homeowner 39.2 63.0 20.3 45.3 26.6 48.7
Type of area
Non-MSA 31.0 51.6 19.4 31.3 17.7 39.7
MSA 36.4 57.8 19.0 42.8 24.4 44.2
Local market concentration (HHI)
less than 1800 35.7 58.4 18.9 42.6 23.5 44.6
1800 or more 34.3 47.6 19.5 30.8 21.5 36.7
