How Well Capitalized

Are Well-Capitalized
Banks?

T
he wave of bank and savings and loan failures in the 1980s and
early 1990s, and the resulting losses to deposit insurance funds,
served to highlight the need for banks to hold sufficient capital to
survive difficult times. In addition, many argued that deposit insurance
made it imperative that banks be better capitalized, since deposit insur-
ance reduces the market discipline that depositors might otherwise
provide. With reduced market discipline, banks have an incentive to
take on greater risks and more leverage than they would if the market
fully reflected the increased risk such actions pose.1 Consequently, recent
bank regulatory initiatives increasingly have emphasized the role of bank
capital as a cushion to allow banks to absorb adverse shocks without
experiencing insolvency.
Recent bank legislation and regulation have sought to implement a
carrot-and-stick approach that penalizes banks that have too little capital,
while reducing the regulation imposed on banks deemed to be well
capitalized. The Federal Deposit Insurance Corporation Improvement
Act of 1991 (FDICIA) has served as the cornerstone of a major overhaul
of banking legislation unprecedented since the Great Depression. The
early intervention component of FDICIA provides for prompt corrective
action (PCA) based on capital ratio thresholds, with supervisory inter-
Joe Peek and vention in undercapitalized banks that becomes increasingly severe as the
Eric S. Rosengren bank’s capital position deteriorates.2 Undercapitalized banks are re-
stricted in their activities, and severely undercapitalized banks are subject
to early closure.3 On the other hand, as long as a bank is deemed to be
Professor of Economics, Boston Col- well capitalized, regulators are not required to take any action.
lege and Visiting Economist, Federal This emphasis on bank capital also is found in newly proposed
Reserve Bank of Boston; and Vice legislation to relax Glass-Steagall restrictions, which would allow ex-
President and Economist, Federal Re- panded activities only at banks that are well capitalized. Nor is the United
serve Bank of Boston, respectively. Leo States the only country to focus on bank capital regulation. With the 1988
Hsu and Carol Greeley provided valu- adoption of the Basle Accord, an international agreement that set com-
able research assistance. mon standards by which to evaluate capital adequacy, many other
countries now place greater emphasis on the role of cient capital to weather the storm. We find that many
bank capital. of the institutions that either failed or required sub-
While regulations are being designed to reward stantial supervisory intervention were well capitalized
banks that are deemed to be well capitalized and prior to the emergence of banking problems in New
restrict those that are not, no clear consensus has been England. In fact, four-fifths of the banks that failed
reached in the academic literature on just how much during the New England banking crisis were still
capital is necessary. Ideally, required capital ratios classified as well capitalized within two years of their
should be related to the degree of risk undertaken by failure. In addition, at one-third of those failed banks,
the individual institution, and a movement in that the capital-to-asset (leverage) ratio declined by more
direction has been taken with the implementation of than 5 percentage points in a single quarter, enough to
risk-based capital ratio requirements. Still, risk mea- wipe out the entire capital of any bank below the
surement remains a herculean task, given that many well-capitalized threshold.
bank assets are difficult to value and risk characteris- The recent increased emphasis on capital levels
tics can change rapidly as banks adjust their on- and has been instrumental in raising the capital ratios of
off-balance-sheet positions. most U.S. banks, and the more frequent examinations
required by FDICIA may result in reported capital
ratios that more accurately reflect bank health. Never-
theless, problems of the magnitude of those recently
Recent bank regulatory initiatives experienced in New England would require greater
increasingly have emphasized capital cushions than the minimum “well-capitalized”
PCA threshold, if widespread bank insolvencies were
the role of bank capital as a to be avoided.
cushion to allow banks to absorb The next section of this article briefly reviews
recent legislation that has attempted to induce banks
adverse shocks without to become better capitalized and examines how the
experiencing insolvency. legislation relates to traditional supervisory oversight.
The second section investigates whether examiners
view well-capitalized banks as posing little threat of
insolvency. The third section considers whether banks
While FDICIA established the capitalization cate- that met the “well-capitalized” threshold were suc-
gories, it was left to bank regulators to assign numer- cessful in avoiding problems during the severe eco-
ical values to capital ratios to serve as the thresholds nomic downturn in New England in the early 1990s
defining those categories. The levels that were even- and shows how quickly banks had their capital base
tually assigned were below the levels proposed by eroded. The final section provides some conclusions.
many of the early proponents of the legislation (Ben-
ston and Kaufman 1994a, 1994b) and originally con-
sidered by regulators (Carnell 1995). Nonetheless, the
capital ratio threshold associated with regulators’ cur- I. Capital Regulation and the
rent definition of a well-capitalized bank is certainly Supervisory Process
higher than that maintained by many banks before Prior to the enactment of FDICIA, the bank su-
FDICIA. pervision and examination process already gave bank
This article examines whether institutions satisfy- supervisors the opportunity both to verify that a
ing the “well-capitalized” criteria before and during bank’s practices and procedures were consistent with
the recent banking crisis in New England had suffi- safety and soundness criteria and to act to correct the
situation if they were not. Both informal and formal
1
An extensive literature exists on the causes of bank failures,
including loss of charter value (for example, Keeley 1990) and moral
3
hazard (for example, Kane 1985; Barth 1991). Numerous studies While reported capital ratios are notoriously inaccurate indi-
about moral hazard helped motivate the recent legislation that cators of the actual financial health of a bank (see, for example, Jones
focused on capital. and King 1995), the early closure provision, by raising the capital
2
Peek and Rosengren (1997) provide a more comprehensive ratio threshold that could trigger bank closure, should be expected
discussion of Prompt Corrective Action and a more technical to reduce but not eliminate the resolution costs of failed banks. In
analysis of its likely impact on the timing of supervisory interven- fact, Billett, Coburn, and O’Keefe (1995) find that overall resolution
tion in problem banks. costs have decreased since the enactment of FDICIA.

42 September/October 1997 New England Economic Review

regulatory actions could be used for early supervisory based on its capital: (1) well capitalized, (2) adequately
intervention, well before a bank reached the point of capitalized, (3) undercapitalized, (4) significantly un-
failure. dercapitalized, and (5) critically undercapitalized.
As the financial condition of a bank deteriorates, Banks in the top two categories essentially are not
the first action taken is usually the memorandum of restricted. Banks then come under progressively more
understanding (MOU), an informal regulatory action severe restrictions as they cross capital thresholds that
that lists recommended actions to improve the bank’s place them in lower categories. The leverage ratio
condition. If bank supervisors determine that a bank’s thresholds assigned by regulators to the top four
problems are more serious, they will institute a formal categories relevant to the early intervention provisions
regulatory action, either a written agreement or a of FDICIA are as follows: 5 percent or higher for well-
cease and desist order. Both actions cover the same capitalized banks; 4 percent or higher for adequately
general areas discussed in a full bank examination or capitalized; below 4 percent for undercapitalized; and
in the MOU. However, because formal actions are below 3 percent for significantly undercapitalized
legally enforceable agreements with civil penalties for institutions. Critically undercapitalized institutions (a
noncompliance, they are viewed as the most drastic ratio of tangible equity to assets of 2 percent or less)
action available to the bank supervisor short of closing face the early closure provisions of FDICIA.
the bank.4
Formal actions are intended to provide specific
recommendations for actions to be taken by banks to
prevent further deterioration in their financial condi- Many of the institutions that
tion. These recommendations may include improved either failed or required
management information systems, greater oversight
of credit risks, and improved reserving procedures. In substantial supervisory
addition to such general management recommenda- intervention were well capitalized
tions, frequently the result of deficiencies found dur-
ing the examination process, several specific quantita- prior to the emergence of banking
tive requirements are usually stated in the formal problems in New England.
action. By far the most common are requirements to
improve capital ratios or at least to maintain them
above a particular level.5 The most common target
While the capital thresholds associated with the
in these actions has been a 6 percent capital-to-asset
prompt corrective action provisions of FDICIA are
(leverage) ratio (Peek and Rosengren 1995c).
stated in terms of both leverage ratios and risk-based
The mandatory provisions instituted against un-
capital ratios, we focus only on leverage ratio thresh-
dercapitalized banks under the PCA guidelines of
olds. First, risk-based capital ratios are not available
FDICIA are similar to the conditions commonly im-
before 1990. Second, for the period in New England
posed on banks under formal regulatory actions.
under study here, leverage ratios rather than risk-
FDICIA requires that bank regulators each quarter
based capital ratios tended to be the binding con-
assign every bank to one of five regulatory categories,
straint on capital-constrained banks. This is consistent
with evidence on nationwide samples that leverage
4
Cease and desist orders signed after the August 9 passage of ratios and not risk-based capital ratios affected bank
the Financial Institutions Reform, Recovery, and Enforcement Act of behavior (for example, Hancock and Wilcox 1994).
1989 generally have been publicly disclosed by regulators, resulting As shown in Table 1, virtually all the PCA man-
in greater public scrutiny of the problems at the bank. Written
agreements have been publicly disclosed only since November 29, datory provisions imposed on undercapitalized and
1990, when the disclosure requirement was amended in the Crime significantly undercapitalized institutions are in-
Control Act of 1990. MOUs are not publicly disclosed by the cluded in formal regulatory actions.6 Only restrictions
regulatory agencies. Of course, the institution receiving the regula-
tory action can choose to announce that it had, or would soon, come
under a regulatory action.
5 6
Numerous studies have found evidence that bank behavior In addition to the mandatory actions under PCA listed in
was altered as a result of the renewed emphasis on achieving Table 1, examiners are allowed optional actions. For example, many
specified capital ratios (for example, Furlong 1992; Hall 1993; of the required provisions for significantly undercapitalized insti-
Hancock and Wilcox 1992; Peek and Rosengren 1995a), although tutions can be applied to undercapitalized institutions at the discre-
this view is not unanimous (for example, Berger and Udell 1994). tion of the bank supervisor, based on a bank’s unsatisfactory
See Burger and Udell (1994) for a detailed survey of this literature. CAMEL rating.

September/October 1997 New England Economic Review 43

 Table 1
Provisions for Prompt Corrective Action
Categories Specified in FDICIA for Prompt Corrective Actiona
Formal Regulatory Adequately Significantly
Action:b No Well Capitalized: Capitalized: Undercapitalized: Undercapitalized:
Explicit RBC $ 10% RBC $ 8% RBC , 8% RBC , 6%
Major Provision Capital Trigger and LR $ 5% and LR $ 4% or LR , 4% or LR , 3%
Increase loan loss reserve Yes No No No No
Increase charge-off of classified
assets Yes No No No No
No renewals or extensions of credit
to borrowers with classified assets Yes No No No No
Capital restoration plan required Yes No No Yes Yes
Suspend dividends Yes No No Yes Yes
Asset growth restricted Yesc No No Yes Yes
Prior approval required for
acquisitions, branching, and new
lines of business Yesd No No Yes Yes
Require recapitalization Yes No No No Yes
Restrict transactions with affiliates Yes No No No Yes
Restrict interest rates paid Noe No No No Yes
Further restrictions on asset growth Yes No No No Yes
Prohibits deposits from
correspondents Noe No No No Yes
Hire or replace senior management Yes No No No Yes
a
RBC represents the risk-based capital ratio. LR represents the leverage ratio. The fifth category, “Critically Undercapitalized,” when regulators can close
banks, is not shown.
b
Provisions are mandatory when included. While these provisions generally appear in formal actions, some formal actions do not include all of the provi-
sions.
c
Asset growth is restricted by requiring that capital-to-asset ratio targets be achieved. While most institutions shrink, asset growth is not explicitly restricted.
d
Usually stated as approval needed for any purchase or any activity influencing the capital plan.
e
Not explictly addressed, but could be restricted by general prohibition on unsafe or unsound banking practices.

on interest rates paid and on deposits from correspon- as the first three provisions typically included in
dents are not generally discussed explicitly in formal formal regulatory actions, as listed in Table 1. Formal
actions, although they could be assumed to be covered actions devote significant attention to classifying, re-
there by general prohibitions of unsafe and unsound serving for, and charging off (and transferring to
practices. It is not until banks become significantly OREO) problem loans, frequently including require-
undercapitalized that PCA provisions require restric- ments for explicit, quantified increases in loan loss
tions roughly equivalent to those contained in formal reserves that directly reduce reported capital. Thus,
regulatory actions, however. Well-capitalized institu- formal actions are generally more comprehensive than
tions are not restricted in any significant way, while the PCA provisions and they include nearly all of the
adequately capitalized institutions need FDIC ap- PCA provisions required of undercapitalized and sig-
proval to hold brokered deposits, a very modest nificantly undercapitalized institutions.
restriction for most adequately capitalized institu-
tions. Thus, under PCA provisions, significant actions
are taken only when an institution becomes undercap-
II. Do Well-Capitalized Banks Raise
italized. Undercapitalized banks must adopt a capital
Supervisory Concerns?
restoration plan, suspend dividends and management
fees, and restrict growth. If supervisors viewed well-capitalized banks as
The PCA guidelines in FDICIA also do not ad- posing little risk of insolvency, most formal actions
dress loan problems that can affect bank capital, such would occur only after a bank’s leverage ratio had

44 September/October 1997 New England Economic Review

 those banks that eventually under-
Table 2
went the most severe form of reg-
Leverage Ratios at New England Banks Operating in ulatory intervention short of clos-
1989:I That Received Formal Actions between 1989:I and ing the bank, as of 1989:I only five
of the 159 banks had a leverage
1993:IV
ratio below the well-capitalized
1989:I 1990:I
threshold of 5 percent; 77 had le-
Assets Assets verage ratios equal to or exceeding
Leverage Ratio Number ($000) Number ($000)
8 percent.
Less than 2.0 1 177,205 8 19,420,254 As of 1989:I, reported leverage
2.0 –2.5 1 64,770
2.5–3.0 1 64,208 2 7,802,880
ratios provided little indication of
3.0 –3.5 1 1,303,973 the extent of the severe banking
3.5– 4.0 3 1,815,926 problems soon to be experienced in
4.0 – 4.5 7 4,668,568 New England. Even as late as
4.5–5.0 3 17,843,223 4 41,538,350 1990:I, only 26 of these institutions
5.0 –5.5 6 26,838,272 8 1,868,188
5.5– 6.0 17 36,583,455 14 18,753,153
had a leverage ratio below 5 per-
6.0 – 6.5 15 16,400,457 16 8,924,377 cent. In addition, 51 of these banks,
6.5–7.0 12 3,878,168 14 6,976,463 almost one-third of the banks that
7.0 –7.5 17 6,057,514 18 4,483,754 would receive a formal action by
7.5– 8.0 10 2,284,694 12 1,989,581 1993:IV, still had a leverage ratio
Greater than or equal to 8.0 77 18,574,841 51 12,217,781
equal to or exceeding 8 percent.
All Banks Receiving Thus, reported leverage ratios did
Formal Actions 159 128,702,037 159 131,828,018
not forecast the extent of the im-
pending problems at these institu-
tions soon to come under formal
regulatory actions.
declined below the 5 percent “well-capitalized” One explanation for the failure of reported capital
threshold. To assess whether this is the case, we ratios to serve as leading indicators of formal actions is
examine the period of supervisory intervention asso- that the formal actions often are taken by supervisors
ciated with the New England banking crisis. This well before banking problems are revealed in reported
episode is particularly suited for such a study because bank data. For this reason, it may be informative to
the banking problems were widespread, it was the examine the level of bank capital ratios at the time
first major regional banking crisis to occur following supervisors imposed formal regulatory actions. This
the renewed emphasis on bank capital in the late would also provide some evidence about how serious
1980s, and, prior to this time, regulators were not a problem “well-capitalized” banks can pose, in the
required to publicly disclose formal regulatory ac- view of bank supervisors. To make this comparison, it
tions. is necessary to date the initiation of a formal action.
Between 1989:I and 1993:IV, the years that span For our purposes, we date the formal action as occur-
the period of severe banking problems in New En- ring at the beginning of the examination that resulted
gland, a large number of New England banks received in the formal action.8
formal regulatory actions. For banks operating in
1989:I that received a formal action, Table 2 shows
information. Because this table focuses on the ability of capital ratios
their number and the volume of their assets as of to foreshadow coming banking problems, the table includes neither
1989:I and 1990:I, grouping the 159 banks according to the seven New England banks that were already operating under a
their leverage ratios.7 Although the table includes only formal regulatory action at the end of 1988 (included in Table 5) nor
the three de novo banks that began operations after 1989:I that
subsequently received a formal action (included in Tables 3 and 5).
7 8
In this study, banks are defined to include all FDIC-insured The standard practice of the Federal Deposit Insurance Cor-
commercial and savings banks. The sample of banks is restricted to poration (FDIC) is to date examinations (which are usually reported
the First District of the Federal Reserve System (New England) for in the formal actions) as of the beginning of the exam. The Office of
three reasons. First, this was the region most severely affected by the Comptroller of the Currency (OCC), on the other hand, often
reduced bank capital. Second, this was the first region to have reports an “as of” date, the date of financial data referred to in the
extensive implementation of formal actions following the new report, often the end-of-quarter call report date immediately pre-
emphasis on capital in the late 1980s. Third, this is the only region ceding the start of the exam. Consequently, when the OCC exam
for which we have a complete set of formal actions and examination date is the last day of the quarter, we denote the subsequent quarter,

September/October 1997 New England Economic Review 45

 Table 3 shows the leverage ra-
tios of New England banks receiv- Table 3
ing formal actions between 1989:I Leverage Ratios at New England Banks That Received
and 1993:IV, both in the quarter Formal Actions between 1989:I and 1993:IV
immediately prior to the exam re-
One Quarter Prior to Exam At Exam Resulting
sulting in the formal action and at Resulting in Formal Action in Formal Action
the end of the quarter in which the
Assets Assets
exam occurred. Of the 162 institu- Leverage Ratio Number ($000) Number ($000)
tions that received formal actions
Less than 2.0 1 118,039 10 11,814,178
during this period (the 159 banks 2.0 –2.5 1 1,018,367 6 2,066,654
in Table 2 plus three de novo banks 2.5–3.0 4 1,130,876 9 2,527,144
not yet operating in 1989:I), 122, or 3.0 –3.5 5 3,458,220 4 815,266
three-quarters of the total number 3.5– 4.0 8 936,097 10 6,030,544
4.0 – 4.5 8 3,541,292 19 13,874,913
(65 percent if measured as a share
4.5–5.0 13 35,559,818 20 59,736,340
of assets), had leverage ratios 5.0 –5.5 18 34,145,499 15 8,094,667
above the 5 percent “well-capital- 5.5– 6.0 17 25,692,396 9 2,486,893
ized” PCA threshold in the quarter 6.0 – 6.5 17 4,731,821 13 6,709,029
prior to the exam. Almost 90 per- 6.5–7.0 16 8,384,222 11 10,082,397
7.0 –7.5 14 3,188,792 8 1,568,646
cent (143 institutions) had capital
7.5– 8.0 7 1,355,508 9 2,567,242
ratios above the 4 percent “ade- Greater than or equal to 8.0 33 7,624,212 19 3,566,124
quately capitalized” threshold, and All Banks Receiving
one-fifth of the banks still had le- Formal Actions 162 130,885,159 162 131,940,037
verage ratios above 8 percent,
twice the minimum leverage ratio
required to be considered ade-
quately capitalized by PCA stan-
dards. It appears that either supervisors do not view longer well capitalized after the exam. The deteriora-
reported capital ratios as sufficient statistics to mea- tion in reported capital ratios appears across the
sure bank health or they believe that a 5 percent board, with many banks experiencing a substantial
capital cushion is not sufficient to protect the deposit drop in their capital ratio in the quarter of the exam
insurance fund from bank failures. that led to the formal action. Thus, examiner enforce-
One could argue that examined data present a ment may be one of the primary factors that cause
more accurate indication of a bank’s health (for exam- banks to cross capital thresholds.
ple, FDIC 1997). If so, the more relevant measure of the The large number of well-capitalized banks re-
leverage ratio associated with the implementation of a ceiving formal actions indicates clearly that examiners
formal action would be the leverage ratio reported do not believe that high capital ratios are a sufficient
subsequent to the associated exam. In fact, as can be statistic for determining the health of the bank. This is
seen in the last two columns of Table 3, the distribu- especially true if the data have not been examined
tion of capital ratios based on examined data presents recently. It appears that examiners find that even
a much less rosy scenario, although over half the banks reporting that they are well capitalized can have
banks remain in the well-capitalized category. While very serious problems that affect their safety and
only 40 banks were not “well capitalized” prior to the soundness.
exam resulting in the formal action, 78 banks were no

in which the exam began, as the exam quarter. According to III. How Did Well-Capitalized Banks
discussions with examiners, banks normally will know they are
likely to receive a formal action at the beginning of an examination, Weather the New England Banking Crisis?
although the actual formal action is often not signed for several
months or even quarters after the completion of the exam. Further- It could be that when supervisors impose regula-
more, many of the provisions of the formal action that are time- tory actions on so-called “well-capitalized” banks,
dependent are dated as of the beginning of the exam. Peek and supervisors are being overly cautious and these banks
Rosengren (1995b) have found that bank behavioral responses, such
as decreases in lending, occur discretely in the quarter of the exam do not pose a significant risk of failure. One way to
resulting in the formal action, consistent with this dating practice. determine if supervisory caution about well-capital-

46 September/October 1997 New England Economic Review

ized banks is warranted is to ask how quickly
banks make the transition from being a well- Table 4
capitalized bank to being a failed bank. If Quarters to Failure from Last Well-Capitalized
bank capital erodes relatively slowly, then Quarter, New England Banks Failing Since
supervisors will have substantial time to take
1989:I
precautionary (and preventive) actions once
Quarters to Failure Cumulative
the well-capitalized threshold has been
from the Last Well- Number Number Cumulative
breached. If, instead, bank capital erodes rap- Capitalized Quarter of Banks Percent of Banks Percent
idly, then banks defined as “well-capitalized”
1 2 2.5 2 2.5
can still fail quickly, and regulatory interven- 2 1 1.3 3 3.8
tion may be necessary well before that capital 3 5 6.3 8 10.1
threshold has been crossed, if early interven- 4 14 17.7 22 27.8
tion policies are to be effective. 5 10 12.7 32 40.5
6 10 12.7 42 53.2
Table 4 examines the number of quarters
7 11 13.9 53 67.1
it took for each New England bank that failed 8 11 13.9 64 81.0
since 1989:I to make the transition from the 9 7 8.9 71 89.9
“well-capitalized” category to failure.9 The 10 3 3.8 74 93.7
results are striking. As recently as two years 11 2 2.5 76 96.2
12 1 1.3 77 97.5
before failure, 81 percent of these banks were
13 1 1.3 78 98.7
still well capitalized. Within one year of fail- 14
ure, 28 percent still were well capitalized. 15 1 1.3 79 100.0
Most of the failed banks moved from the
well-capitalized classification to failure dur-
ing a four- to eight-quarter period. Thus, once
a bank’s leverage ratio breached the 5 percent
well-capitalized threshold, regulators had rel-
atively little time to intervene before failure occurred. associated with the normal operations of a de novo
Given the quick erosion of bank capital during the bank, rather than a decline in the bank’s health,
New England banking crisis, effective early interven- declines in the leverage ratio during the initial eight
tion may require that regulators take corrective ac- quarters of operations are not considered.
tions to prevent bank failures well before the 5 percent The table separates the sample along two dimen-
well-capitalized threshold is crossed. This finding is sions: (1) banks that failed and banks that did not and
consistent with the observed pattern for formal actions (2) banks that did and did not receive a formal
taken by regulators. regulatory action. Of the banks that did not fail and
Table 5 shows the largest one-year decline in the did not receive a formal action, those least impaired by
leverage ratio during the period 1988:I to 1996:IV for New England’s banking problems, roughly one-third
every New England bank in operation in 1989:I. The experienced a decline of 2 percentage points or more
endpoints of the one-year subperiods correspond to in their reported capital-to-asset ratio over the course
the 1989:I to 1996:IV sample period that has been the of a single one-year period. Among banks that were
focus of this study. For de novo banks, the table troubled enough to receive a formal regulatory action,
includes the largest one-year decline in the leverage but strong enough to avoid failure (perhaps because of
ratio subsequent to the first two years of their opera- the guidance provided by the formal action), nearly
tions. A different standard is needed because a de two-thirds experienced a leverage ratio decline of 2
novo bank typically begins operations with a very percentage points or more in a one-year period, and 14
high capital ratio that declines over time as the bank percent experienced a decline of more than 5 percent-
expands its operations. Since such a decline would be age points, enough to wipe out the entire capital of a
bank at the 5 percent well-capitalized threshold.
Banks that failed exhibited an even higher pro-
9
The elapsed time between the last well-capitalized date and portion with extremely large one-year declines in their
the failure date is measured as the number of quarters between the leverage ratio. Of the 19 banks that failed before they
dates of the last call report at which the bank had a leverage ratio
equal to or greater than 5 percent and the last call report filed by the could receive a formal action, all experienced declines
failed bank. of more than 3 percentage points in their leverage

September/October 1997 New England Economic Review 47

 Table 5
Largest One-Year Leverage Ratio Decline for Each New England Bank, 1988:I to 1996:IV
Banks That Did Not Fail Failed Banks
No Formal Action Formal Action No Formal Action Formal Action
Percentage Point Decline Number Percent Number Percent Number Percent Number Percent
Less than 1 140 43.6 9 8.3
1 to 2 79 24.6 31 28.4 1 1.7
2 to 3 38 11.8 28 25.7 2 3.3
3 to 4 27 8.4 14 12.8 1 5.3 6 10.0
4 to 5 13 4.0 12 11.0 2 10.5 10 16.7
5 to 6 13 4.0 5 4.6 2 10.5 11 18.3
6 to 8 7 2.2 8 7.3 3 15.8 10 16.7
8 to 10 2 .6 1 .9 4 21.0 5 8.3
Larger than 10 2 .6 1 .9 7 36.8 15 25.0
Total 321 100.0 109 100.0 19 100.0 60 100.0
Note: Percent columns may not sum to 100 due to rounding errors.

ratio, and over 80 percent of these banks experienced was wiped out by the time of its resolution by the
declines greater than 5 percentage points. Among the FDIC, each of these banks experienced a further de-
banks that failed after receiving a formal action, two- cline in its capital ratio subsequent to that call report.
thirds experienced declines in their leverage ratio of This is most prevalent among those banks in the
more than 5 percentage points. table exhibiting the smallest one-year declines in their
As Table 5 shows, both failed and non-failed leverage ratios. For example, 19 of the 22 failed banks
banks in New England experienced substantial one- with a one-year decline of less than 5 percentage
year declines in their leverage ratios. And even these points, and nine of the 10 with a one-year decline of
statistics understate the severity of the declines in the less than 4 percentage points, reported a positive level
capital-to-asset ratios experienced by New England of capital at their last call report. Furthermore, two of
banks that failed. The full extent of the decline in the the three failed banks that experienced the smallest
one-year leverage ratio declines (less than 3 percent-
age points) each had a final reported leverage ratio in
excess of 6 percent, having failed only as a conse-
quence of the failure of larger affiliates that caused the
The capital ratio threshold failure of the entire holding company.
associated with the current The thrust of recent regulatory and legislative
proposals has been to treat well-capitalized banks as
definition of a well-capitalized having sufficient capital to absorb any unanticipated
bank may be set too low for losses. The reward for thus posing little risk to the
deposit insurance fund is that they receive less regu-
effective early intervention. latory oversight and they can more easily enter busi-
nesses prohibited to less well-capitalized banks. The
results in Table 5 suggest that the capital ratio thresh-
old associated with the current definition of a well-
capital ratios of many of these banks had yet to be capitalized bank is too low for effective early interven-
reflected in their reported balance sheet information. tion, a conclusion that concurs with earlier criticism of
For example, at the time of their final call report, a the way FDICIA was implemented (Benston and
number of these banks still reported positive levels of Kaufman 1994a, 1994b). Some banks in New England,
capital. Since the capital of each of the failed banks as a result of bad luck, bad management, or bad

48 September/October 1997 New England Economic Review

monitoring, lost more than 5 percentage points of their tinuities that occur between examined and nonexam-
capital-to-asset ratio in a single year. In fact, among ined quarters, and make the capital thresholds in
banks that eventually failed, nearly one-third experi- FDICIA more meaningful.
enced a decline in their leverage ratio in excess of 5
percentage points in a single quarter. Many well-
capitalized banks failed too quickly for early interven-
IV. Conclusions
tion policies to have any chance to be effective in
reversing their problems. Geographic and product barriers in banking are
These sharp declines in capital ratios reflect not breaking down as a result of financial innovations,
only the severity of the banking problems in New improvements in information technology, and
England, but also a failure to fully indicate the extent changes in legislation and regulation. These changes
of the problems on a bank’s books in a timely fashion. should enable banks to become more diversified and
This may be due in part to the fact that exams were better meet the financial needs of their customers.
relatively infrequent at many institutions during the Nonetheless, the changes can also impose risks on
late 1980s and early 1990s (FDIC 1997). This made the banks. As banks expand into new markets and prod-
data available to the public and the regulators poten- ucts, they may encounter problems unforeseen by
tially much less reliable. In fact, numerous studies management. FDICIA has required more frequent
have found that banks tend to be slow to recognize bank examinations and encouraged banks to be better
troubled loans and to add enough to loan loss reserves capitalized, but it is important not to assume that
to fully reflect the risks and problems in their portfo- because of the capital requirements in FDICIA, so-
lios (see, for example, Jones and King 1992; Gilbert called “well-capitalized” banks pose little or no risk of
1993; Dahl, Hanweck, and O’Keefe 1995). failure.
The problems with infrequent examinations may The experience in New England has shown that
have been remedied at least in part by FDICIA, which even many banks with reported capital ratios well
requires regulators to examine banks more frequently. beyond the 5 percent threshold still failed or required
Frequent exams will limit the discretion that banks can regulatory intervention. Capital ratios were not a
use with respect to the timing of reports of problem leading indicator of potential problems, frequently
loans and provisions for loan loss reserves, and they changing only after bank examiners forced an increase
will force banks to keep reported capital ratios more in in loan loss reserves following an examination or
line with the underlying condition of the bank. formal regulatory action.
During the New England banking crisis, when In addition, the capital cushion was not sufficient
most of the large bank losses were due to commercial to provide much lead time for regulators. Formal
real estate loans, one might have expected that loan regulatory actions frequently occurred while the bank
loss reserves would increase as the loan portfolio was well capitalized, and four-fifths of failed banks
deteriorated, in line with the impaired collateral. One failed within two years of having been well capital-
might also have expected that increases in loan loss ized. Their quick demise was in part the result of the
reserves would be highly correlated across institu- large, discontinuous declines in capital-to-asset ratios
tions, as deterioration in real estate markets forced all that occurred at troubled banks. It was not uncommon
banks to increase reserves. Instead, we observe large for the most seriously troubled banks to have their
declines in capital ratios in a single quarter, occurring capital-to-asset ratio decline by more than 5 percent-
at different times across institutions. This reflects the age points in a single year and, in a number of
practice of many banks during this period of deferring instances, in a single quarter. With such sharp de-
the realization of problems until bank examiners pres- clines, a 5 percent threshold was not sufficient to avoid
sured them to make provisions for impaired loans.10 insolvency at many New England banks.
Thus, one of the most beneficial requirements in Some of the large discontinuous declines in bank
FDICIA may very well be the mandating of more capital may be avoided in the future as institutions
frequent bank exams, which will improve the accu- become better capitalized, more diversified, and better
racy of reported bank capital ratios, reduce the discon- able to monitor risks, and as bank supervisors are
better able to monitor and correct risky activities at
10
banks. However, the experience of the New England
The New England experience may not generalize to future
periods to the extent that legislation and examiner attitudes discour- banking crisis should make us cautious about the
age forbearance. appropriate level of capital.

September/October 1997 New England Economic Review 49

References
Barth, James R. 1991. The Great Savings and Loan Debacle. Washing- Economic Review, January/February, pp. 35–52.
ton, D.C.: AEI Press. Hall, Brian J. 1993. “How Has the Basle Accord Affected Bank
Benston, George J. and George G. Kaufman. 1994a. “The Intellectual Portfolios?” Journal of the Japanese and International Economies, vol.
History of the Federal Deposit Insurance Corporation Improve- 7, pp. 408 – 40.
ment Act of 1991.” In Reforming Financial Institutions and Markets Hancock, Diana and James A. Wilcox. 1992. “The Effects on Bank
in the United States, edited by George G. Kaufman, pp. 2–17. Assets of Business Conditions and Capital Shortfalls.” In Credit
Norwell, MA: Kluwer Academic Publishers. Markets in Transition, Proceedings of the 28th Annual Conference
———. 1994b. “Improving the FDIC Improvement Act: What Was on Bank Structure and Competition, Federal Reserve Bank of
Done and What Still Needs to be Done to Fix the Deposit Chicago, May, pp. 502–20.
Insurance Problem.” In Reforming Financial Institutions and Markets ———. 1994. “Bank Capital and the Credit Crunch: The Roles of
in the United States, edited by George G. Kaufman, pp. 99 –120. Risk-Weighted and Unweighted Capital Regulations.” Journal of
Norwell, MA: Kluwer Academic Publishers. the American Real Estate and Urban Economics Association, March,
Berger, Alan N. and Gregory F. Udell. 1994. “Did Risk-Based pp. 59 –93.
Capital Allocate Bank Credit and Cause a Credit Crunch in the Jones, David S. and Kathleen Kuester King. 1992. “The Implemen-
U.S.?” Journal of Money, Credit and Banking, vol. 26, Part 2 tation of Prompt Corrective Action.” In Credit Markets in Transi-
(August), pp. 585– 628. tion, Proceedings of the 28th Annual Conference on Bank Struc-
Billett, Matthew T., Jane F. Coburn, and John P. O’Keefe. 1995. ture and Competition, Federal Reserve Bank of Chicago, May, pp.
“Acquirer Gains in FDIC-Assisted Bank Mergers: The Influence of 68 –100.
Bidder Competition and FDIC Resolution Policies.” In The New ———. 1995. “The Implementation of Prompt Corrective Action: An
Tool Set: Assessing Innovations in Banking. Federal Reserve Bank of Assessment.” Journal of Banking and Finance, vol. 19, June, pp.
Chicago, Proceedings of the 31st Annual Conference on Bank 491–510.
Structure and Competition, pp. 276 –94. Kane, Edward J. 1985. The Gathering Crisis in Federal Deposit Insur-
Carnell, Richard Scott. 1995. “A Partial Antidote to Perverse Incen- ance. Cambridge MA: The MIT Press.
tives: The FDIC Improvement Act of 1991.” Annual Review of Keeley, Michael C. 1990. “Deposit Insurance, Risk, and Market
Banking Law, pp. 317–71. Power in Banking.” The American Economic Review, vol. 80, De-
Dahl, Drew, Gerald A. Hanweck, and John P. O’Keefe. 1995. cember, pp. 1183–1200.
“Audits, Exams and Accounting Integrity in Banking.” Unpub- Peek, Joe and Eric S. Rosengren. 1995a. “The Capital Crunch:
lished manuscript. Presented at Academy of Financial Services Neither a Borrower Nor a Lender Be.” Journal of Money, Credit and
Annual Meeting, October. Banking, August, pp. 625–38.
Federal Deposit Insurance Corporation. 1997. “Bank Examination ———. 1995b. “Banks and the Availability of Small Business
and Enforcement, 1980 –1994.” In History of the Eighties—Lessons Loans.” Federal Reserve Bank of Boston Working Paper Series
for the Future: Proceedings of the FDIC Symposium. No. 95-1, January.
Furlong, Frederick T. 1992. “Capital Regulation and Bank Lending.” ———. 1995c. “Bank Regulation and the Credit Crunch.” Journal of
Federal Reserve Bank of San Francisco Economic Review, no. 3, pp. Banking and Finance, vol. 19, June, pp. 679 –92.
23–33. ———. 1997. “Will Legislated Early Intervention Prevent the Next
Gilbert, R. Alton. 1993. “Implications of Annual Examinations for Banking Crisis?” Southern Economic Journal, July.
the Bank Insurance Fund.” Federal Reserve Bank of St. Louis

50 September/October 1997 New England Economic Review