Insider trading around bank failures.
Jalbert, Terrance J. ; Rao, Ramesh P. ; Bathala, Chenchuramariah T. 等
INTRODUCTION
In this paper we examine insider trading around the failure of
banks. Between 1980 and 1994, 1,617 banks failed in the United States.
This number constitutes 9.14 percent of all U.S. chartered banks. On a
total asset basis, the failed banks held 8.98 percent of the total
banking system assets (FDIC, 1997). While no geographic area of the
United States was immune from the failures, some were hit harder than
others. Nearly 60 percent of the failures occurred in 5 states: Texas,
Oklahoma, California, Kansas and Louisiana (FDIC, 1997). Despite the
economic significance of these bank failures, to date no paper has
examined insider trading behavior around bank failures. We examine
insider trading around four largest commercial bank failures that took
place in the 1980's. The four failures that are examined are
Continental Illinois National Bank and Trust Company, First City
Bancorp., First Republic Bank and Mbank.
Given the lack of evidence regarding insider trading for banks and
other financial institutions, we will discuss the empirical literature
on insider trading behavior in non-financial firms. These studies have
attempted to answer three important questions: (1) Is insider trading
beneficial to capital markets? (2) Do insiders earn excess returns?, and
(3) What kind of trading patterns do insiders exhibit relative to
significant corporate news announcements? While all of these issues are
interesting, our focus is on answering the third question for the
banking industry.
Elliott, Morse and Richardson (1984) found evidence that insiders
with private information purchased stock in their firms before value
increasing announcements and sold stock before value decreasing
information releases. Oppenheimer and Dielman (1988) examined insider
trading patterns during the twelve months prior to announcements of
dividend resumption (omission) and possible abnormal returns associated
with such insider trading. They found evidence suggesting that insiders
engaged in extensive net buying (selling) activity prior to dividend
resumption announcements (omission). However, these pre-announcement
insiders purchases failed to consistently earn excess returns, while
pre-announcement insider sales enabled the insider to avoid negative
abnormal returns. Furthermore, officers (high-information insiders) were
found to earn larger profits than all insiders considered as a broad
group.
The literature relating to non-financial firms indicates that the
announcements of voluntary liquidations tend to benefit stockholders,
whereas the announcement of filing for bankruptcy reduces equity prices.
On that basis, Eysell (1991) hypothesized that the announcement of
voluntary liquidations would be associated with an increase in insider
purchases and that announcements concerning bankruptcy filings would
lead to increased insider sales. As expected, when he empirically tested
these hypotheses he found that corporate insiders were in fact heavy
purchasers prior to liquidation announcements and during the period when
the value of firms' stock kept rising. Furthermore, as expected,
insiders of firms that eventually filed for bankruptcy were heavy net
sellers. Additionally, those with the greatest access to information
(high-information insiders) were found to be the heaviest purchasers
before liquidation announcements and the heaviest sellers before
bankruptcy announcements. These findings indicate that insider trading
is prevalent in a manner consistent with the exploitation of private
information. In a recent article, Hu and Noe (2001) derive the
conditions which allow insiders to trade on their on their own behalf
and in a manner which increases shareholder wealth.
Karpoff and Lee (1991) examined the insider trading patterns before
new equity offerings. The authors found that insiders are net sellers of
their firms' common stock prior to the announcements of new common
stock and convertible debt issues. On that basis, they concluded that
penalties for illegal insider trading did not deter insiders from using
their superior information. Teal (1993) examined insider holdings, as
opposed to insider trading, around the failure of savings and loan associations (S&Ls) during the period from June 1977 to March 1991.
Teal argues that managers maintain a dual role. In the management role,
the manager strives to maximize the value of his salary. In the
shareholder role the manager strives to maximize his equity value. He
suggests that these objectives often conflict, and consequently, we
should expect managers with higher shareholdings to act on behalf of
shareholders in general by assuming greater risk in order to increase
the value of the deposit insurance option. Consistent with this, he
found evidence that the insiders of S&Ls which subsequently failed
owned 23.1% of the outstanding equity, compared to only 11.0% for
S&Ls that did not fail. Bank failure announcements are generally
found to have a negative effect on the stock prices of the failing banks
as well as other commercial banks. For example, using weekly data Swary
(1986) found that all banks experienced a 3% decline in the value of
their common stock in reaction to the Continental Illinois failure. In a
similar fashion, Lamy and Thompson (1986) show that all banks
experienced a 1% decline in stock value in reaction to the Penn Central
failure.
In this paper we examine insider trading around the failure of
financial institutions. We examine insider trading around four major
bank failures that took place in the 1980's. We begin by discussing
the theory related to bank contagion and insider trading. We continue by
proposing testable hypotheses and discussing the data and methodology.
Next, the test results are presented followed by concluding comments.
THEORY RELATED TO BANK CONTAGION AND INSIDER TRADING
Banks are unique in that a significant proportion of their
liabilities are payable at par on demand or at very short maturities. To
meet these deposit withdrawals the bank may have to quickly liquidate
assets at current market value. While banks are required to hold a small
percentage of reserves against transaction or demand deposits, these
reserves are not sufficient to meet a massive withdrawal of deposits. To
help protect again such a "bank run" the FDIC insures bank
deposits up to $100,000. This deposit guarantee, plus the fact that
historically the FDIC charged a flat-rate insurance premium regardless
of bank risk, encouraged banks to assume increased levels of risk. This
moral hazard condition has been well documented in the literature (Kane,
1986). In effect, bank management has been encouraged to take excessive
risks which, if successful, benefits bank shareholders as well as bank
management. If the risky investments are unsuccessful and the solvency of the bank should be in jeopardy, the FDIC and tax payers will
ultimately bare the financial burden. In addition, to prevent a massive
run on the banking system, the Federal Reserve and the FDIC instituted a
"too-big-to-fail" doctrine which lead to the protection of all
liability holders at Continental Illinois banks regardless of amount or
type of deposit or indebtedness. The argument put forth at that time was
that, should Continental Bank fail, hundreds of other banks with
correspondent banking relationship with Continental might also fail.
This could then lead to widespread panic or contagion within the banking
system that could eventually migrate to non-bank financial institutions.
The size of the FDIC insurance fund was felt to be inadequate to deal
with such a massive bank run.
Aharony and Swary (1983) make a distinction between pure or
industry-specific contagion and firm or bank specific contagion. With
industry-specific contagion the market has difficulty differentiating
between individual banks and views the banking industry as more or less
homogeneous. Given the difficulty that the market has in evaluating the
quality of bank assets, in particular private loan contracts, problems
at one bank are viewed as a likely problem for all banks and may illicit an irrational response toward the entire industry. With bank-specific
contagion an adverse event at one bank is seen as being indicative of
problems at other banks which share common characteristics, such as
similar types of loan exposures or common product lines. In this case
the market reaction is rational in the sense that bank performance is
correlated to a common set of industry or macro economic variables.
Industry-specific contagion is viewed as being more damaging than
bank-specific contagion as it potentially impacts all banks, rather than
shifting deposits and loans within the system from weak to healthy
institutions. Furthermore, as Kaufman (1994) points out, the news may
not be all bad since the failure of a major bank may significantly
benefit its competitors. This was certainly the case for First Chicago
Bank, the primary competitor of Continental Bank. In reviewing the
evidence surrounding a number of major bank failures, Kaufman concludes
that bank-specific rather than industry-specific contagion is more
likely.
In a recent paper, Bougheas (1999) develops a model of bank
contagion and concludes that bank failure alone is not sufficient to
induce a "bank run" but must be accompanied by periods of
economic instability. In terms of international bank failures, Jayanti,
Whyte, and Do (1996) examine several bank failures in Britain and Canada and find no evidence of industry-specific contagion in Britain, while
Canadian banks reacted adversely to several domestic bank failures. The
authors conclude that the Canadian stock market primarily may have
reacted to new bank regulations proposed immediately following the
failures and that deposit insurance by itself may not be sufficient to
prevent contagion effects.
Overall, the majority of evidence suggests that bank managers have
little to worry about industry-wide contagion effects and may have an
incentive to act on insider information relating to an impending bank
failure. This inside information may encourage them to sells shares of
their own stock prior to a bank failure if they perceive that
bank-specific contagion will adversely impact their bank. Conversely,
the impending failure may provide an incentive to purchase shares based
upon inside information if the failed bank is a major a competitor. As
in the case of First Chicago and Continental Bank, it seems likely that
in certain cases both effects could be operative. Which effect dominates
and leads to either insider net sales or purchases is an empirical
question and is the central focus of this research.
TESTABLE HYPOTHESES
We argue that insiders have better information than other market
participants regarding both the timing and likely effect of bank
failures. Inter-bank payments and credit transactions, plus informal
contacts allow insiders in one bank to be aware of problems occurring in
other institutions. This is private information that generally would not
be available to other market participants during the same time frame.
Furthermore, insiders in healthy banks are in a better position to
evaluate both the extent and ultimate consequences of the problems at
troubled institutions and determine more accurately how their bank will
be affected. Thus, its seems likely that insiders of non-failed banks
might view the consequences of another bank failing quite differently
than the general market. For example, if the general market has trouble
differentiating between banks, a major bank failure might be perceived
as a negative announcement impacting all commercial banks. On the other
hand, insiders at the non-failed bank may view the failure of a
competitor as an opportunity to obtain additional business. Based on
this private information, in comparison to less informed investors, we
expect that bank insiders will have an incentive to trade their own
shares based on their perception of the impact of the impending failure
of other another bank. An increase in insider purchases indicates that
insiders view the failure as having a net positive impact. Insiders
would treat the reduction in price caused by other market participants
trading as an opportunity to purchase shares at a bargain. A reduced
level of purchases or net selling by insiders would suggest that
insiders are less optimistic about the impact of the failure on their
bank than are other market participants. In this case, insiders would
view the lower than expected decline in price as an opportunity to sell
their shares at a high price.
THE DATA AND METHODOLOGY
The four largest bank failures that took place between 1982 and
1989 are identified from the annual reports of the FDIC on failed
institutions. We selected these four failures because they were markedly
larger than other failures that took place during that time period.
First City, the smallest of the four banks examined, was nearly twice as
large as the next largest failed commercial bank (FDIC, 2002). Because
of their size, a more pronounced insider reaction is expected within the
failed bank as well as in non-failed banks around the time of the
failure. The four banks and selected information about the banks are
provided in Table 1 (FDIC, 2002). In order to analyze the insider
trading in other banks, a listing of national commercial banks is
obtained from Standard and Poor's Compustat Data Files. All banks
denoted as a national commercial bank, commercial bank, or commercial
bank - not chartered are selected. Data regarding the open market trades
of insiders in these banks is obtained from the Ownership Reporting
System Master File compiled by the US Securities and Exchange Commission
(SEC). Insider trading data is collected for the three years prior and
the one year subsequent to each failure.
More specifically, we examine insider trading in the time period
surrounding the failure. In any event study, it is necessary to define
and event period. Loh and Rathinasamy (1995) select a six-month event
period in their study of insider trading around dual class
recapitalizations. Gosnell, Keown, and Pinkerton (1992) find that the
level of insider selling increases over the final 5 months leading to
the first public bankruptcy announcement of OTC firms. Sanders and
Zdanowicz (1992) examine takeovers and find that target firm insiders
know of impending takeovers as much as 542 days and on average 79.1 days
prior to the first public announcement. In this paper, the
"pre-event" period is defined as the four quarters prior to
the failure (Quarters -1 to -4). The "post-event" period is
defined as the four quarters subsequent to the event period (Quarters +1
to +4). In order to estimate how the failure has impacted the direction
and extent of insider trading, an eight-quarter insider trading
"comparison period" is identified which includes the two years
just prior to the beginning of the pre-event period. These timeframes
are selected to be consistent with the previous literature as well as to
best utilize a limited amount of data available. While we use these
timeframes, tests are conducted in a quarterly fashion, thereby
permitting the identification of anomalies that occur at other time
intervals. Figure 1 illustrates these definitions:
[FIGURE 1 OMITTED]
In order to differentiate the behavior of large volume insiders
from small volume insiders, two measures of insider behavior are
incorporated into the analysis. The "number of transactions"
approach for measuring insider behavior is based on the number of trades
that insiders made, regardless of size. Of interest here is the
proportion of insider trades that were sales. On the other hand, the
"shares traded" approach is based on the actual number of
shares traded. Of interest is the net number of shares traded by
insiders that were sales. This combination of measures is similar to
that of Loh and Rathinasamy (1995) and Hirschey and Zaima (1989). Other
measures have been suggested by Penman (1985), however, he finds no
appreciable difference in the predictive ability of the measures.
THE EMPIRICAL RESULTS
The first step in the analysis is to compare the number of
individual insider sales transactions as well as the total number of
shares traded in each of the eight quarters surrounding the bank failure
date to the number of insider sales transactions in the comparison
period. In order to perform the analysis a dummy variable is assigned to
each trade indicating whether the transaction represents a purchase or a
sale. Trades that involve a sale of shares are assigned a value of
1,while trades that involved a purchase of shares are assigned a value
of 0. A t-test is used to determine if the proportion of sales
transactions for each of the eight event quarters is significantly
different from the proportion of sales transactions during the
comparison period. Given the small number of transactions in certain
quarters, a similar test is performed for both one-year pre- and post-
event periods.
Consistent with the literature which indicates that geographic
proximity and differences (or similarities) in bank regulation may play
an important role in insider behavior, each of the above mentioned tests
are performed on the following two test samples and reported in Table 2.
Panel A in Table 2 reports the results of a sample of non-failed banks
located outside the state in which the failed bank is located. Panel B
reports the findings for the sample of non-failed banks located within
the same state as the failed bank.
The series of t-tests discussed above are then estimated for each
of the two samples and for each of the four bank failures. Thus a total
of 8 sets of statistical tests are made. The first two of the following
four analyses compare the proportion of insider sales transactions and
the average net size of these transactions between various groups of
non-failed banks during various time periods. The final two analyses
compare the proportion of insider sales transactions and the average net
size of these transactions between the set of failed and non-failed
banks during the same time periods.
Proportion of Sales Transactions: Non-failed Bank Comparisons
As mentioned above, the results of the first set of tests based
upon the proportion of insider sales transactions or trades are
presented in Table 2. In each of the two panels, insider trading at
non-failed banks in various geographic categories (e.g., out-of-state
and within-state) is being compared at different points in time as
measured by the four bank failure event dates. In Panel A of Table 2,
the results of the test on the out-of-state sample of non-failed banks
are presented. To illustrate, examining the Continental Illinois bank
failure, during the comparison period there were a total of 4,408
insider trades reported nationwide and 27.86% of them were insider
sales. On the other hand, during quarter +1 of the post-event period
there were a total of 749 insider trades of which approximately 20.96%
were sales. This reduction in the proportion of sales suggests that on
an out-of-state basis, insiders had become more optimistic about the
effects of the Continental failure than other market participants, hence
the reduction in share sales. A t-test on the difference in these
proportions is significant at the one percent level. This procedure is
then repeated for each of the remaining seven pre- and post- event
quarters as well as for the one-year pre- and post-failure periods. The
latter results indicate a reduction in the proportion of sales in both
periods relative to the comparison period but a slight recovery during
the post-event period.
In general, similar reductions in the proportion of insider sales
are found in other event quarters surrounding the Continental Bank
failure. Comparable results are presented for First Republic and First
City banks. However, during two quarters both before and following the
failure of Mbank, bank insiders nationwide increased the proportion of
times they sold shares, suggesting that insiders were less optimistic
about the effects of the Mbank failure on their own bank than was true
for the other three bank failures. Overall, the evidence suggests that
insiders in non-failing banks changed their trading behavior around bank
failures, although the direction of the change is not identical for each
failure.
In Panel B, the in-state non-failed banks test results are
presented. Given that the number of insider trades is quite small for
certain quarters, the aggregate one-year pre- and post- event
comparisons are undoubtedly more reliable. The results for Continental
Bank indicate a significant increase in the proportion of net sales
during the one-year post-event period. For First Republic Bank the
results are similar during both the pre- and post-event period, with a
modest decrease in the proportion of net sales. On the other hand, Mbank
reported a substantial decline during the post-failure period. First
City Bank reported a dramatic drop in the proportion of net sales during
the pre-failure period, followed by a significant recovery during the
post-failure period. Thus, once again there is evidence of a change in
trading behavior around the time of a bank failure.
Number of Share Transacted: Non-Failed Bank Comparisons
In this analysis we test for differences in the average number of
shares traded. We do this by creating an index of purchases/sales. The
index is created by multiplying the number of shares traded by +1 if the
transaction is a purchase and by -1 if the transaction is a sale. Thus a
purchase of 100 shares is given an index value of +100 and a sale of
shares is given a index value of -100. A positive index value indicates
that in the aggregate share purchases exceeded share sales. A negative
total indicates the opposite. As before, a t-test is performed to
determine if there is a significant difference between the average size
of net sales or purchases between the comparison period and the each
individual event quarter, as well as over the one-year pre- and
post-event periods.
The test results for the average net number of shares traded are
presented in Table 3. Generally, the tests show little change in insider
trading around the time of the failures for the out-of-state sample of
non-failed banks. However, there is evidence to suggest a change in
insider trading patterns among in-state non-failed banks. This result is
perhaps not surprising, since one would expect that in-state banks would
be affected more by a failure than out-of state banks. For example, in
the case of Continental Illinois the average number of shares for
in-state bank trades was +1,404, indicating that the number of purchases
exceeded the number of sales. During the four pre-event quarters (-1 to
- 4) the balance starts to shift in the direction of either an increase
in insider sales or a large reduction in the size of net purchases.
During the four post-event quarters (+1 to + 4) the average net number
of shares traded is now negative for quarters 2,3 and 4, indicating a
significant shift towards insider selling. A comparison of the one-year
pre- and post-failure periods dramatically illustrates these results as
of the number of net shares purchased declines from approximately 519 to
91. An even more drastic shift takes place for First Republic and First
City as large net purchase positions during the comparison period shift
towards sizeable net sale positions during the pre- and post-periods. At
the same time, there is a noticeable reduction in net sales during the
post-event period relative to the pre-event period.
The reader will notice a marked difference between the results from
the proportion of insider sales and the average number shares traded
approaches. Recall that the proportion of insider sales treats each
trade equally, regardless of the number of shares involved. The average
net number of shares traded assigns those transactions that involve a
larger number of shares more importance in the analysis. Thus, the
differing results are an indication that insiders behave differently
based upon the volume of shares that they trade.
Proportion of Transactions Approach: Failed Bank vs. Non-Failed
Bank Comparisons
The comparisons presented in Table 4 employs the proportion of
sales transactions approach in a manner analogous to the tests performed
in Table 2 except that we are now comparing failed banks with non-failed
banks. We perform significance tests for each of the four failing banks.
Due to the limited amount of available data and the close time proximity
of the three Texas bank failures, we combine the Texas bank failures
into a single sample (labeled ALLTX). The event day for the combined
Texas sample is the time period from April 20, 1988 (the date that First
City Bank failed) to March 29, 1989 (the date that Mbank failed). The
pre-event quarters and comparison period are defined relative to April
20,1988, while the post-event quarters are defined relative to March
29,1989. (Note: because of the sampling procedures employed the total
number of observations in the three Texas failures may not equal the
total number of observations in the combined test).
The values for N1, N2, PS1, and PS2 are found in Table 2. Table 4
simply tests for a statistically significant difference between PS1 and
PS2, using N1 and N2 degrees of freedom. In Panel A, insider trading in
the failed bank is compared to insider trading in non-failed banks
nationwide. In the comparison period and in both the pre- and post-
event periods, Continental Illinois consistently had a higher proportion
of insider sales compared to the national proportion (the once exception
being quarter - 4). For example, in the comparison period, the
proportions were 50.7% and 27.7% for Continental and the sample of
nationwide banks, respectively. The proportions of sales for failed and
non-failed banks for the one-year pre-event period are 34.9% and 22.9%,
respectively. For the post-event period the proportions are 47.8% and
23.5, respectively. In general, insiders at the failed banks were more
frequent sellers than insiders at non-failed institutions.
The trading differences have a greater degree of statistical
significance during the post- event period, suggesting that the
dominance of insider selling increased somewhat after the failure event.
In Panel B, trading in the failed bank is compared to trading in
non-failed banks in the same state. The results are somewhat similar
indicating that insiders at Continental were more likely to be selling
share compared to insiders at in-state non-failed banks, although the
results are not as statistically consistent. A similar set of results
for the combined set of failed Texas banks is observed but here the
trading differences are statistically significant.
Average Number of Shares Approach: Failed Bank vs. Non-Failed Bank
Comparisons
The final comparison presented in Table 5 is based on the average
number of shares traded by insiders in a manner analogous to the tests
performed for Table 3. In the case of Continental Bank for both the
comparison period and quarter -4 the average number of shares sold per
trade was positive indicating that the number of purchases exceeded the
number of sales. The reverse was true for the nationwide sample of
non-failed banks for the comparison period. During the remaining three
pre- event periods the average number of shares sold turns negative for
Continental indicating that insider sales were growing. The same was
true of Continental during the post- event period although the results
were generally not statistically significant. For the entire pre-event
period, the average net number of shares (66) is close to zero. Compared
to the post-event period where the average net number of shares traded
(-605) is decidedly negative. For the Texas banks and general and Mbank
in particular, the average number of net sales increases dramatically
between the one-year pre- and post-event periods.
CONCLUSIONS
In this paper we examine insider trading around the failure of
financial institutions. We examine insider trading around four major
bank failures that took place in the 1980's. The four failures that
are examined are Continental Illinois National Bank and Trust Company,
First City Bankcorp., First Republic Bank, and Mbank. We find that bank
insiders changed their trading behavior in the time period surrounding
these large bank failure announcements relative to a comparison time
period. Insider trading patterns suggest that bank insiders are
generally more optimistic than other insiders around the time of large
bank failures. We find that large volume insiders behave differently
than small volume insiders. Small volume traders appear to react more
strongly to the bank failure than large volume insiders. We find that
insiders in banks which are located in the same state as the failing
bank behave differently than insiders in out-of-state banks around the
time of a failure. Finally, we find that insiders in failing banks trade
differently than insiders in non-failed banks. Insiders in the failed
bank are larger sellers of their banks stock around the time of the
failure than insiders in other banks.
REFERENCES
Aharony, J. & I. Swary. (1983). Contagion effects of bank
failures: evidence from capital markets. Journal of Business, 56,
305-322.
Bougheas, S (1999). Contagious bank runs. International Review of
Economics and Finance 8, 131-146.
Jayanti, S.V., A. M. Whyte & A. Q. Do. (1996). Bank failures
and contagion effects: evidence from Britain and Canada. Journal of
Economics and Business 48, 103-116.
Elliott, J., D. Morse & G. Richardson. (1994). The association
between insider trading and information announcements. Rand Journal of
Economics 15(4), 521-536.
Eysell, T. H. (1991). Corporate insiders and the death of the firm:
Evidence on the incidence of insider trading in corporate
dissolution's. Financial Review 26(4), 517-533.
FDIC. (2002). Historical statistics on banking, bank and thrift failure reports. Retrieved March 15, 2002
from:http://www2.fdic.gov/hsob/SelectRpt.asp?EntryTyp=30.
FDIC Division of Research and Statistics. (1997). History of the
eighty's--lessons for the future: volume I: an examination of the
banking crises of the 1980s and early 1990s, 14-15.
Gosnell, T., A. J. Keown & J. M. Pinkerton. (1992). Bankruptcy
and insider trading: differences between exchange-listed and OTC firms.
Journal of Finance, 47(1), 349-362.
Hirschey, M. & J.K. Zaima. (1989). Insider trading, ownership
structure, and the market assessment of corporate selloffs. Journal of
Finance 44, 401-438.
Hu, J. & T. Noe. (2001). Insider trading and managerial
incentives. Journal of Banking and Finance 25 (4), 681-716.
Kane, E. (1986). Appearance and reality in deposit insurance: The
case for reform. Journal of Banking and Finance 10 (2), 175-188.
Karpoff, J.M. & D. Lee. (1991). Insider trading before new
issue announcements. Financial Management 20(1), 18-26.
Kaufman, G. (1994). Bank contagion: A review of the theory and
evidence. Journal of Financial Services Research 13(2), 123-150.
Lamy, R. & R. G. Thompson. (1986). Penn square, problem loans,
and insolvency risk. Journal of Financial Research 9(2), 103-111.
Loh, C. & R.S. Rathinasamy. (1995). Insider trading and dual
recapitalization. Journal of Financial and Strategic Decisions, 8(3),
61-68.
Oppenheimer, H.R. & T. E. Dielman. (1988). Firm dividend policy
and insider activity: Some empirical results. Journal of Business
Finance and Accounting 15(4), 525-541.
Penman, S.H. (1985). A comparison of the information content of
insider trading and management earnings forecasts. Journal of Financial
and Quantitative Analysis 20, 1-17.
Sanders, R. W. & J. S. Zdanowicz. (1992). Target firm abnormal
returns and trading volume around the initiation of change in control
transactions. Journal of Financial and Quantitative Analysis 27,
109-129.
Swary, I. (1986). Stock price reaction to regulatory action in the
Continental Illinois crises. Journal of Business 59(3), 451-473.
Teal, J. L. (1993). Shareholder control and financial distress in
the thrift industry. Journal of Business Research 26, 161-170.
Terrance J. Jalbert, University of Hawaii at Hilo
Ramesh P. Rao, Oklahoma State University
Chenchuramariah T. Bathala, Cleveland State University
Alan Reichert, Cleveland State University
Table 1: Bank Failure Summary Information
Bank State Data of Failure
Continental Illinois IL July 26, 1984
First City Bankcorp TX April 20, 1988
First RepublicBank TX July 29, 1988
Mbank TX March 29, 1989
Bank Total Deposits Total Assets
Continental Illinois 28,580,214,000 39,956,956,000
First City Bankcorp 9,275,170,000 12,374,490,000
First RepublicBank 19,946,416,000 31,277,123,000
Mbank 11,937,990,000 15,640,651,000
Bank Failure Type
Continental Illinois Assistance Transaction
First City Bankcorp Purchase and Assumption
First RepublicBank Purchase and Assumption
Mbank Assistance Transaction
This table shows the four largest bank failures that occurred
between 1982 and 1989 along with summary statistics
about the failed bank.
Table 2: Analysis of Insider Trading in Event Quarters relative to
Insider Trading in the Comparison Period Using the Transactions
Methodology
Panel A: Out of State Non-Failed banks
Q C. I. F. R. M. B. F. C.
CP N 4408 5556 7676 4633
PS 0.2786 0.2527 0.1998 0.2614
-4 N 648 1226 881 915
PS 0.2531 0.1631 0.2111 0.1956
T 1.39 7.43 *** -0.78 4.50 ***
-3 N 659 1365 757 926
PS 0.2352 0.1055 0.2483 0.2192
T 2.43 ** 14.49 *** -2.96 *** 2.80 ***
-2 N 754 961 914 1525
PS 0.2599 0.2206 0.2495 0.0852
T 1.07 2.2 ** 3.30 *** 18.28 ***
-1 N 728 851 962 1006
PS 0.1676 0.2009 0.1954 0.2018
T 7.20 *** 3.47 *** 0.33 4.19 ***
-1to-4 2790 4409 0.356 4443
PS 0.2283 0.1649 0.2242 0.1636
T 4.82 *** 10.87 *** 2.91 *** 11.48 ***
1 to 4 2936 3525 4302 3521
PS 0.2333 0.2267 0.1934 0.2247
T 4.39 *** 1.28 0.85 3.85 ***
1 N 749 762 884 902
PS 0.2096 0.2848 0.276 0.1929
T 4.22 *** -1.85 * -4.85 *** 4.68 ***
2 N 640 1055 870 735
PS 0.2109 0.2 0.246 0.2667
T 3.87 *** 3.87 *** -3.01*** -0.3
3 N 868 823 1069 999
PS 0.2465 0.2272 0.1637 0.2272
T 1.99 ** 1.62 2.96 *** 2.32 **
4 N 679 885 1479 885
PS 0.2636 0.2644 0.1346 0.2192
T 0.82 -0.73 6.54 *** 2.75 ***
Panel B: In-State Non Failed Banks
Q C. I. F. R. M. B. F. C.
CP N 78 184 175 173
PS 0.1795 0.4325 0.4286 0.5087
-4 N 13 17 14 8
PS 0.6923 0.2353 0.5714 0.5
T -3.66 *** 2.41 ** -1.00 0.04
-3 N 11 31 15 19
PS 0 0..3226 0.4667 0.2632
T 4.10 *** 1.97 * -0.28 2.22 **
-2 N 8 14 27 32
PS 0.125 0.3571 0.5185 0.2813
T 0.41 1.08 -0.86 2.55 **
-1 N 12 12 15 16
PS 0.1667 0.6667 0.2667 0.375
T 0.11 -1.10 1.31 1.02
-1to-4 44 74 71 76
PS 0.2727 0.3649 0.4648 0.3158
T -1.15 2.09 ** -0.51 2.93 ***
1 to 4 53 74 61 72
PS 0.3208 0.3784 0.2951 0.4306
T -1.81* 1.88 * 1.91 * 1.12
1 N 18 19 16 9
PS 0.2222 0.5789 0.375 0.5556
T -0.39 -0.6 0.41 -0.26
2 N 7 21 14 22
PS 0.4286 0.381 0.2143 0.6364
T -1.2 1.09 1.79 * -1.14
3 N 21 26 20 19
PS 0.2857 0.2692 0.45 0.3684
T -0.97 2.46 ** -0.18 1.17
4 N 7 8 11 22
PS 0.5714 0.25 0 0.2273
T -1.9 1.52 11.42 *** 2.84 ***
Table 3: Analysis of Insider Trading in Event Quarters relative
to Insider Trading in the Comparison Period
Using the Shares Traded Methodology
Panel A: Out-of-State Non-Failed Banks
Q C. I. F. R. M. B. F. C.
CP N 4408 5556 7676 4633
SH -2337 1427.2 60.19 1671.8
-4 N 648 1226 881 915
SH 265.15 97.4 157.46 219.11
T -0.31 1.03 -1.38 0.94
-3 N 659 1365 757 926
SH 541.04 293.63 99.64 56.4
T -0.34 0.88 -0.5 1.04
-2 N 754 961 914 1525
SH 527.96 95.06 61.6 230.24
T -0.34 1.03 -0.04 0.93
-1 N 728 851 962 1006
SH 927.7 119.08 255.35 191.12
T -0.38 1.01 -0.94 0.95
-1to-4 2790 4409 3560 4443
SH 574.15 161.74 145.52 180.72
T -0.34 0.98 -1.32 0.96
1to4 2936 3525 4302 3521
SH 3527.2 513.11 662.48 249.17
T -0.68 0.69 -2.33 ** 0.92
1 N 749 762 884 902
SH 2818.2 70.12 1633.6 130.31
T -0.59 1.05 -1.34 0.99
2 N 640 1055 870 735
SH 2205.8 234.35 106.99 83.47
T -0.52 0.91 -0.45 1.02
3 N 868 823 1069 999
SH 7632.4 1778.4 722.92 48.46
T -1.08 -0.19 -1.95 * 1.05
4 N 679 885 1479 885
SH 306.95 50.15 365.15 734.49
T -0.31 1.06 -3.75 *** 0.59
Panel B: In-State Non-Failed Banks
Q C. I. F. R. M. B. F. C.
CP N 78 184 175 173
SH 1403.8 -749.3 191.57 -648
-4 N 13 17 14 8
SH 372.5 660.71 -8.5 8
T 2.71 *** -2.70 *** 1.05 -1.59
-3 N 11 31 15 19
SH 2545 465.74 -156.9 564.84
T -0.6 -2.77 *** 1.82 -2.30 **
-2 N 8 14 27 32
SH -179.3 7.79 -65.81 575.06
T 2.13 ** -1.88 * 1.26 -2.65 ***
-1 N 12 12 15 16
SH 92.5 -4.67 867 -170.3
T 2.05 * -2.03 ** -1.41 -1.02
-1to-4 44 74 71 76
SH 518.84 347.61 123.32 348.99
T 1.13 -2.96 *** 0.35 -2.40 **
1to4 53 74 61 72
SH 90.7 223.08 145.93 152.56
T 1.49 -2.73 *** 0.26 -1.98 **
1 N 18 19 16 9
SH 720.28 -41.16 252.81 129.78
T 0.35 -2.04 ** -0.24 -1.89 *
2 N 7 21 14 22
SH -348.9 -106.1 383.07 -91.18
T 2.38 ** -1.73 * -0.89 -1.4
3 N 21 26 20 19
SH -87.62 640.08 -181.1 -122.13
T 2.33 ** -3.15 *** 1.76 * -1.24
4 N 7 8 11 22
SH -533.7 359.5 283.18 643
T 2.66 ** -2.52 ** -0.54 -2.59 **
In this table trading in the event quarters is compared to trading
in the comparison period using the shares traded methodology. N is
the number of insider trades. SH is the mean number of shares traded;
CP indicates the comparison period. T is the test statistic for the
two-sample T-test. *** indicates significance at the 1 percent level,
** indicates significance at the 5 percent level and * indicates
significance at the 10 percent level. I.D. indicates that there was
insufficient data to perform the test.
Table 4: Comparison of Insider Trading in the Failed Bank to
Insider Trading in Non-Failed Banks using the Transactions
Methodology
Panel A: Nationwide Non-Failed Banks
Q C. I. F. R. M. B. F. C. ALLTX
CP N1 69 I.D. 41 3 5
PS1 0.5072 0.4146 0.6667 0.20
N2 4486 7851 4806 4809
PS2 0.2769 0.2049 0.2703 0.2705
T 3.78 *** 2.69 ** 1.19 -0.35
-4 N1 17 I.D. I.D. I.D. 6
PS1 0.1176 0.1667
N2 661 923
PS2 0.2617 0.1983
T -1.75 * -0.19
-3 N1 21 13 I.D. I.D. 23
PS1 0.4286 0.9231 0.6522
N2 670 1396 945
PS2 0.2313 0.1103 0.2201
T 1.76 * 10.50 *** 4.22 ***
-2 N1 15 1 9 I.D. 14
PS1 0.5333 0 1 .7143
N2 762 975 941 1557
PS2 0.2585 0.2226 0.2572 0.893
T 2.05 * . 52.11 *** 4.98 ***
-1 N1 10 I.D. 1 I.D. 7
PS1 0.30 . 0.2857
N2 740 977 1022
PS2 0.1676 0.1965 0.2045
T 0.86 . 0.44
-1 to -4 63 14 10 I.D. 50
PS1 0.3492 0.8571 1.00 0.5600
N2 2834 4483 3631 4519
PS2 0.229 0.1682 0.2289 0.1662
T 1.97 * 7.09 *** 110.59 *** 5.54 ***
1 to 4 46 1 10 I.D. 10
PS1 0.4783 1.00 1.00 1.00
N2 2989 3599 4363 4364
PS2 0.2349 0.2437 0.1948 0.1948
T 3.25 *** . 134.27 *** 134.30 ***
1 N1 4 I.D. 6 I.D. 6
PS1 0.25 1 1.00
N2 767 900 901
PS2 0.2099 0.2778 0.2775
T 0.16 48.35 *** 48.41 ***
2 N1 16 1 2 I.D. 2
PS1 0.3750 . 1 1
N2 647 1076 884 884
PS2 0.2133 0.2035 0.2455 0.2455
T 1.28 . 52.10 *** 52.10 ***
3 N1 20 I.D. 1 I.D. 1
PS1 0.5500 1 1.00
N2 889 1089 1089
PS2 0.2475 0.169 0.169
T 2.63 ** . .
4 N1 6 I.D. 1 I.D. 1
PS1 0.6667 1 1.00
N2 686 1490 1490
PS2 0.2668 0.1336 0.1336
T 1.89 . .
Panel B: In State Non-Failed Banks
Q C. I. F. R. M. B. F. C. ALLTX
CP N1 69 I.D. 41 3 5
PS1 0.5072 0.4146 0.6667 0.20
N2 78 175 173 176
PS2 0.1795 0.4286 0.5087 0.5114
T 1.70 ** -0.16 0.47 1.53
-4 N1 17 I.D. I.D. I.D. 6
PS1 0.1176 0.1667
N2 13 8
PS2 0.6923 0.500
T -3.69 *** 1.32
-3 N1 21 13 I.D. I.D. 23
PS1 0.4286 0.9231 0.6522
N2 11 31 19
PS2 0 0.3226 0.2632
T 3.87 *** 5.23 *** 2.68 **
-2 N1 15 1 9 I.D. 14
PS1 0.5333 . 1 0.7143
N2 8 14 27 32
PS2 0.125 0.3571 0.3171 0.2813
T 2.23 ** . 4.91 *** 2.91 ***
-1 N1 10 I.D. 1 I.D. 7
PS1 0.300 1 0.2857
N2 12 15 16
PS2 0.1667 0.2667 0.375
T 0.70 . -0.40
-1 to -4 63 14 10 I.D. 50
PS1 0.3492 0.8571 1.00 0.5600
N2 44 74 71 76
PS2 0.2727 0.3649 0.4648 0.3158
T 0.84 4.39 *** 8.98 *** 2.75 ***
1 to 4 46 1 10 I.D. 10
PS1 0.4783 1.00 1.00 1.00
N2 53 74 61 62
PS2 0.3208 0.3784 0.2951 0.2903
T 1.60 . 11.97 *** 12.21 ***
1 N1 4 I.D. 6 I.D. 6
PS1 0.25 1.00 1.00
N2 18 16 17
PS2 0.2222 0.3750 0.3529
T 0.10 5.00 *** 5.42 ***
2 N1 16 1 2 I.D. 2
PS1 0.3750 . 1 1.00
N2 7 21 14 14
PS2 0.4286 0.381 0.2143 0.2143
T -0.23 . 6.90 *** 6.90 ***
3 N1 20 I.D. 1 I.D. 1
PS1 0.55 1 1.00
N2 21 20 20
PS2 0.2857 0.45 0.45
T 1.73 * . .
4 N1 6 I.D. 1 I.D. 1
PS1 0.6667 1 1.00
N2 7 11 11
PS2 0.5714 0 0
T 0.33 . .
In this table, insider trading in the failed bank is compared to
insider trading in the non-failed banks using the transactions
methodology. N1 and N2 are the number of insider trades in the
failed bank and non-failed banks respectively. PS1 and PS2 are
the proportion of trades in the failed bank and non-failed banks
that were sales respectively. T is the test statistic for the
two-sample T-test. *** indicates significance at the 1 percent
level. ** indicates significance at the 5 percent level and
* indicates significance at the 10 percent level. I.D. indicates
that there was insufficient data to perform the test.
Table 5: Comparison of Insider Trading in the Failed Bank to
Insider Trading in Non-Failed Banks using the Shares Traded
Methodology
Panel A: Nationwide Non-Failed Banks
Q C. I. F. R. M. B. F. C. ALLTX
CP N1 69 I.D. 41 3 8
SH1 452.3 -1283 81.67 70.87
N2 4486 7851 4806 4806
SH2 -2272 63.12 1588.3 1588.3
T 0.32 -3.11 *** -0.95 -1.01
-4 N1 17 I.D. I.D. I.D. 6
SH1 2700.8 138.33
N2 661 923
SH2 252.61 217.28
T 1.35 -0.76
-3 N1 21 13 I.D. I.D. 23
SH1 -1100 -166.8 -2424
N2 670 1396 945
SH2 573.94 297.45 66.62
T -3.39 *** -6.86 *** -3.63 ***
-2 N1 15 1 9 I.D. 14
SH1 -1053 100 -1569 -74.21
N2 762 975 941 1557
SH2 520.54 93.81 5794 237.32
T -2.86 *** . -3.63 *** -4.71 ***
-1 N1 10 I.D. 1 I.D. 7
SH1 -286.7 -861 140
N2 740 977 1022
SH2 914.16 264.74 185.46
T -1.77 * . -0.28
1-to-4 63 14 10 I.D. 50
SH1 66.02 -147.8 -1499 -1100
N2 2834 4483 3631 4519
SH2 573.29 164.81 145.09 183.55
T -0.88 -5.27 *** -3.99 *** -3.57 ***
1to4 46 1 10 I.D. 11
SH1 -604.6 -50 -3040 -2755
N2 2989 3599 4363 4363
SH2 3466.3 507.14 655.26 655.26
T 2.09 ** . -3.35 *** -3.27 ***
1 N1 4 I.D. 6 I.D. 7
SH1 -11.5 -1243 -1051
N2 767 900 900
SH2 2768.9 1609 1609
T -1.07 -2.36 ** -2.20 **
2 N1 16 1 2 I.D. 2
SH1 -704.9 . -3122 -3122
N2 647 1076 884 884
SH2 2178.2 -12132 111.36 111.36
T -0.79 . -3.66 -3.66
3 N1 20 I.D. 1 I.D. 1
SH1 -481.4 -5000 -5000
N2 889 1089 1089
SH2 7450 706.32 706.32
T -2.25 ** . .
4 N1 6 I.D. 1 I.D. 1
SH1 -1144 -11700 .
N2 686 1490 1490
SH2 298.17 364.54 364.54
T -1.24 . .
Panel B: In State Non-Failed
Q C. I. F. R. M. B. F. C. ALLTX
CP N1 69 I.D. 41 3 8
SH1 452.3 -1283 81.67 70.88
N2 78 175 173 173
SH2 1403.8 191.57 -648 -648
T -0.87 -3.21 *** 1.15 1.73 *
-4 N1 17 I.D. I.D. I.D. 6
SH1 2700.8 138.33
N2 13 8
SH2 -372.5 8.00
T 1.70 0.79
-3 N1 21 13 I.D. I.D. 23
SH1 -1100 -166.8 -2424
N2 11 31 19
SH2 2545 465.74 564.84
T -1.97 * -2.19 ** -3.86 ***
-2 N1 15 1 9 I.D. 14
SH1 -1053 100 -1569 -74.21
N2 8 14 27 32
SH2 -179.3 7.78 -65.81 575.06
T -1.38 . -3.22 *** -2.50 **
-1 N1 10 I.D. 1 I.D. 7
SH1 -286.7 -861 140
N2 12 15 16
SH2 92.5 867 -170.3
T -0.76 . 1.01
1-to-4 63 14 10 I.D. 50
SH1 66.02 -147.8 -1499 -1100
N2 44 74 71 76
SH2 518.84 347.61 123.32 348.99
T -0.62 -2.95 *** -3.82 *** -372 ***
1to4 46 1 10 I.D. 11
SH1 -604.6 -50 -3040 -2755
N2 53 74 61 61
SH2 90.99 223.08 145.93 145.93
T -0.54 . -2.96 ** -2.86 **
1 N1 4 I.D. 6 I.D. 7
SH1 -11.5 -1243 -1051
N2 18 16 16
SH2 720.28 252.81 252.81
T -0.29 -3.61 *** -3.15 **
2 N1 16 1 2 I.D. 2
SH1 -704.9 -50 -3122 -3122
N2 7 21 14 14
SH2 -348.9 -106.1 383.07 383.07
T -0.11 . -3.93 -3.93
3 N1 20 I.D. 1 I.D. 1
SH1 -481.4 -5000 -5000
N2 21 20 20
SH2 -87.62 -181.1 -181.1
T -1.54 . .
4 N1 6 I.D. 1 I.D. 1
SH1 -1144 -11700 -11700
N2 7 11 11
SH2 -553.7 283.18 283.18
T -0.49 . .
In this table, insider trading in the failed bank is compared to
insider trading in non-failed banks using the shares traded
methodology. N1 and N2 are the number of insider trades in the
failed bank and non-failed banks respectively. CP indicates the
comparison period. SH1 and SH2 are the mean number of shares
traded in the failed banks and non-failed banks respectively. T
is the test statistic for the two-sample T-test, ***, **, and
* indicate significance at the 1, 5 and 10 percent levels
respectively. I.D. indicates that there was insufficient data to
perform the test.