Revised Draft: January 25, 2000
INCENTIVES FOR BANKING MEGAMERGERS: WHAT MOTIVES MIGHT
REGULATORS INFER FROM EVENT-STUDY EVIDENCE?
Edward J. Kane
Finance Department
Boston College
Abstract
Methodologically, this paper frames the opportunity cost of any merger as the
value of the alternative deals it precludes or defers. This challenges the standard eventstudy hypothesis that stock markets benchmark the value of a merger deal by the profits
the partners would have earned in stand-alone activity. Substantively, the paper finds
that megamergers in banking show two size-related exceptions to the prototypical result
that acquirer stock value tends to be unaffected or to fall when a merger is announced.
Giant U.S. banking organizations gain value from becoming more gigantic and gain
additional value when they absorb an in-state competitor.
JEL Classification Codes: G14, G21, G34
Keywords: Banking Megamergers; Event Studies; Financial Regulation; Too Big to Fail
INCENTIVES FOR BANKING MEGAMERGERS: WHAT MOTIVES MIGHT
REGULATORS INFER FROM EVENT-STUDY EVIDENCE?*
Edward J. Kane
Finance Department
Boston College
Financial-institution mergers and acquisitions are dramatically remolding at least
five dimensions of U.S. and global banking competition. Dimensions changing most
markedly include the number of leading players, their product lines, their front-office and
back-office technologies, their geographic reach, and their average size.
In the U.S., the Bank Holding Company Act assigns central bankers the duty of
assessing the private and societal benefits that derive from combinations of very large
financial institutions. To fulfill this duty, staff economists seek to identify the motives of
each would-be acquirer (Amel, 1997). Unfortunately, merger partners often have good
reason to hide their true motives from government officials.
Virtually every merger application projects improvements in efficiency or
portfolio diversification. Authorities must routinely doublecheck the plausibility of these
projections. They must also search for private benefits that may flow from increases in
monopoly power, managerial entrenchment, or political clout.
To evaluate a given merger, regulators must solve three problems. First, they
must decide which of the eight possible reasons for merging listed in Table 1 actually
apply. Next, they must estimate whether each proposed combination promises enough
social benefits to offset social costs from enhanced political influence and monopoly
power. Finally, authorities must figure out whether and how they might intervene in the
deal to mitigate undesirable effects. Although authorities have been reluctant to block
banking mergers during the last decade, numerous deals have been modified to assuage
official concerns (Amel, 1997).
2
It is often presumed that stock-market reactions to merger announcements reliably
predict the post-acquisition performance of the combining firms. To the extent that this
standard event-study hypothesis is true, a resource-saving three-stage process can be used
to pinpoint megamergers that might deserve special scrutiny. The first round of tests
would seek to interpret the size and distribution of the changes in stockholder value that
occur on or around the merger’s announcement date. These first-round tests would look
for evidence consistent with the hypothesis that increased size per se might generate
regulatory or monopoly rents. When such evidence emerges, closer analysis would be
called for, and two further rounds of investigation would kick in.
Second-round tests would investigate whether regression models could fairly
allocate a substantial portion of announcement-created stockholder value to regulatory or
monopoly-power benefits. If so, staff analysts would be directed to undertake a detailed
third-round econometric investigation to isolate the technical-efficiency and
diversification benefits that the proposed combination might generate. The third-round
investigation would formally audit the partners’projections and develop additional data
needed to carry out a comprehensive cost-benefit analysis of how to ameliorate the
merger’s undesirable effects.
This paper limits itself to indicative first-round and second-round tests. It begins
by motivating and conducting event-study tests of the stock market’s reaction to the
largest bank mergers and acquisitions announced in the U.S. during each of the years
1991 to 1998. Previous studies have found that the value of the target’s stock usually
increases, but that this increase tends to occur at the expense of the shareholders of the
acquiring institution. In what we may call the standard corporate merger, the price of the
acquirer tends either to be unaffected or to decline somewhat when merger plans are
announced. For example, in a study of 101 1968-1986 tender offers, Lang, Stulz, and
Walkling (1999) report that the mean deflections of stock prices caused by merger
announcements (so-called “abnormal returns) are slightly negative for acquirers and
markedly positive for target firms. Using a five-day pre-announcement window and a
1985-1991 sample of 153 large-bank mergers, Houston and Ryngaert (1994) found
significantly positive abnormal returns for targets and significantly negative abnormal
returns for acquirers, with little evidence that the average bank merger creates any value.
3
In a study that uses an innovative statistical method to analyze 154 banking organizations
that acquired other banks during the slightly different sample period 1989-95, Cyree and
DeGennaro (1997) also found that bank acquirers seldom earned an announcement
return.
The tests conducted here identify some size-related megamerger “exceptions” to
the finding that merger announcements in banking seldom create statistically significant
net increases in stock market value. In interpreting this result, the paper challenges the
standard event-study hypothesis that announcement-day excess returns reliably predict
post-acquisition improvements in firm performance. It argues that event studies weigh an
announced deal against previous market evaluations of the partners’other potential
targets and acquirers, so that a poor acquirer return may indicate that speculators had
previously identified a better hypothetical partner for the acquirer than the particular
target it actually chose to marry.
Our regressions establish that patterns of value creation differ with the nature of
the target firm. When a deposit institution is taken over, acquirer value grows with target
size in worrisome ways. Giant banking organizations gain value from becoming more
gigantic and gain additional value from absorbing bank targets that were previously
competing in-state.
The larger a target is, the larger and more complex its acquisition is apt to make
the banking organization that acquires it. With increased size, complexity, and market
share come increases in market power and political clout. Increased market power
engenders monopoly rents and increased market share enhances an institution’s
significance to members of the congressional delegations of the states in which it
operates. Increased political clout intensifies incentive conflicts facing top regulators.
The Long Term Capital Management rescue shows that increased portfolio complexity
makes it harder for regulators to monitor and enforce limits on an institution’s risk
exposures. In particular, any merger that strengthens market presumptions that a
megabank acquirer is “Too Big to Fail and Unwind” –or, more accurately, too big to
discipline adequately (TBTDA)-- lowers that entity’s financing costs. Funding costs fall
because the institution’s increased size enhances its access to unpriced de facto taxpayer
guarantees of its uninsured debt (Todd and Thomson, 1990). How much power and clout
4
is created and how these capacities are expected to be exercised are important questions
that event-study experimentation cannot directly answer.
The simplest explanation of our results is that, when a large bank absorbs a very
large target bank, stockholders expect the acquirer to price its products less competitively
in the future and to use its expanded political clout to extract additional regulatory
forbearances. The unpleasant social costs this simple explanation implies make the costs
and benefits of banking megamergers an urgent area for public-policy research.
Although the technology-driven contestability of modern banking markets may be
expected to discipline postmerger efforts to exercise monopoly power, mechanisms for
reining in TBTDA rents are crude and unreliable. Moreover, the demise of GlassSteagall and Banking Holding Company Act restrictions on U.S. banks’ability to affiliate
with other types of financial firms threatens to make TBTDA subsidies more widely
accessible than ever before.
It should be clear that TBTDA and monopoly benefits are not the only possible
explanations for the results we observe. Our profession prides itself on being able to
weave sophisticated counterintuitive explanations for what a layperson would deem to be
straightforward phenomena. In line with this predilection, models could be devised that
would attribute the correlation between target size and announcement-day value creation
to nonlinearities in scale and scope economies (particularly in extending a megabank’s
brand to new products and customers) and attribute the incremental value that emerges in
in-state deals to the avoidance of distance-related operational expense. However, until a
more sophisticated model has been validated econometrically, Occam’s Razori and
Bayesian inference both imply that officials should adopt the straightforward
interpretation as a working hypothesis, and this implication is reinforced by asymmetries
in the size of the costs that banks, customers and taxpayers face if the simple
interpretation does not apply. The TBTDA and network-control dangers implied by the
simple reading of announcement-day patterns of value creation transmit to top regulatory
officials duties of vigilant regulatory response (Hoenig, 1999) that extend far beyond
traditional efforts to protect household and small-business customers by postmerger
branch-office divestitures that limit market concentration in particular locales.
5
I. What Benchmark Evaluations Drive Announcement-Day Movements in Stock Prices?
Banks are in the information and dealmaking business. In negotiating and
executing deals, information is simultaneously an input, an intermediate product, and a
final output. The informational function of the bank works to support its dealmaking
function by helping to identify and value the pluses and minuses of every deal the bank
holds or considers.
The Importance of Intangibles
The information a bank uses is neither freely available nor completely reliable and
verifiable. Each bank strives to build up and maintain its capacity to access, store,
process, and transfer information at low cost. In principle, the discounted value of the
future dealmaking profits that may be fairly attributed to the bank’s ability to exercise its
special informational capacities represents an intangible asset of the bank.
The value of a bank is the net value of its intangible and tangible assets and
liabilities. Tangible assets and liabilities (TA and TL) may be defined as balance-sheet
positions that may be traded individually to another institution. Intangible items consist
of rights granted by a government or private party and positions that --although lacking a
separate physical existence-- may be used in conjunction with a bank’s other assets to
generate profits. A bank’s intangible assets (IA) include its firm-specific informational
and dealmaking capacities and various real options (RO) that these capacities create.
In the U.S., Generally Accepted Accounting Principles (GAAP) permit very few
intangible items to be booked as part of a bank’s accounting net worth. However,
economic net worth (ENW) is the difference between TA+IA and TL+IL. Equivalently,
ENW (i.e., a bank’s market capitalization) may be expressed as the sum of the bank’s
tangible and intangible net worth (TNW + INW):
ENW = (TA+IA) – (TL+IL) = (TA-TL) + (IA-IL) = TNW + INW.
(1)
Megabank mergers and acquisitions are merely a particular form of dealmaking.
An acquirer (A) buys the tangible and intangible balance sheet of the target (T) from its
shareholders for a bundle (BA) of tangible assets (or liabilities) and combines the
resulting positions with its own. Figure One illustrates the economic balance sheets of
the premerger entities.
6
The tangible net worth of the combined postmerger entity (TNWC) is easy to
define and relatively verifiable:
TNWC = TNWA + TNWT – BA.
(2)
The value of C’s intangible net worth (INWC) is more problematical. This is because the
merger opens some real options that would not exist without the merger and closes some
of the options the individual partners’previously enjoyed. Williams (1993) emphasizes
that changes in ownership structure can greatly influence the value of real options.
Options created include opportunities to apply superior technologies, to marry the
newly combined entity down the line with still other firms and to cross-sell to the
customer base of both firms products that the premerger banks had not offered before.
Options closed include opportunities to merge with some other partners and (perhaps) to
operate some branches whose market areas overlap.
In purchase accounting, GAAP requires the target’s tangible items and so-called
“identified intangibles” to be marked to market (MM) and the difference between the
purchase price and MM is booked as ”goodwill” (G). By itself, the value of G provides
no clue as to whether a merger creates or destroys stockholder value. Figure Two
clarifies that whether a hypothetical merger creates or destroys value depends on whether
the change in intangible net worth (∆IA-∆IL) exceeds or falls short of the value of the
assets that are being conveyed to target shareholders:
∆ENW = ENWC – (ENWA+ENWT) = ∆IA-∆IL-BA.
(3)
A richer question that the stock market must address is whether the premerger
resources of A or T could earn even better returns if they were combined with the
resources of a different merger partner. The most predictable postmerger changes in
intangible value would be those that would be available or “common” to any potential
acquirer of T. We denote market estimates of these common benefits by k. We assume
that these are predominantly efficiency benefits from installing new management and that
estimates of k do not change during the process of negotiating the merger deal. Other
estimated intangible benefits would be specific to the particular strategies and
opportunities each of m potential acquirers the market has identified would bring to the
negotiating table. We denote these partner-specific benefits as ki, i=1, … , m. In an
announced merger of A with T, this notation implies that ∆IA-IL=k+kA.
7
Ownership of nearly all large U.S. banks resides in a holding company. This
pattern of layered control smoothes the way for changes in governance at individual
institutions. In an active and informationally efficient market for corporate control,
speculators would seek to identify --in the Bayesian sense of establishing subjective
probabilities based on available evidence-- banks that were ripe for takeover. Assuming
speculators are risk-neutral, they would bid the expected benefits of taking over each
value-creating target into the stock prices of every target and all of its potential acquirers
well in advance of every actual deal. Of course, speculators would recognize that a bank
could be simultaneously an acquirer and a target in different potential deals, that
managers of A may subsequently mismanage the merged institution, and that the best
potential deals for either A or T may not be the one they are announcing. In this
informational environment, a combination of banks that improved the economic
efficiency of the stand-alone firms could nevertheless lower the announcement-day stock
price of an actual acquirer or target. This would occur when speculators believed that the
deal announced precluded a superior combination or fell short of the hypothetical benefits
attributable to other better combinations that speculators had previously identified.
Our analysis makes a number of convenient assumptions. We analyze
announcement-day price movements under the assumption that –although the acquirer
might be forced to raise its bid-- every announced deal is certain to be completed. In this
case, each announcement reduces uncertainty about three things: what targets will be
acquired, who will acquire them, and whether a deal for T is currently being negotiated.
Differences in the amount of uncertainty to be resolved can explain the James and Weir
finding that, in 60 bank acquisitions announced during 1972-83, acquirer returns proved
positively related to the number of alternative targets and negatively related to the
number of other potential bidders in the target’s market.
Prior to the announcement, we presume that risk-neutral speculators assign a
probability qT to the possibility that bank T will be taken over. The lower is qT, the more
“surprising” the market would regard the takeover announcement. We also presume that
speculators assign each target’s potential acquirers a probability Π i of success and an
expected cost of Di for merger-related charges and revenue displacements. We also
suppose that a potential acquirer’s probability of gaining the target would rise with the
8
partner-specific values it can create with the target. Finally, to avoid complicating the
exposition by discount factors, we exclude from the k’s and the Di the wedge of present
value that is surrendered in waiting for a deal to unfold.
Deals Between An Unambiguous Target and Unambiguous Acquirer
We start by looking at a target bank that is simultaneously considered certain to
be acquired and certain never to be an acquirer. On the day before T’s takeover is
announced, the takeover premium TPT(t-1) imbedded in its stock price would be the
expected value of the successful bid BA:
m
Et-1(BA) = TPT(t-1) = k +
∑
i =1
Π i (t − 1)ki −
m
∑
i =1
Π i (t − 1) Di (t − 1) .
ii
(4)
Unless another acquirer (AA) is expected to surface, the expected value of the bid on the
announcement day t would move to BA, the compensation set by A. In this case, the
announcement-day change in the target’s takeover premium would be:
∆TPT = BA – Et-1 (BA) = BA – k – Et-1 (ki) + Et-1 (Di).
(5)
Equation (5) tells us that announcing that firm A has won the competition for the
target resolves uncertainty about the values to the acquirer of specific intangible benefits
ki and dealmaking costs Di. It also sets a minimum value on the bundle of assets target
stockholders are to receive. If better acquirers exist, (5) would include the conditional
expected value of subsequently having to outbid the incremental compensation the best
alternative partner might offer. This incremental value would equal at least the amount
B* by which speculators believe that the (ki-Di) of the second-best acquirer exceeds BADA.
If we assume that speculators do not view A as a potential target in any alternative
combination, the change in the best acquirer’s potential takeover benefits from acquiring
only T can be calculated in a similar manner. On announcement-day eve, TPAT (t-1)
would equal the market’s expected value of A’s net return from merging with T in due
time:
TPAT (t-1) = Π A [k+kA(t-1)-DA(t-1)] - Π A E(BAA=acquirer).
(6)
It is reasonable to assume that merging with T puts A’s managers in what we may call a
“digestive state” during the time it takes to assimilate the operations of T. During this
interval, A’s prospects for merging with other targets are reduced. This adverse effect on
9
the acquirer’s other takeover options (∆MOA) might be negligible for very small targets.
On the other hand, although the diversion of management time from other deals might be
expected to grow with the complexity of the merger, predeal planning and due diligence
can limit the duration of the digestion period.
The unobservability of prior effort makes the effect of complexity on projected
dealmaking costs and other merger prospects ambiguous. Uncertainty about kA probably
also increases with a deal’s complexity. Our second-round statistical tests use the onbalance-sheet asset size of the target (TAT) as a proxy for partner-specific benefits and
complexity, so that we write the announcement-day takeover premium for A as:
TPAT (t) = ft(TAT) = k + kA(t) – DA(t) – BA - ∆MOA,
(7)
with the sign of ft ' (TAT ) to be determined empirically. The announcement-day change in
TPAT can be found by subtracting equation (6) from equation (7). If kA and DA don’t
change on the announcement day:
∆ TPAT = (1-Π A) [kA(t-1) – DA(t-1)] – [BA-Π A E(BAA=acquirer)] - ∆MOA.
(8)
Difficulty of Interpreting the Signs of Changes in TPT and TPA
Day-to-day changes in TPT and TPAT are not directly observable in stock-market
data. What we can observe is the announcement-day percentage change [RT(t) and RA(t)]
in the stock prices of the target and the acquirer. The simplest way to render changes in
TPT and TPAT observable is to assume that, on the announcement day, these changes
account for the deviation of RT(t) and RA(t) from the percentage change in the rate of
increase experienced by the stock market as a whole [RM(t)]. One standard event-study
test uses the so-called market-adjusted returns [YT(t) and YA(t)] as proxies for ∆TPT(t)
and ∆ TPAT (t), respectively:
YT(t) = RT(t) – RM(t),
(9)
YA(t) = RA(t) – RMt).
(10)
The standard interpretation of event-study outcomes presumes that stock markets
value each deal against the benchmark of the profit opportunities the partners would
possess if they were left to operate in a stand-alone mode. But the true opportunity cost
of any merger is the value of the deals it precludes or defers. Opportunity-cost analysis
10
clarifies that the benchmark that should govern announcement-day reactions to a
megadeal is how target and acquirer resources would perform in the best alternative
megadeal whose execution the announced merger either precludes or makes much less
likely to occur.
In part, the asymmetry in the announcement-day movement of acquirer and target
stock prices that researchers typically observe reflects asymmetry in alternative partners’
ability to improve on a suboptimal deal after it is announced. Although the acquirer’s bid
may easily be forced higher, it is unlikely to go down. This is because each merger
announcement conveys information about the target to any superior acquirer and gives
that acquirer a hard value to bid against. Acquirer A’s managers know that any
alternative acquirer (AA) who sees substantial value in the target is apt to woo target
shareholders by offering a richer deal and that the process of taking a shareholder vote on
A’s offer creates an urgent deadline for AA to do so. In contrast, potential partners in
better deals that the acquirer A’s managers may have overlooked or even sought to thwart
have less incentive to throw themselves immediately into the acquirer’s arms. Whether
they are prospective targets or acquirers, they know that completing and digesting the
announced megadeal will, at least for awhile, restrict acquirer A’s ability to contemplate
a second megamerger.
The due-diligence process may well give insiders better estimates of kA and DA
than speculators possessed on the preannouncement day. Assuming that this better
information leaks out as part of the announcement makes it hard to interpret the sign of
changes in TPT and TPAT , even if these variables could be observed without error.
∆TPT could be positive for a number of reasons: (1) because A’s bid reflected its
ability to make better postmerger use of T’s tangible and intangible portfolios than other
acquirers; (2) because A’s informational and dealmaking skills gave it a lower-thanaverage dealmaking cost; (3) because speculators had underestimated kA or overestimated
DA on t-1 [a situation which would bias E(BA) downward], or (4) because (as supposed in
the theory of the winner’s curse) A’s managers overpaid for the target.
Similarly, we might observe a decline in TPAT for a number of different reasons:
(1) because A’s managers surfaced adverse information about kA and DA during the duediligence process, (2) because A’s managers were pursuing personal benefits at
11
stockholder expense, (3) because A paid a surprisingly high price for T, or (4) because
the merger options being set aside were extremely valuable.
Deals Between a Switch-Hitting Target and Acquirer
Speculators may envision real-world targets and acquirers as occupying alternate
sides of different hypothetical deals. We label such targets and acquirers as switch hitters
(ST and SA). Switch-hitting introduces an additional term into the takeover premiums
and expands the set of merger options that an acquirer closes when a deal is announced.
For switch-hitters, the announcement-eve takeover premium is the sum of two
conditional expectations: (1) the expected value of its net takeover premium given that it
is a target and (2) expected value of its net takeover premium given that it is an acquirer.
Each conditional expectation is the product of the probability (q) of ending up on a
particular side of a merger deal (A or T) and the expected net benefits that speculators
believe will occur in that particular state.
Before the deal is announced, four dealmaking states of the world exist: ST=T,
ST=A, SA=T, and SA=A. At t-1, speculators assign each state a corresponding
probability. We denote these probabilities as qTST , q AST , qTSA , and q ASA , respectively.
Using this notation, the announcement-eve takeover premium for ST and SA become:
TPST(t-1)= qTST {k+Et-1[(kA-Di)ST=T]}+ q AST {kA+Et-1[kA-DA-BAST=A]}.
(11)
TPSA(t-1)= qTSA {k+Et-1[(kA-Di)SA=T]}+ q ASA {k+Et-1[kA-DA-BASA=A]}.
(12)
The announcement-day value of each premium replaces the unrealized
conditional expectation with the dealmaking values that actually obtain and subtracts the
value of the real options surrendered in the deal. For a switch-hitting target,
TPST(t) ≥ BA - TPST(t-1).
(13)
As before, the option of other would-be acquirers to bid on the target remains
open. The inequality in (13) allows for the possibility that another acquirer may
subsequently enter the bidding. Even if we assume SA is the best acquirer for ST, TPST
will be enriched and TPSA(t) reduced by an amount slightly in excess of the market’s
estimate of the difference between an amount B* equal to (ki-Di) of the second-best
acquirer and SA’s initial bid B A.
TPSA (t ) = ft (TAT ) = k + k A (t ) − DA (t ) − ( BA + B ) − ∆MO
*
12
(14)
II. Focus of Announcement-Day Event-Study Tests
Event-study test procedures turn on two benchmarks for announcement-day
stock-price returns. Excess returns relative to either benchmark might indicate that a
particular megamerger promises partner-specific regulatory or monopoly-power benefits.
To express these benchmarks, we let MCA(t) and MCT(t) represent the respective values
of acquirer and target stock on the date the merger is announced. MCA(t-1) and MCT(t-1)
represent these stocks’aggregate pre-announcement values.
Analyses of standard corporate takeovers usually find that competition for a
desirable target from other potential acquirers leads an acquirer to offer a bid BA that
drives up the announcement-day price of the target firm without usually experiencing a
favorable movement in its own stock price. However, in a banking megamerger, we
might expect partner-specific benefits to prove important for two reasons. First,
competition for targets is curtailed because few target banks are large enough, diverse
enough, or risky enough to make much difference in an organization’s postmerger
profitability. Second, substantial regulatory and monopoly-power benefits can be
expected to accrue reliably to only a small number of acquirers.
The benchmark for what we call “strong case” evidence of the possibility of
undesirable private benefits occurs when the acquirer’s and the target’s stock both rise on
the announcement day. We characterize this case as evidencing “Hypersynergy”:
MC A (t ) − MC A (t − 1) ≥ 0
MC T (t ) − MC T (t − 1) ≥ 0.
(15)
A weaker case of “Aggregate Value Creation” occurs when the acquirer’s stock price
falls, but does not fall far enough to prevent the pro-forma market capitalization of the
combined entity from increasing:
MCA(t)+MCT(t)>MCA(t-1)+MCT(t-1).
(16)
The hypersynergistic case is consistent with the idea that the source of acquirer’s
stock-price appreciation is something that cannot easily be captured by other bidders and
that may benefit each partner through either their joint domination of a transactions
and/or information network or improved opportunities to increase their leverage and
other risk exposures. The strength of hypersynergy criterion is that, to the extent that the
13
particular benefits of a merger are unique to the acquirer, A can outbid T’s next-best
partner without having to transfer A’s benefits entirely to T’s stockholders. In particular,
whenever a merger’s incremental value lies in strengthening safety-net guarantees or
integrating control over an important informational or transactional network, the
announcement of the deal ought to favor both stocks. In contrast, whenever a deal is
motivated merely by acquirer hubris or other types of managerial agency costs, we may
reasonably expect competition among rational alternative bidders to force the empirebuilding acquirer into making an overbid that would depress MCA(t). Empirical studies
of postmerger experience suggest that speculators ought to treat partner-specific savings
in operating costs as doubtful and slow to materialize (Berger, Demsetz, and Strahan,
1999; Houston, James, and Ryngaert, 1999). We adopt this view as a working hypothesis
and also hypothesize that partner-specific increases in monopoly power that do not trace
to TBTDA-enhanced branding or to improved control of a transactional or informational
network are unlikely to be durable. In markets that are globalizing because of increased
contestability, the threat of new entry makes nonnetwork sources of monopoly power
hard to sustain.
The weaker-case Value Creation benchmark is harder to interpret. As we have
seen, unless the merger completely surprises speculators, announcement-day price
movements omit some of the economic value the merger conveys to target and acquirer
stockholders. To the extent that a bank is seen to be a highly likely target, most of the
economic value it has to offer an average acquirer will have already been impounded into
its preannouncement market capitalization (Calomiris and Karceski, 1998). This means
that announcement-day hypersynergy is a sufficient, but not necessary indicator that
substantial economic value is created by the combination. The potential importance of
unmeasured value creation makes it prudent to subject mergers that manifest weak-case
value creation to at least our second round of tests. For mergers that either meet or nearly
meet the weak-case criterion, it is incumbent on regulatory staff economists to look for
evidence of increased monopoly power or enhanced regulatory avoidance.
III. Event-Study Test Results
14
To test the hypothesis that stockholders derived monopoly or regulatory benefits
from recent U.S. megamergers, we must first develop operational definitions for
“megabanks” and “megamergers.” We initially define a megabank as one whose parent
holding company ranks among the top 12 U.S. bank holding companies (BHCs) either in
asset size or in market capitalization. One justification for focusing on the top 12 BHCs
is that this criterion for being Too Big to Fail (TBTDA) was stated by the Comptroller of
the Currency in Congressional testimony explaining the motivation for the Continental
Illinois bailout (O’Hara and Shaw, 1990). Table 2 lists the top 12 BHCs in each category
on June 30, 1998. Although the rankings vary, the overlap between the two lists is
substantial. Only 14 different names appear.
In the next subsection, we define a 1998 banking megamerger as a merger or
acquisition involving one of these 14 BHCs or a thrift of similar size that increases the
pro-forma size or market capitalization of the merged organization by at least half the
amount of assets or market capitalization shown for the 12th-ranking BHC. More
extensive time-series testing is undertaken in the subsequent subsection, in which the
identities of large BHCs and the size of qualifying pro-forma increments vary year by
year.
Value Creation in the “All-Time Top Ten Merger Deals”
During the first half of 1998, five deals met our megamerger criteria:
1. Travelers Group, Inc. – Citicorp
2. NationsBank Corp. – BankAmerica Corp.
3. Norwest Corp. – Wells Fargo & Co.
4. Banc One Corp. – First Chicago NBD Corp.
5. Washington Mutual Inc. – H.F. Ahmanson & Co.
In 1998, these deals were listed among the all-time top 10 bank and thrift deals in market
value by Nickerson (1998). Listed in order, the other five all-time deals are:
1. First Union Corp. – CoreStates Financial (11-18-97)
2. NationsBank Corp. – Barnett banks (8-29-97)
3. Wells Fargo & Co. – First Interstate (1-24-96)
4. Chemical Banking – Chase Manhattan (8-28-95)
5. NationsBank Corp. – Boatmen’s Bankshares (8-30-96)
15
Table 3 adjusts announcement-day stock returns for merger partners in these topten deposit-institution deals to remove the effect of the average price appreciation earned
that day on the S&P 500. The empirical literature on bank mergers finds that the stock
prices of acquirers seldom appreciate on the announcement day (e.g., Berger, Demsetz,
and Strahan, 1999; Calomiris and Karceski, 1998; Clark, 1988; Cornett and Tehranian,
1992; Cyree and DeGennaro, 1997; Houston and Ryngaert, 1994; Houston, James and
Ryngaert, 1999; and Willmarth, 1995). This supports our working hypothesis that, unless
acquirer-specific regulatory or monopolistic benefits exist, a very small proportion of
acquirers would experience a favorable announcement-day price movement. Moreover,
one may infer that adverse regulatory responses to megabank price movements are
expected to be softened by lobbying activity. Otherwise, the fear that antitrust and
regulatory issues would lead conscientious authorities to prevent a blockbuster deal’s
completion would exert offsetting downward pressure on the acquirer’s stock price
whenever its announcement-day value advanced strongly.
In contrast to results for ordinary banking mergers, the calculations presented in
Table 3 indicate that the acquirer’s market-adjusted stock price increased in three banking
megamergers and that the value of the combined enterprise increased substantially in
these cases. These hypersynergistic cases all involve partners that ranked among the
nation’s top-five banks. Two other instances show better than a 5 percent weightedaverage return. Both of these entail the takeover of a giant thrift institution. Only the
Norwest-Wells merger fails the weaker value-creation criterion outright, while the Banc
One-First Chicago NBD and First Union-CoreStates mergers produce only a 0.1 percent
weighted-average return.
The wide range of weighted-average returns shown in Table 3 clarifies that
sometimes the benefits of banking megamergers either are fully anticipated or are
regarded by the market as inferior to some of the acquirer’s other feasible deals. Market
participants habitually worry about disinformation, acquirer empire building, and the
sustainability of projected improvements in profitability. Houston, James, and Ryngaert
(1999) show that analysts are particularly skeptical about allegations that merged
operations will generate large operating-cost savings. Keefe, Bruyette & Woods found
16
that only two of the top 15 banking megamergers of 1997 and 1998 met their earnings
projections for 1999 (Padgett, 1999). Figure Three displays the six worst outcomes.
To the extent that some uncertainty is apt to be resolved in every megamerger
announcement, targets may be expected to experience value creation. That in four of the
all-time deals the combined mega-enterprise created little weighted-average net
announcement-day value is consistent with the hypothesis that the deals these acquirers
selected may have closed more-valuable real options. Nevertheless, this analysis of the
all-time top-ten U.S. banking deals suggests that partner-specific regulatory and
monopolistic benefits emerge in mergers in which the target is one of the nation’s very
largest banks or thrifts.
Value Creation in the Top Deals of 1991-1998
Procedures we have labeled as “second-round” tests seek to identify some of the
determinants of excess announcement-day returns. Hypotheses of this kind cannot be
tested effectively using a sample of only ten deals. To expand the investigation
efficiently, we need to define a megamerger differently. We begin by identifying the
largest 15 banking mergers that took place in each of the eight years 1991-1998. Table 4
summarizes the means, standard deviations, and number (N) of observations collected on
characteristics of targets, acquirers, and mergers whose effect on announcement-day
returns we intend to investigate.
In six of the potential 120 cases, we were not able to assign a precise
announcement date. Appendix Table 1 identifies three 1998 observations we eliminated
because we couldn’t locate the announcement-day return for the acquirer on CRSP tapes.
In second-round tests, we dropped a few more cases because of difficulties in finding
accounting data on asset size for target thrifts and securities firms.
The average asset size of the targets is half the size of their acquirers. In line with
other studies of bank mergers, the mean announcement-day return for targets ( X T ) is
large and statistically significant, while the mean excess return for acquirers X A is
negative and significant. However, what is unusual in our results is that the weightedaverage excess return for the combination ( X W ) is a marginally significant 0.83 percent.
17
Looking at the frequency distribution of XA, XT, and XW reveals that, in 29 of the
megamergers for which we can calculate all three variables, the acquirer’s excess return
(XA) is positive. Moreover, in about half of these hypersynergistic megamergers, XA
exceeds 1.0 percent. The target’s excess return is positive in 82.3 percent of the cases.
Finally, the weighted-average excess return of target and acquirer (XW) is positive 52.2
percent of the time.
Appendix Table 1 lists the particular megamergers that survived the target excessreturn screening process. The third and fifth columns note the rank of partners who
would meet the top-12 BHC size criterion employed in the previous subsection.
The purpose of constructing the larger sample is to test whether and how
projected megamerger benefits change with a few easily verified characteristics of the
partner firms. Tables 5 and 6 present the results of a series of parsimonious second-round
regression experiments. These models account for and confirm the potentially dilutive
financial-structure impact of the deal on acquirer stockholders, by introducing a binary
variable STK that equals one when the acquirer uses stock in its bid package and equals
zero otherwise. Table 5 explores alternate regression models for acquirer excess returns,
while Table 6 seeks to explain the weighted-average return found for the combined
enterprise.
The first two panels of Table 5 look only at acquisitions of deposit-institution
targets. The panels tell us that the asset size of a deposit-institution target is the most
important single determinant of acquirer excess returns. The second-panel and thirdpanel equations indicate that the value creation we observe is not rooted in benefits
generated by diverisifying into new markets. This reinforces results compiled by
DeLong (1999a and 1999b) and Cornett, Hovakimian, Palia, and Tehranian (1999) on
differences between interstate and intrastate banking mergers. Solving the third-panel
equation for the target size at which the positive slope term first overcomes the negative
 .20 
intercept =
 tells us that we may expect XA to be positive whenever an in-state
 .075 
target’s asset size exceeds about $3 billion.
18
Some lines report t-values without showing a corresponding regression
coefficient. This is done to communicate the effects of introducing potential regressors in
stepwise fashion.
These solo t-values support four further inferences. First, out-of-state acquisitions
create less acquisition-day value than in-state deals. In-state deals may be interpreted as
eliminating a competitor and strengthening an institution’s standing with state politicians.
Out-of-state deals are often designated as market-extension mergers. A straightforward
potential explanation for the differential result would be that in-state deals create or
enhance acquirer monopoly power and Congressional access. Second, when the number
of employees is introduced into the first-panel runs, it receives a negative sign. This is
inconsistent with the idea that the opportunity to cut jobs per se reliably creates acquirer
value. However, the employees variable loses significance once the differential effects of
the size of out-of-state targets are controlled for. Third, although five observations
cannot provide compelling evidence, the securities-firms targets (SFT) in our sample
generate significantly less incremental and weighted-average value than bank and thrift
targets as their asset size increases. As with narrow banking market-extension mergers,
this is consistent with the idea that partner-specific benefits are less in mergers that
diversify or extend a banking organization’s scope than in those which increase a
megabank’s share of its traditional markets. It suggests that megabanks may be paying to
acquire proven investment-banking talent, but are not expected to retain at least some of
their target’s key players. Fourth, individually introducing the assets of the combined
enterprise or a binary intercept-shift dummy variable to indicate when a merger involves
a top-5 BHC does not significantly improve the model’s fit.
To double-check the potential importance of the binary variables, we re-estimated
the model of panel one, interacting both the slope and intercept of target assets with each
dummy variable. As before, B5 shows no significant effect for either parameter. Because
a significantly negative slope shift did emerge for out-of-state and securities-firm
acquisitions, we designate the regressions displayed in panels two and three as our
working models of XA.
Except that the intrastockholder transfers from including stock in the bid package
wash out (as they ought ) across the combined enterprise, regression experiments
19
attempting to explain RW produce parallel results. As shown in Table 6, the asset size of
an acquirer’s deposit-institution target is the single most important variable, and its effect
is significantly less in out-of-state acquisitions than in in-state deals. The intercepts and
slopes found in these experiments imply that net stock market value is created even for
combinations in which deposit-institution target assets are barely in the neighborhood of
$1 billion.
IV. Reasons for Fearing that TBTDA Remains Alive and Well
Our findings raise regulatory and antitrust questions that public-policy economists
should investigate closely. The central issue is how much of the increased stock-market
value is generated in the private real economy and how much is generated by
presumptions that the enterprise’s improved postmerger interface with government
officials can reduce the costs of its debt financing (Kane, 1991; Boyd and Graham, 1998).
Because the standard microeconomic theory of firm value treats a firm’s government
interface as a sideshow, an economist’s instinct is to presume that in banking
megamergers, too, the value created would trace to improvements in an enterprise’s
operating revenues and costs. The argument that follows suggests that competitive forces
at work in banking today make antitrust activity and interventions to assure efficient
resource use potentially less important than policies that might lessen the political
pressures that make megabanks “too big to discipline adequately.”
Limitations on Operating Sources of Net Economic Value
In the absence of TBTDA benefits, a merged enterprise can create net economic
value in three broad ways. First, the merger may enhance future banking revenue by
increasing monopoly power over either network operations or loan and deposit prices.
Second, the merger may enhance banking revenue by opening new territory: reaching
new customers or making additional portfolio diversification and product lines feasible.
Third, the merger may generate operating-cost savings by consolidating locations,
equipment, and staffs.
In a megamerger, increased monopoly power is most likely to develop from fuller
network control and from brandable products like credit cards. Other forms of enhanced
monopoly power figure to be short-lived. Partly this is because a megamerger cannot go
20
forward, until antitrust officials and banking regulators determine that, in the various
markets in which the partners have previously competed, adverse competitive effects are
mitigated, for example, by branch sales. Countervailing pressure against monopoly
pricing will also be exerted by de novo entry in the postmerger marketplace (Berger,
Bonime, Goldberg and White, 1999). In recent decades, technological and regulatory
change has greatly extended the geographic reach of nonlocal banks and the ability of
nonbank financial institutions to offer bank-like products. Mergers and countermergers
are vehicles by which technological change is steadily integrating financial markets
across regions and countries and strengthening the actual and potential competitive
pressure exercised by nonlocal banks and by firms with nonbank charters. Credit cards
are increasingly national and global products, while ATM networks and interstate
branching are pushing checking accounts in a similar direction. Hence, from a publicpolicy point of view, the substantial contestability of modern banking markets can be
relied upon to generate strong counterpressure against aggressive monopoly pricing.
Although far from irrelevant, the public-policy implications of increased monopoly
power in banking mergers are apt to prove of second-order importance.
Empirical studies show that the importance of projected reductions in operating
costs is habitually exaggerated (Houston, James, and Ryngaert, 1999). In all mergers,
press releases routinely project operating-cost savings from consolidating the office
networks and staffs of the partner institutions. Although empirical studies sometimes
find significant cost savings from consolidation (Calomiris and Karceski, 1998; Cornett
and Tehranian, 1992), projected improvements in operational efficiency often fail to be
achieved ex post. Although technological change may lessen the relevance of past data,
empirical studies indicate that alleged economies of scope are problematical and that
operating economies of scale tend to peter out well before a bank reaches the TBTDA
size range (Clark, 1988; pp. 16-33; Berger and Humphrey, 1990; Noulas, Ray and Miller,
1990, pp. 94-108). Tables 5 and 6 support these doubts by finding a perverse impact for
work-force size in explaining XA and XW. Skepticism about economies of scale and
scope interact with concerns about empire building and managerial entrenchment (Gorton
and Rosen, 1995) to lead outsiders to expect announcement-day press releases to “spin”
the sources of consolidation profits in creative ways. Rational investors must be expected
21
to steeply discount alleged operating-cost savings to allow for the difficulty of carrying
out the necessary adjustments.
Realistically, the rate of discount applied to projected cost savings should be high
and the rosy projections participants offer should be revised downward even when the
quality and integrity of management at the acquiring institutions is perceived to be very
high. First, because corporate customers and wealthy households can benefit from
maintaining multiple banking relationships, the premerger customer bases of individual
partners often exhibit hard-to-sustain overlaps. Knowledgeable customers commonly
seek to play off one bank against another as a way of bargaining for better interest rates
on loans and deposits. Many customers with joint relationships will move one or another
segment of their premerger banking relationships to a competing entity. Second,
downsizings typically entail expensive employee buyouts, while branch closings and
most other forms of cost reduction threaten parallel reductions in the quality of at least
some services customers value. The point is that aggregate customer requirements for
floor space, front-office servicing facilities, tellers, automated teller machines, loan
officers, and many aspects of back-office support cannot be cut much without creating
unwelcome service deterioration or delays. Cost “savings” achieved by service cutbacks
undermine customer relationships. For customers that greatly value service quality,
implicit interest rates on loans rise and/or implicit interest rates on deposits fall. If the
projected postmerger economies entail noticeably longer queues, evidence that affected
customers are seeking better service elsewhere will rapidly temper the enterprise’s initial
cost-cutting zeal.
Why TBTDA Benefits Are Potentially Important
TBTDA may be described as a breakdown in ordinary regulatory discipline and
insolvency resolution that occurs when regulators encounter problems at an institution
whose size, complexity, or political clout is substantial. The bureaucratic effort and
reputational penalties top regulators experience from confronting and resolving an
individual insolvency increase with all three of these variables. However, the
presumption that complexity and clout correlate strongly with size explains why the name
given the doctrine adopts asset size as a catch-all condition for qualifying status.
22
Incentives that undermine regulators’ability to discipline TBTDA banks are
rooted in inescapable public-policy asymmetries (Kane, 1995). A top banking regulator’s
reputation is harmed less by the banking insolvencies that are allowed to develop during
his or her watch than by the surfacing of news that reveals to the public at large the
longstanding weakness of a major bank. The attraction of TBTDA to policymakers is
that forbearance makes insolvencies and risk exposures at the nation’s most important
institutions less visible to the newsmedia.
To make the TBTDA policy strategy less attractive to banking regulators, Title I
of the FDIC Improvement Act of 1991 (FDICIA) placed new constraints on insolvency
resolution. It imposed on federal banking regulators duties of prompt corrective action
and least-cost resolution. Section E of that Act’s Title I specifies that resolution costs
should be calculated “on a present-value basis, using a realistic discount rate.” Arguably,
the operative effect of FDICIA’s constraints has been to increase the size threshold A* at
which access to TBTDA subsidies kicks in.iii
However, recognizing regulators’conflicting social missions, the Act authorizes a
“systemic risk” exception for cases where least-cost resolution “would have serious
adverse effects on economic conditions or financial stability.” The systemic risk posed by
a troubled firm grows with its size and complexity. To keep the systemic-risk exception
from being used routinely, the law requires that the exception can only be triggered when
the FDIC, the Federal Reserve, and the U.S. Treasury jointly determine that such adverse
effects exist. As a disincentive to making such determinations lightly, the General
Accounting Office is required to evaluate the justification offered for every such
determination that is made.
The Fed’s 1998 intervention in resolving the capital shortage of Long Term
Capital Management shows that, even in the face of FDICIA restraints, TBTDA
incentives remain strong (Financial Economists Roundtable, 1999). Whenever a highlylevered large institution threatens to become insolvent, financial regulators face serious
conflicts in their social missions. Besides maintaining regulatory discipline, they must
worry about sustaining public confidence in the robustness of the financial system as a
whole and minimizing the community disruption that would follow from winding up a
mega-institution’s tangled affairs. In addition to the bureaucratic and personal costs of
23
prompt insolvency resolution, day-to-day efforts to promote these supplementary
regulatory missions give federal authorities a predictable preference for helping to
conceal economic insolvencies at important banks and for handling large-bank
insolvencies in ways that keep uninsured creditors whole. The complexity of a
liquidation and payout, the degree of community disruption, the risk of undermining
confidence in other institutions, and even the agency’s exposure to after-the-fact political
criticism all increase with the size and complexity of a troubled institution. Hence, the
larger an institution is, the smaller the chance that the Federal Deposit Insurance
Corporation (FDIC) will resort to asset liquidation and deposit payout in the event of
closure.
Empirical research confirms the contention that, prior to FDICIA, TBTDA
subsidies gave mega-institutions access to government-contributed risk capital (O’Hara
and Shaw, 1990; Todd and Thomson, 1990; Black, Collins, Robinson, and Schweitzer,
1997; Angbazo and Saunders, 1995). In principle, government risk capital reduces the
private cost of funding a TBTDA institution’s on-balance-sheet assets and off-balancesheet risk exposures. Routinely extending regulatory forbearances to stockholders and
creditors of large, complex, or politically powerful institutions subsidizes any activity
(including mergers) that makes an institution larger, riskier, more complex, or politically
more powerful. TBTDA benefits distort patterns of competition between banks and
nonbanks, between large and complex institutions and small and straightforward
institutions, and even between domestic and foreign financial venues.
The present value of TBTDA-generated government risk capital, GC, is part of
any mega-institution’s market capitalization, MC. At each mega-institution, the market
value of stockholder-contributed capital, SC, may be defined as:
SC≡MC-GC.
(17)
GC varies over time with a bank’s size and its borrowers’aggregate repayment prospects.
It also varies with elements of bank risk exposure that can be negotiated quickly, such as
exposures to relative movements in interest rates and in foreign exchange rates.
Derivatives trading, in particular, allows inadequacies in risk-based capital regulation and
in pricing deposit insurance and intraday credit to convey opportunities for megabanks to
increase the value of GC rapidly if and when they wish. Ceteris paribus, GC increases
24
whenever an institution increases the volatility of its future earnings or increases its
leverage in clever ways that authorities either do not monitor in timely fashion or do not
yet appropriately discipline.
A straightforward way to model the option-generating effects of TBTDA on
megamerger incentives is to treat GC as kicking in at a particular asset-size threshold
A=A* and increasing noticeably whenever a megabank’s asset size increases
substantially. A convenient approximation is to assume further that, for a given portfolio
structure, either authorities or private creditors require each bank to maintain a market
capitalization that is at least x percent of their total assets, where x increases with
perceptions of the bank’s overall exposure to loss:
MC GC + SC
x
=
≥
.
A
A
100
(18)
In this model, TBTDA implies that the percentage requirement to hold stockholdercontributed capital would decline whenever an institution managed to increase its asset
size substantially. Moreover, the decline would occur faster, the more sensitive TBTDA
benefits are to increases in asset size.
Rearranging (18), the effective requirement for stockholder-contributed capital
may be expressed as:
SC
x
GC
≥
−
.
A 100
A
(19)
Since stockholder-contributed capital ratios decline with bank size, the consolidation of
banking assets into very large institutions serves to economize on the amount of private
capital in the industry. Unless small banks’ leverage disadvantage is offset by
advantageous tax and reserve-requirement treatment, TBTDA implies that stockholders
in small and medium-size banks can create predictable value for themselves by planning
to sell out to any very large institution. We have denoted this value by k and assumed it
is imbedded in MC(t-1).
Model (19) implies the testable hypothesis that lower capital requirements and
funding costs for TBTDA institutions in the U.S. encourage megamergers in banking and
could influence the amount and distribution of the net value each megamerger creates.
When a post-merger institution ends up ranked more firmly among the largest financial
25
organizations in the country, the merger improves the institution’s access to de facto
guarantees of its uninsured debt. Other things equal, the existence of TBTDA benefits in
megamergers implies the hypothesis that the stock price of an acquiring megabank would
go up roughly in proportion to opportunities for the postmerger enterprise to increase its
leverage and to hold volatile portfolio positions. For stockholders of the two merging
institutions, an enhanced entitlement to capital forbearance should immediately increase
the capitalized value of implicit government guarantees of uninsured bank liabilities.
This implies that some of the unique integrative value a megamerger creates for acquirers
should pass through to holders of their uninsured obligations. Although our limited data
set does not allow us to investigate this phenomenon, it is consistent with the Fitch-IBCA
credit-rating agency’s 1999 decision to equalize the debt ratings of BHCs and their
subsidiary banks.
V. Conclusion
As do all event studies, this paper raises questions it cannot decisively answer. It
shows only that in the banking megamegers of 1991-98, stockholders of large-bank
acquirers gained value when a deposit-institution target was large and that acquirers
gained more value when a deposit-institution target was previously headquartered in the
same state.
Without extensive further testing, one cannot rule out any number of socially
comforting explanations for this result. That said, it must be recognized that our
megabank results differ from those for ordinary corporate mergers and for smaller
banking deals. The effect of size underscores the possibility that Too-Big-To-Discipline
subsidies have distorted dealmaking incentives for megabanks and that the high leverage
these firms maintain makes it easy for them to shift unaccounted risk onto taxpayers.
Governments everywhere have difficulty disciplining large, complex, and global
financial enterprises. This simple truth supports the unpleasant working hypothesis that
the banking megamergers of the 1990s may have emerged as a dialectical market
response to FDICIA’s effort to curtail TBTDA forbearances. On this argument,
megamergers hope to create institutions so large or complex that their creditors can
26
continue to count on qualifying for FDICIA’s systemic-risk exception as a matter of
course.
The increased accountability that the FDIC Improvement Act imparts to banking
regulators might have been enough to minimize TBTDA merger incentives in the 1980s.
In those bygone days, bank balance sheets were easier to monitor, interstate and crossindustry acquisitions were more strictly regulated, and the largest U.S. bank was much
smaller and less globally involved than it is today.
Several ways exist for regulatory economists who have fuller access to financialinstitution data to test more decisively for TBTDA effects. The first place to look is in
markets for uninsured bank and bank holding-company debt. The size of TBTDA
benefits can arguably be measured by analyzing premerger and postmerger spreads
between the market yields on acquired and target banks’direct debt and market yields on
the debt their parent holding companies issue. TBTDA benefits from megamergers
should produce significant announcement-day effects on the price of acquirer and target
debt. Support for this hypothesis is found in evidence assembled by Flannery and
Sorescu (1996) that during 1983-1991 conjectural government guarantees and bankspecific risks were impounded into the prices of the subordinated debentures of bank
holding companies.
It would also be useful to use accounting datasets and regression techniques to
separate TBTDA effects from putative measures of the monopoly rents that might be
associated with size. TBTDA benefits are apt to be revealed by postmerger surges in
leverage, uninsured liabilities, nonperforming loans, and other risk exposures. Monopoly
rents are apt to be associated with premerger market overlap and postmerger increases in
the share of credit-card business and control of trading, ATM, and other payments
networks.
27
FIGURE ONE
Premerger Economic Balance Sheets of an Acquirer and its Target
Acquiring Bank
TAA
TLA
IAA
ILA
ENWA
Target Bank
TAT
TLT
IAT
ILT
ENWT
FIGURE TWO
Pro Forma Economic Balance Sheet of Combined Firm
TAC (=TAA + TAT)
TLC (=TLA + TLT + BA)
IAC (=IAA + IAT + ∆IA)
ILC (=ILA + ILT - ∆IL)
TNWC (=TNWA+TNWI-BA = TAC – TLC – BA)
INWC [=INWA + INWT + (∆IA - ∆IL)]
28
FIGURE THREE
Worst Shortfalls in Post-Merger Accounting Performance Among Top 15 1997-1998
Banking Megamergers
Merger Partners
Shortfalls Re Earnings Target*
First Union – CoreStates
-23.8%
Washington Mutual – Great Western
-17.3%
NationsBank – BankAmerica
-16.8%
First American – Deposit Guaranty
-15.9%
Bank One – First Chicago
-14.9%
First Bank System – U.S. Bancorp
-10.9%
* Adapted from Oct. 12, 1999 American Banker presentation of Keefe, Bruyette &
Woods data.
29
Table 1
Dermine’s List of Economic Justifications for Bank Mergers
Cost-Based Economies of Scale
Brand-Based Economies of Scale
Revenue-Based Economies of Scale
Safety-Net-Based Economies of Scale
Revenue-Based Economies of Scope
X-Efficiency
Market Power
Managerial Agency Costs
• Pursuit of size to strengthen managerial entrenchment
• Pursuit of size-based increases in executive salaries
Source: adapted from Dermine (1999).
30
31
32
33
34
35
36
REFERENCES
Akhavein, Jalal D., Allen Berger, and David B. Humphrey, 1997. “The Effects of
Megamergers on Efficiency and Prices: Evidence form a Bank Profit Function,”
Review of Industrial Organization, 12, 95-139.
Amel, Dean F., 1997. “Antitrust Policy in Banking: Current Status and Future
Prospects,” Proceedings of the 33rd Annual Conference on Bank Structure and
Competition. Chicago: Federal Reserve Bank of Chicago, 166-179.
Angbazo, Lazarus, and Anthony Saunders, 1995. “Additional Evidence on the Value of
‘Too-Big-To-Fail’Guarantees,” New York University (Feb.).
Berger, Allen N., 1998. “The Efficiency Effects of Bank Mergers and Acquisition: A
Preliminary Look at the 1990s Data,” in Y. Amihud and G. Miller (eds.), Bank
Mergers & Acquisitions, 79-111.
Berger, Allen N, Rebecca S. Demsetz, and Philip E. Strahan, 1999. “The Consolidation
of the Financial Services Industry: Causes, Consequences, and Implications for
the Future,” Journal of Banking and Finance, 23(February), 135-194.
Berger, Allen N., Seth Bonime, Lawrence Goldberg, and Lawrence 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,” Washington: Board
of Governors of the Federal Reserve System (July).
Berger, Allen N. and David B. Humphrey, 1992. “Megamergers in Banking and the Use
of Cost Efficiency as an Antitrust Defense,” The Antitrust Bulletin, 37(Fall), 541600.
_____, “Measurement and Efficiency Issues in Commercial Banking,” Federal Reserve
Board Finance and Economics Discussion Series No. 151, December 1990.
Black, Harold A., M. Cary Collins, Breck L. Robinson, and Robert L. Schweitzer, 1997.
“Changes in Market Perception of Riskiness: The Case of Too-Big-To-Fail,”
Journal of Financial Research, 20(Fall), 389-406.
Boyd, John, and S. Graham, 1998. “Consolidation in U.S. Banking: Implications for
Efficiency and Risk,” in Y. Amihud and G. Miller (ed.), Bank Mergers &
Acquisitions. Norwell, MA: Kluwer Academic Publishers, 113-135.
37
Calomiris, Charles W., and Jason Karceski, 1998. Is the Bank Merger Wave of the 1990s
Efficient?, Washington: American Enterprise Institute for Public Policy Research.
Clark, Jeffrey A., “Economies of Scale and Scope at Depository Financial Institutions: A
Review of the Literature,” Federal Reserve Bank of Kansas City Economic
Review, 73, September/October 1988.
Cornett, Marcia M., and Hassan Tehranian, 1992. “Changes in Corporate Performance
Associated with Bank Acquisitions,” Journal of Financial Economics, 31(April),
211-234.
Cornett, Marcia M., Gayane Hovakimian, Darius Palia, and Hassan Tehranian, 1999.
“The Impact of Manager-Shareholder Conflict on Acquiring Bank Returns,”
unpublished working paper, Boston College.
Cyree, Ken B. and Ramon P. DeGennaro, 1997. “Wealth Effects for Acquiring Firms in
the Banking Industry: Evidence from a New Approach,” University of Tennessee
(May).
DeLong, G.L., 1999a. “What do Investors Look for When U.S. Banks Announce
Mergers?,” working paper (City University of New York: New York, NY).
DeLong, G.L., 1999b. “Domestic and International Bank Mergers: Stockholder Gains
from Focusing versus Diversifying,” working paper (City University of New
York: New York, NY).
Dermine, Jean, 1999. The Economics of bank Mergers in the European Union: A
Review of the Public Policy Issues, Report Commissioned by Dutch Ministry of
Finance, Fontainebleau: INSEAD.
Financial Economists Roundtable, 1999. Statement on Long-Term Capital Management
and the Report of the President’s Working Group on Financial Markets, Chicago
(September).
Flannery, Mark, 1998. “Modernizing Financial Regulation (Again),” Prepared for a Sept.
17-18 Conference on Financial Modernization and Regulation, Jointly sponsored
by the Federal Reserve banks of Atlanta and San Francisco.
Flannery, Mark, and Sorin M. Sorescu, 1996. “Evidence of Market Discipline in
Subordinated Debenture Yields, 1983-1991,” Journal of Finance, 51(September),
1347-1377.
38
Gorton, Gary and Richard Rosen, 1995. “Corporate Control, Portfolio Choice, and the
Decline of Banking,” Journal of Finance, 50(December), 1377-1420.
Hoenig, Thomas M., 1999. “Financial Industry Megamergers and Policy Challenges,”
Federal Reserve Bank of Kansas City, Economic Review, 84(third quarter), 7-13.
Houston, Joel F., and Michael D. Ryngaert, 1994. “The Overall Gains from Large Bank
Mergers,” Journal of Banking and Finance, 18(December), 1155-1176.
Houston, Joel F., Christopher M. James, and Michael D. Ryngaert, 1999. “Why are
Value Enhancing Mergers in Banking So Hard to Find?,” Gainesville Department
of Finance, Insurance and Real Estate, University of Florida.
James, Christopher M., and Peggy Wier, 1987. “Returns to Acquirers and Competition in
the Acquisition Market: The Case of Banking,” Journal of Political Economy,
95(April), 355-370.
Kane, Edward J., 1991. “Taxpayer Losses in Deposit Insurance: The Incentive-Conflict
Hypothesis,” in Richard Herring (ed.), Reforming the American Banking System,
Philadelphia: Wharton Financial Institutions Center.
_____, 1995. “Three Paradigms for the Role of Capitalization Requirements in Insured
Financial Institutions,” Journal of Banking and Finance, 119(June), 431-459.
Lang, Larry, Rene M. Stulz, and Ralph A. Walkling, 1999. “A Test of the Free cash
Flow Hypothesis: The Case of Bidder Returns,” Unpublished working paper,
Ohio State University.
Nickerson, Cynthia A., 1998. “To Merge or Not to Merge: An Analysis of Second
Quarter M&A Activity,” Banking Policy Report, 17 (August 3), 1 & 12-16.
Noulas, Athanasios G., Subhash Ray and Stephen Miller, “Returns to Scale and Input
Substitution for Large Banks,” Journal of Money, Credit, and Banking, 22,
February 1990.
O’Hara, Maureen, and W. Shaw, 1990. “Deposit Insurance and Wealth Effects: The
Value of Being ‘Too Big to Fail’,” Journal of Finance, 45 (Dec.), 1587-1600.
Padgett, Tania, 1999. “Big Mergers Yielding Profit Let Downs, Study Finds,” American
Banker, 164(Oct. 12), 1 & 37.
39
Prager, R.A., and T.H. Hannan, 1998. “Do Substantial Horizontal Mergers Generate
Significant Price Effects?: Evidence from the Banking Industry,” Journal of
Industrial Economics, 46, 433-452.
Seidman, William, 1986. Presentation by FDIC Chairman to the National Press Club,
Washington, DC, (October 16).
Todd, Walker F., and James B. Thomson, 1990. “An Insider’s View of the Political
Economy of the Too Big to Fail Doctrine,” Federal Reserve Bank of Cleveland
Working Paper No. 9017.
Williams, Joseph T., 1993. “Equilibrium and Options on Real Assets,” Review of
Financial Studies, 6(Winter), 825-850.
Wilmarth, Arthur E., Jr., 1995. “Too Good to Be True? The Unfulfilled Promises Behind
Big Bank Mergers,” Stanford Journal of Law, Business & Finance, 2(Fall), 1-88.
40
41
42
43
44
*
The author is grateful to Tara Rice and Debarshi Nandy for outstanding research assistance and to Marcia
Cornett, Ramon De Gennaro, Ron Feldman, Mark Flannery, Stephen Kane, James Moser, Evren Ors,
George Pennacchi, Hassan Tehranian, James Thomson, and an anonymous referee for valuable comments
on earlier drafts.
i
Occam’s Razor is the philosophical principle that assumptions should not be multiplied unless they are
necessary to improve an explanation.
ii
Actually, speculators would expect some of the incremental value by which the ki of the best potential
acquirer i exceeds the kj of the second-best acquirer j to be held out of i’s bid. This idea is important in
interpreting event-study returns intuitively, but we fear that putting it into these equations would obscure
the substantive thrust of the argument.
iiiiii
Without extending our data set backward in time, we cannot meaningfully test for a FDICIA-induced
regime shift in our regression models for XA and XW.
45