1 Public Law 103-328, 108 STAT.2338-2381
(September 29, 1994).
2 In states that permit statewide branching, a
single state or federal bank charter would be all that is needed to
operate within the state. In states that place geographic
restrictions on branching within the state, several separately
chartered banks might be required to operate a larger statewide
3 See Billett, Coburn and O'Keefe (1995) for a
discussion of the FDIC's resolution methods.
4 The number of mergers and consolidations was
obtained from the FDIC's corporate database on the structure of the banking
and thrift industries. Detailed records are maintained for each
merger and consolidation transaction from both the acquirer's and
the target bank's perspective. The total number of merger and
consolidation transactions an acquirer is involved in over a
specific time period is, therefore, equal to the number of
individual banks assumed. For example, the consolidation of four
affiliated banks into an assuming bank's charter would be counted
as four consolidations. Mergers of nonaffiliated banks were
5 See Rose (1987), (1988) for comprehensive studies
of merging banks' characteristics. Rose and others have sought to
identify merger motives and the traits of acquirers and target
banks through financial ratio analysis. Mean financial ratios of
groups of acquirers (targets) are compared to those of peer groups
to identify areas of importance. In addition, statistical logit
analysis is used to identify significant differences between
acquirers and nonacquiring banks. These two techniques are
employed in this study for similar purposes.
6 See Houston and Ryngaert (1994) for a
comprehensive study of the overall gains in large- bank mergers.
Houston and Ryngaert find that for a sample of large-bank mergers
the positive merger-related returns to targets are offset by
negative returns to acquirers. This result does not support the
efficiency gains hypothesis.
7 A proof of this statement can be found in most
textbooks on investment theory. See, for example, Financial Theory
and Corporate Policy, by Thomas E. Copeland and J. Fred Weston,
Addison Wesley Publishing Co., Philippines, (1979): pp. 153-154.
8The DOJ and the FTC first published their merger
guidelines in 1968 and have periodically revised and clarified
those guidelines. See, Horizontal Merger Guidelines,
April 2, 1992, issued by the DOJ and the FTC.
9 See Scherer (1980) for a discussion of merger
motives and the potential for third-party influences on merger
10 For a discussion of the possible distortions of
financial ratios resulting from merger accounting see O'Keefe
(1992), pp. 23-24.
11 See Billett, Coburn and O'Keefe (1995) for a
discussion of FDIC-assisted mergers and the unique nature of the
12 Significance tests were done using Student's "t"
statistics of mean differences in financial ratios. A 95-percent
confidence level was used as the criterion in all significance
13 A discussion of logit analysis can be found in
14 The HHI is a measure of product market
concentration and measures market shares across competitors in a
market. The HHI is defined as the sum of squared market shares for
all competitors in a well-defined geographic or product market.
High HHI values occur when one or a few sellers possess large
market shares, while lower HHI values occur with a larger number of
competitors with lower individual market shares.
15 Several variations of equation 1 were also
tested. The additional terms tested included measures of regional
location, alternative measures of profitability (return on assets
and return on equity), nonperforming assets, total noninterest
expense, and composite CAMEL ratings. The results of those models
were similar to those found with equation 1, which yielded the
strongest predictive results.
16 In logit estimation the estimated coefficient,
k, indicates how a change in the independent variable it is associated with, xk,
changes the natural logarithm of the ratio of the probability of merging (either
as an acquirer or target bank), to the probability of not merging in the
estimation time interval.
17 The pseudo R2 statistic is equal to 1 minus the
ratio of the log of the likelihood function maximized with all the
explanatory variables included in the model, to the log of the
likelihood function maximized in the model with none of the
explanatory variables included (except the intercept). See Maddala
18 See Judge, et. al. (1985), p. 767.
19 See Billett, Coburn, and O'Keefe (1995) for a
discussion of these assurances.