The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
FDK
Federal Deposit Insurance Corporation
Washington.
DC 20429
Office of the Chairman
March 27, 1995
Dear Chief Executive Officer:
Because of the extraordinary interest that banks and savings
associations have expressed in the Federal Deposit Insurance
Corporation’s proposals on deposit insurance premiums, I am
sending you the attached copy of the testimony that I submitted
to the House Financial Institutions subcommittee on March 23,
1995.
The testimony focuses on the undercapitalization of the
Savings Association Insurance Fund.
Beginning on page 27, it
discusses three standards that should be applied to any proposed
solution to that undercapitalization.
It then discusses the wide
range of solutions that have been proposed, applying the three
standards to a discussion of each in turn.
The testimony does
not take a position on any of the proposed solutions.
The comment period on the deposit insurance premium
proposals closes April 17, 1995. All the interested parties are
urged to make their views known.
Please send your written
comments to the Office of the Executive Secretary, Federal
Deposit Insurance Corporation, 550 17th Street NW, Washington,
D.C. 20429.
Sincerely
Ricki Tigert Helfer
Chairman
Attachment
TESTIMONY OF
RICKI TIGERT HELFER
CHAIRMAN
FEDERAL DEPOSIT INSURANCE CORPORATION
ON
THE CONDITION OF THE BIF AND THE SAIF AND RELATED ISSUES
BEFORE THE
SUBCOMMITTEE ON FINANCIAL INSTITUTIONS
AND CONSUMER CREDIT
COMMITTEE ON BANKING .AND FINANCIAL SERVICES
U.S. HOUSE OF REPRESENTATIVES
THURSDAY, MARCH 23, 1995
ROOM 2129 RAYBURN OFFICE BUILDING
INTRODUCTION
Madam Chairman and Members of the Subcommittee, I am here
today to present the views and analyses of the Federal Deposit
Insurance Corporation (FDIC) concerning the condition of the Bank
Insurance Fund (BIF) and the Savings Association Insurance Fund
•
We face a compelling problem —« and one that has grown
more compelling this year.
prospects appear favorable.
The BIF is in good condition and its
Despite the general good health of
the thrift industry, however, the SAXF is troubled.
to the SAIF problem requires action by the Congress.
Any solution
Indeed, the
need for Congressional action is more urgent today than ever
before.
Beginning later this year, a substantial disparity between
the deposit insurance premiums paid by BIF members and SAIF
members is likely to occur.
The disparity is mandated by current
statutory provisions.
The FDIC cannot avoid bringing the
disparity into being.
Only Congress can change the laws that
will soon require the FDIC to promulgate significantly different
assessments for the two deposit insurance funds.
Like the tip of
an iceberg, the premium disparity is only the visible
manifestation of a larger difficulty, most of which lies beneath
the surface.
2
This difficulty —
which most recently has been described in
depth in a report by the General Accounting Office —
has three
dimensions.
One, as Chart 1 shows, the SAIF is significantly
underfunded.
At year-end 1994, the SAIF had a balance of $1.9
billion —
or 28 cents in reserves for every $100 in insured
deposits.
This amounts to six percent of the assets of SAIF-
insured "problem" institutions.
The $21.8 billion BIF, in
contrast, amounts to 52 percent of the assets of BIF-insured
problem institutions.
Assuming that loss experience from failed
thrifts does not increase significantly from today's levels, the
SAIF is not expected to be fully capitalized at $1.25 in reserves
for every $100 in insured deposits until at least 2002.
Two, an ongoing fixed draw of $779 million on SAIF revenue
arises from the obligation to pay interest on bonds issued by the
Financing Corporation (FICO) in the 1980s.
This draw alone
creates a premium differential between BIF members and SAIF
members that likely would persist for 24 years until the bonds
are repaid.
This differential, at least 11 basis points, could
provoke further shrinkage in the SAIF assessment base and a
shortfall of assessment revenue to pay the FICO obligation, which
would lead to default on the bonds.
If you have ever tried to
fill a bucket with a hole in it, you understand what I mean.
3
Three, for the first time, the SAIF will assume
responsibility for resolving failed thrifts after June 30 of this
year.
Given the underfunding of the SAIF, significant insurance
losses in the near-term could render the SAIF insolvent and put
the taxpayer at risk.
This risk stems from the fact that deposit
insurance carries with it an implicit U.S. Government guarantee.
THE SEARCH FOR A SOLUTION
To establish parity between the BIF and the SAIF today would
require about $15.1 billion, or about 25 percent of the total
equity capital of SAIF members.
Of this total, $6.7 billion
would be needed to increase the SAIF from its unaudited year-end
1994 balance of approximately $1.94 billion to $8.66 billion, the
amount that currently would achieve the designated reserve ratio
required by Congress of 1.25.
The remaining $8.4 billion of the
$15.1 billion is the amount that would be necessary at current
interest rates to defease the FICO obligation.
That is to say,
it is the amount that would have to be invested today to generate
an income stream sufficient to service the FICO bonds until
maturity between the years 2017 and 2019.
Requiring these amounts to be collected entirely through
SAIF insurance premiums"raises difficult questions.
What will be
the effect on the ability of SAIF members to raise new capital,
to prosper, and to compete effectively?
will erosion of the SAIF
4
assessment base and changes in its composition jeopardize the
ability of the FICO to meet its obligations?
burden be shared?
Should some of the
And by whom?
There is no magic answer to these questions.
No matter how
the $15.1 billion cost is borne, there will be an outcry by at
least one constituency that a great injustice is being done.
There is no way for the FDIC to resolve this issue through the
exercise of its regulatory authority.
For two reasons, the need to find solutions to the problems
grows more urgent.
One, as mentioned earlier, starting July 1,
1995, the cost of all new thrift failures must be paid out of the
SAIF.
Two, recently announced efforts by some SAIF-insured
institutions to transfer deposits into BIF-insured institutions
raises the specter that the insured deposit base of the SAIF
could shrink so rapidly that, under current assessment rates,
debt service on the FICO bonds would quickly run into trouble.
Although the need for immediate Congressional action
concerning the SAIF is evident, there is considerable
disagreement over precisely what action should be taken and
whether it should be taken this year or later.
The most
frequently mentioned sources of money to address SAIF's needs
include the thrift industry, the banking industry, and the U.S.
Treasury.
Others have been mentioned, too, as having an interest
5
in resolving the problems.
None of the possible sources of
funding is happy about the prospect of footing the bill for
capitalizing the SAIF and funding the FICO interest payments.
The first section of this testimony describes the conditions
of the BIF and the SAIF and the reasons for the coming disparity
in their assessment rates.
The second section of the testimony
summarizes the statutory constraints that prevent a regulatory
solution to the problems.
The third section of the testimony
discusses the unprecedented public hearing on this subject held
on March 17 before the Board of Directors of the FDIC.
This is
followed by an analysis of the various proposals for addressing
the SAIF problem, measured against three standards set out in the
testimony.
THE CONDITION OF THE BIF AND THE SAIF
Bank Ipaiiranee Fund
The good news in this testimony is about the Bank Insurance
Fund.
The fund balance is rapidly approaching the
recapitalization level specified in the Financial Institutions
Reform, Recovery, and Enforcement Act of 1989 (FIRREA) and
confirmed in the Federal Deposit Insurance Corporation
Improvement Act of 1991 (FDICIA).
the designated reserve ratio —
As noted before, that level -
is 1.25 percent of insured
6
deposits
or $1.25 for every $100 insured deposits.
At year-
end 1994, the BIF had a balance of $21.8 billion, which was 1.15
percent of insured deposits.
The BIF has made a remarkable recovery.
Three years ago, at
year-end 1991, the BIF had a negative balance of $7.0 billion.
From this nadir, the lowest level in the bank fund's six decades
existence, the balance improved to a negative $100 million at
year-end 1992 and a positive $13.1 billion at year-end 1993.
In other words, since year-end 1991, the BIF has grown by
almost $29 billion.
of the BIF.
Two factors contributed to the restoration
One, fewer banks failed than had been anticipated.
While the number and assets of failed banks reached record levels
in the late 1980s and early 1990s, both fell sharply in the last
two years.
As a result, declining insurance losses enabled the
FDIC to recapture reserves that had been set aside before 1992.
In fact, over the last three years (1992 through 1994) reversing
provisions for insurance losses increased BIF net income by $12.8
billion.
Second, banks have paid significantly higher premiums to the
BIF than they paid previously.
rates were increased sharply.
Beginning in 1990, assessment
Rates are now almost three times
than the rate paid in 1989.
In the last three years,
7
insured institutions have paid nearly $17 billion in assessments
to the BIF.
The recovery of the BIF reflects the recovery of the banking
industry from the problems of the late 1980s and early 1990s.
Since 1990, the earnings of the industry have been on an
impressive upward trend:
$16.1 billion for 1990, $18.6 billion
for 1991, $32.2 billion for 1992, $43.1 billion for 1993, and
$44.7 billion for 1994.
The results for 1992, 1993, and 1994
were successive earnings records.
Ninety-one percent —
more than nine of every ten —
BIF-
member institutions are currently in the lowest risk category and
pay the lowest assessment rates.
percent of all BIF-member assets.
These institutions hold 88
They meet the highest
regulatory capital standards and have the strongest examiner
ratings*
These institutions are not expected to cause losses to
the BIF in the near-term.
As bank earnings have improved, bank failures have declined
dramatically.
The number of BIF-insured failures in 1994 was 13,
the lowest total since 1981.
These 13 failures marked the
continuation of a seven-year downward trend:
221 in 1988, 207 in
1989, 169 in 1990, 127 in 1991, 122 in 1992, and 41 in 1993.
The
estimated costs for these 13 failures last year is $139 million,
all of which had been reserved in prior years.
Consequently, no
8
additional expenses for failures were incurred by the BIF in
1994.
As a result of the recovery of both the banking industry and
its insurance fund, the BIF is projected to reach the 1.25
statutory designated reserve ratio between May and July of this
year.
Thereafter, absent a factual basis for a higher reserve
ratio, the FDIC has a statutory mandate to set deposit insurance
assessments to maintain the balance of the fund at the 1.25
ratio, at the same time retaining a risk-related system of
premiums and assessing each BIF member at least $1,000
semiannually.
Therefore, when the designated reserve ratio for
the BIF is reached —
an event that appears imminent —
the law
requires the FDIC to reduce assessments for BIF members.
In January of this year, the FDIC Board of Directors issued
a proposal to lower assessment rates for all but the riskiest BIF
members once the fund attains the designated reserve ratio.
Because the SAIF is significantly undercapitalized, the FDIC
Board proposed maintaining assessment rates for SAIF members at
current levels.
If the two proposals are adopted, a significant
disparity will exist between the assessment rate schedule for
BIF-insured institutions and the assessment rate schedule for
SAIF-insured institutions, regardless of whether the Board
retains the current SAIF rate schedule or reduces SAIF
assessments to the statutory minimum weighted average of 18 basis
9
points.
The FDIC has asked for public comments on the assessment
rate proposals, and the 60-day comment period extends until April
17.
The FDIC also held an unprecedented public hearing on issues
related to the BIF and SAIF assessment rate proposals, as
discussed in the next section.
Savinas Association Insurance Fund
There is also good news about the health of the savings and
loan industry.
Eighty-seven percent of all SAIF-member
institutions with 71 percent of SAIF—member assets are in the
lowest risk category and pay the lowest assessment rates.
Despite the good news in the savings and loan industry, the
SAIF —
as noted earlier —
underfunded.
is troubled.
It is significantly
Assessment revenue is constantly being diverted to
meet obligations from savings and loan failures in the 1980s.
The SAIF must begin paying for thrift failures that occur after
mid-year.
This testimony discusses each of these three issues in
turn.
First, the SAIF is undercapitalized.
As noted earlier, the
SAIF had a balance of $1.9 billion, or only 0.28 percent of
insured deposits at year-end 1994.
Thus, the current insurance
reserve amounts to only six percent of the assets of SAIF-insured
"problem" institutions.
The $21.8 billion BIF balance, in
10
contrast, amounts to 52 percent of the assets of BIF-insured
problem institutions.
At the current pace, and under reasonably
optimistic assumptions, the SAIF would not reach the minimum
reserve ratio of 1.25 percent until at least the year 2002.
Consequently, it would be impossible to lower SAIF premiums to
the proposed levels for the BIF for at least seven years, and
because of the continuing need to fund interest payments on the
FI CO bonds, probably much longer.
Second, SAIF assessments have been diverted to purposes
other than the fund.
This problem was described in detail in the
recent General Accounting Office report.
1994, $7 billion —
—
In short, from 1989 to
approximately 95 percent of SAIF assessments
was diverted from the SAIF to pay off obligations from thrift
failures in the 1980s through the Resolution Funding Corporation
(REFCORP), the Federal Savings and Loam Insurance Corporation
Resolution Fund (FRF), and the Financing Corporation (FICO)
Attachment B ) •
(see
Of the $9.3 billion in SAIF assessment revenue
received from 1989 to 1994, a total of $7 billion was diverted:
$1.1 billion was diverted to REFCORP; $2 billion was diverted to
FRF, and $3.9 billion to date, was diverted to FICO.
SAIF
assessment revenue currently amounts to just over $1.7 billion a
year, while FICO interest payments run $779 million a year, or
about 45 percent of all SAIF assessments.
Without these
diversions, the SAIF would have reached its designated reserve
ratio in 1994.
The REFCORP and FRF no longer have claims on SAIF
11
assessments, but the FICO claim will remain as an impediment to
capitalizing SAIF for 24 years.
Third, the SAIF will be under stress beginning on July 1,
1995, when it takes over responsibility for resolving all new
failures of SAIF-insured savings associations.
One large or
several sizable thrift failures could bankrupt the fund.
funding sources may be available to pay for losses:
Two
(1) an
authorization for payments from the U.S. Treasury of up to $8
billion for losses incurred by the SAIF in fiscal years 1994
through 1998; and (2) unspent RTC money during the two years
following the RTC's termination on December 31, 1995.
To obtain
funds from either of these sources, the FDIC must certify to
Congress that an increase in SAIF premiums would reasonably be
expected to result in greater losses to the Government, and that
SAIF members are unable to pay assessments to cover losses
without adversely affecting their ability to raise and maintain
capital or maintain the assessment base.
Congress required these certifications in an effort to
ensure that SAIF members pay the highest rates possible before
taxpayer funds are used to cover losses.
Of course, this would
have the effect of exacerbating the impending premium
differential.
It may require extremely grave conditions in the
thrift industry in order for the FDIC to certify that raising
SAIF assessments would result in increased losses to the
12
Government•
Moreover, these sources of funds cannot be used to
capitalize the fund —
that is, to provide an insurance reserve,
which was the original purpose of requiring a 1.25 reserve ratio.
A detailed discussion of the legislative history of the SAIF
funding scheme is contained in Attachment A.
By far the largest of the drains on SAIF assessment income,
the FI CO was established by Congress in 1987 in an attempt to
recapitalize the defunct Federal Savings and Loan Insurance
Corporation (FSLIC).
The FICO was provided with approximately
$3.0 billion in capital by the Federal Home Loan Banks.
The
capital was used by the FICO to purchase zero-coupon U.S.
Treasury securities.
These securities in turn served as
collateral for the issuance of 30-year interest-bearing debt
obligations by the FICO.
The proceeds from these obligations
were channeled by the FICO to the FSLIC.
From 1987 to 1989, the
FICO issued approximately $8.1 billion in bonds.
When they
mature, the principal values, or face amounts, will be paid with
the proceeds of the simultaneously maturing zero-coupon Treasury
securities.
No FICO bonds were issued after 1989, and the FICO's
issuing authority was terminated in 1991.
The Competitive Equality Banking Act of 1987 made FSLICi nsuredinstitutions responsible for the annual interest
payments.
FIRREA abolished the FSLIC, created the SAIF, and
13
reaffirmed the FICO's first priority to assess SAIF members.
The
FICO bonds do not mature until 2017 to 2019 and are not callable.
In enacting FIRREA, Congress in 1989 recognized that draws
on the SAIF by the FRF, REFCORP, and FI CO would delay the
capitalization of the insurance fund.
At that time, the GAO
notes, the Administration projected annual thrift deposit growth
of six to seven percent.
Since SAIF's inception, however, total
SAIF deposits have declined an average of five percent annually.
FIRREA authorized the appropriation of funds to the SAIF in
an aggregate amount of up to $32 billion to supplement assessment
revenue by ensuring an income stream of $2 billion each year
through 1999 (not to exceed $16 billion in the aggregate) and to
maintain a statutory minimum net worth through 1999 (not to
exceed $16 billion in the aggregate).
Subsequent legislation
extended the date for receipt of Treasury payments to 2000.
Despite requests by the FDIC to the Department of the Treasury
and the Office of Management and Budget, the Treasury never
requested any appropriations for these purposes, and the SAIF
never received any of the authorized funds.1
The issue of the SAIF's need for appropriated funds to reach
mandated reserve levels has been recognized by the FDIC since the
creation of the SAIF.
It was raised on January 10, 1992, in a
letter from William Taylor, Chairman of the FDIC, to Richard
Darman, Director, U.S. Office of Management and Budget, and it was
raised again in a letter, dated February 20, 1992, from Stanley J.
Poling, Director,
FDIC Division of Accounting and Corporate
Services, to Jerome H. Powell, Assistant Secretary for Domestic
14
The outlook for the SAIF is further complicated by the fact
that the law limits SAIF assessments that can be used for FICO
payments to assessments on insured institutions that are both
savings associations and SAIF members.
Two types of institutions
that pay assessments to the SAIF, Oakar and Sasser institutions,
are not savings associations that are SAIF members.
An Oakar is
a BIF member that has acquired SAIF*-insured deposits and
therefore pays deposit insurance premiums to the BIF and the
SAIF.
Between late 1989 and year-end 1994, 715 banks had
purchased $180 billion of thrift deposits —
or 25 percent of
year—end 1994 SAIF domestic deposits.
A Sasser institution is a commercial bank or a state savings
bank that has changed its charter from a savings association to a
bauik but remains a SAIF member.
There are 319 "Sasser" banks
holding deposits of $53 billion —
or 7.4 percent of SAIF
domestic deposits.
Because assessment revenue from Oakar banks and from Sasser
banks cannot be used to meet debt service on FICO bonds, almost
33 percent of SAIF-insured deposits were unavailable to meet FICO
Finance, U.S. Treasury. More recently, the issue was addressed at
the time Congress was considering the RTC Completion Act in a
letter dated September 23, 1993, from Andrew C. Hove, Jr., Acting
Chairman, to the House and Senate Banking Committee Chairmen and
Ranking Minority Members.
(Copies of this correspondence are
appended in Attachment C.)
See also the Testimony of Andrew C.
Hove, Jr., on "The Condition of the Banking and Thrift Industries,"
before the United States Senate Committee on Banking, Housing and
Urban Affairs, September 22, 1994.
15
payments in 1994 (see Chart 1) .2
percent at the end of 1993.
This portion was up from 25
This shift contributed significantly
to a 7.9 percent decline in 1994 in the SAIF assessment base
available to service FICO, even though the overall insured
deposit base of the SAIF declined by only 1.1 percent in 1994.
At current assessment rates, an assessment base of $325 billion
is required to generate revenue sufficient to service the FICO
interest payments.
The FICO-available base at year-end 1994
stood at $486 billion.
The difference of $161 billion can be
thought of as a cushion which protects against a default on the
FICO bonds.
If the 7.9 percent rate of shrinkage in the SAIF
assessment base available to FICO were to continue, this FICOcushion would be eliminated within five years.
The disparity that would arise from the FDIC's premium
proposals would further complicate the outlook for SAIF.
The
proposed assessment rate schedules for BIF and SAIF members are
shown in Table 1.
The proposals would result in SAIF members
2See Notice of FDIC General Counsel's Opinion No. 7, 60 FR
7055 (February 6, 1995), confirming a 1992 opinion of the FDIC
Legal Division that assessments paid by banks on deposits acquired
from SAIF members should remain in the SAIF and not be allocated
among the FICO, REFCORP, or FRF. In a letter to the FDIC Board of
Directors, dated May 11, 1992, the Comptroller General described
this conclusion and treatment of Oakar assessments as "reasonable."
§ee letter from Charles A. Bowsher, Comptroller General of the
United States, to the FDIC Board of Directors, dated May 11, 1992.
In addition, the FDIC General Counsel's opinion states the FDIC
Legal Division's position that assessments paid by any former
savings association that has converted to a bank and remains a SAIF
member are not available to the FICO.
See GAO Report 95-84,
Deposit Insurance Funds. March 1995, p. 15.
16
paying an average assessment rate of 24 basis points, 19.5 basis
points higher than the average rate of 4.5 basis points for BIF
members.
This premium differential could adversely affect SAIF
members in a number of ways, including increasing the cost of
remaining competitive, impairing their ability to generate
capital internally or externally, and causing marginally higher
rates of failure.
Historically, savings associations have paid somewhat higher
deposit insurance premiums than have banks.
From 1935 to 1980,
this differential was 4 to 5 basis points, and from 1980 to 1991
the differential ranged as high as 12.5 basis points.
the differential was zero.
In 1992,
Since 1992, under risk-related
assessments, SAIF members have paid an average rate about 1 to 2
basis points above the average rate for BIF members.
It is not
clear that these historical differentials are instructive when
evaluating the impact of the differential that would result from
the current assessment-rate proposals.
Previous premium
differentials were smaller and the marketplace is widely
considered to be more competitive today.
By way of background, from 1966 until 1984, thrifts were
allowed to pay slightly higher rates of interest on deposits
under Regulation Q.
This interest rate differential was most
frequently set at 25 to 50 basis points and was justified by the
advantage that banks had in accepting interest-free demand
17
deposits and engaging in commercial lending.
The Regulation Q
advantage may have lessened the burden of higher insurance
premiums for thrifts.
All these advantages were eventually
dissipated by innovation, market forces and legislation.
We have considered the effect of a differential on pricing,
on capital and on failures.
Pricing.
If BIF-members pass all or some of their
assessment reductions to their depositors by paying higher
interest rates or to their borrowers by charging lower rates,
SAIF members would be forced to incur higher costs in order to
remain competitive.
It is difficult to predict the eventual size
of the effective differential because this will be determined by
and SAIF-member management.
In the extreme case where SAIF
members absorb all of the differential, pretax earnings in the
aggregate would be reduced by $1.4 billion.
For the 25 percent
of SAIF members earning a return on assets of 1.13 percent or
higher in 1994, a differential of 20 basis points would reduce
pretax earnings by 6.8 percent.
For SAIF members with the median
ROA of 0.86 percent in 1994, pretax earnings would be reduced
about 12 percent.
significant.
dramatic.
Earnings reductions this large would be
The likely impact, however, promises to be less
BIF members are likely to use some portion of their
assessment savings to increase dividends or otherwise enhance
shareholder value, and SAIF members can offset some portion of
18
the differential by increasing revenues or reducing other
expenses.
Capital.
To the extent SAIF members7 earnings are reduced
by a premium differential, their ability to generate or raise
capital could be impaired.
Thrifts7 average returns on assets
and equity already lag significantly below those of banks, and
the industry faces longer-term structural problems that will be
difficult to overcome.
This is primarily due to the fact that
the business of mortgage lending has become increasingly
competitive, reducing the profitability of holding mortgage loans
to maturity.
However, current tax laws require thrifts to
maintain a certain percentage of their tangible assets in
"qualified thrift investments" in order to realize the tax
benefits available under a thrift charter.
In recent years, we
have seen some thrifts successfully raise new capital, even in
some instances where the institutions were unprofitable, and we
must conclude that the potential for a future premium
differential was known at the time of issue.
However,
investors
cannot be expected to suffer low returns indefinitely.
Failures.
We are particularly concerned about the possible
effects a premium differential could have on weaker institutions
and whether a differential would cause any increase in failures.
We analyzed the group of SAIF-insured institutions with FDIC
supervisory ratings of 3 or higher and projected their
19
performance for a five-year period, incorporating a 20-basis
point differential and a variety of interest-rate and assetquality assumptions.
The results showed a slight increase in
failures attributable to the differential, but we feel these
additional failures should be manageable by the SAIF provided
there is no unforeseen spiking of losses attributable to other
factors, such as an economic downturn.
In fact, in our
projections the factors relating to interest rates and asset
quality had a greater effect on failure rates than did a premium
differential.
The potential cumulative effect of all three
factors could be substantial.
Our analysis is included as
Attachment C.
Most recently, the outlook for the SAIF has been further
clouded by dramatic new developments.
On March 1, 1995, Great
Western Financial Corporation, the parent company of a SAIFmember federal savings bank with offices in California and
Florida, announced that it had submitted applications for two
national bank charters.
Under the applications these commercial
banks would share Great Western's existing branch locations.3
By mid-March,
five other SAIF-insured institutions announced that
they were considering similar actions.
3
In its press release of March 1, 1995, Great Western noted the
proposed premium differential and said the company's plan would
"ensure its ability to offer deposit products at rates which will
be competitive with commercial banks."
20
If these or other efforts in converting SAIF-insured
deposits to BIF-insured deposits are successful, others are
likely to follow.
That would mean the SAIF assessment base could
shrink significantly —
and quickly.
These six institutions have
approximately $80 billion in SAIF deposits, which represent 50
percent of the FICO-cushion mentioned earlier.
Removal of those
deposits from the SAIF would result in a significantly smaller
base from which to generate the fixed FI CO assessment.
Such a large shift in deposits would also have ramifications
for the BIF. An additional $80 billion in BIF-insured deposits
would require an additional $1 billion in BIF reserves —
percent of $80 billion.
1.25
While these announcements are unlikely
to result in a large enough shift in insured deposits from the
SAIF to the BIF by midyear to delay recapitalization of the BIF,
such a shift could ultimately push the reserve ratio below 1.25
percent.
If this were to occur, premiums paid by banks would
have to be increased in order to again reach and maintain the
1.25 target ratio.
The six new BIF members would begin
contributing assessments to the BIF, but other BIF members would
pay the preponderance of the needed $1 billion addition to
reserves•
It is estimated that many more thrift institutions are
considering ways of shifting deposits to the BIF.
The announced
proposals require various approvals associated with chartering
21
new institutions, but there are other means to achieve the same
ends that do not require such approvals, and are likely to lead
to a further shrinkage in the SAIF assessment base.
For example,
existing affiliations between BIF and SAIF members enable
deposit—shifting without the need for new charters or approvals
by regulators.
In general, we can expect the market to respond
to cost differences, and those who suggest that regulators can
prevent the movement of deposits out of the SAIF appear to
underestimate the market's ability to innovate around
constraints.
If the rate of shrinkage in the SAIF assessment
base increases 4 percent per year as a result of all available
techniques, debt service on the FI CO bonds is threatened as early
as 2001.
If the rate of shrinkage in the SAIF assessment base
increases to 10 percent per year, debt service on the FICO bonds
is threatened as early as 1977 (see figure 4 of Attachment C ) .
CONSTRAINTS
A number of legal constraints prevent a regulatory solution
to the SAIF problem and, therefore, require Congressional action
if the problem is to be addressed.
•
Among the constraints:
The law requires that the FDIC Board set assessments to
maintain each deposit insurance fund's reserve ratio at
the minimum designated reserve ratio (DRR) of 1.25
percent of estimated insured deposits once that ratio
has been achieved.
22
•
The FDIC Board may increase the DRR above 1.25 percent
for any year only if the Board determines that
circumstances exist raising a significant risk of
substantial future losses to the fund for the y e a r .
•
Assessment rates and the DRR of the BIF and SAIF must
be set independently.
•
The BIF and the SAIF must be maintained separately,
with no commingling of assets, liabilities, revenues or
expenses.
•
The FDIC Board must maintain a risk-based assessment
system and assess each fund member at least $1,000
semiannually after a fund is capitalized.
•
Until January 1, 1998, the FDIC Board is required to
set SAIF assessments to increase the reserve ratio to
the designated reserve ratio.
Beginning January 1,
1998, the FDIC is required to promulgate a SAIF
recapitalization schedule that achieves the DRR.
•
As long as the SAIF remains undercapitalized, until
January 1, 1998, SAIF assessments must average at least
18 basis points; thereafter, SAIF assessments must
average at least 23 basis points.
•
Assessment revenue from SAIF deposits that have been
purchased by BIF members (Oakar banks)
and from savings
associations that have converted to bank charters
(Sasser banks) is deposited in the SAIF and is not
available to the FICO.
23
•
FICO bonds are not an obligation of the FDIC, but of
the FICO.
Although the FICO is a mixed-ownership U.S.
government agency, FICO bonds do not carry the full,
faith and credit of the United States.
•
Until 2019, the last maturity date of FICO's bonds,
with the approval of the FDIC Board, the FICO has first
priority to assess savings associations that are SAIF
members to cover FICO's debt service needs.
•
In setting SAIF assessments, the FDIC Board is required
to consider the fund's expected operating expenses,
case resolution expenditures and income, the effect of
assessments on members' earnings and capital, and any
other factors the Board determines to be appropriate.
•
FICO assessments is a relevant "other factor" that the
FDIC Board may consider in setting SAIF assessments.
GOING FORWARD
Public Hearing
On Friday, March 17, the FDIC Board of Directors held an
unprecedented public hearing on the agency's proposals to reduce
deposit insurance premiums for most banks while keeping insurance
rates unchanged for savings associations.
These proposals were
issued for public comment on January 31, and although written
24
comments are not due until April 17, more than 600 comment
letters already have been received.
The FDIC Board decided that a public hearing would provide a
unique opportunity to explore all of the issues relevant to its
consideration of the proposed assessment rates, the problems
facing the SAIF, and the need for Congressional action.
The
format consisted of an open dialogue with representatives of both
BIF-insured and SAIF-insured institutions and other interested
parties.
We heard not only from the major financial institution
trade associations, but also from private citizens and individual
bank and thrift executives from both large and small
institutions.
I think I speak for the entire FDIC Board, as well as our
witnesses and many observers, when I characterize these
discussions as enlightening, thought-provoking, and extremely
beneficial.
In general there was agreement that while there is
no easy solution, there is a very real problem.
A problem that
needs to be addressed sooner, rather than later.
There was not unanimous agreement on the timing of problems
for the SAIF and the FICO bonds.
The majority of the
participants, however, conceded that a very real crisis looms on
the horizon.
One of our witnesses characterized himself as an
historian and urged us not to repeat mistakes of the past "where
25
policymakers have avoided decisions and waited for crises to
occur."
In a similar vein, others cautioned against temporizing.
I will not attempt to summarize the positions of all parties
who spoke at the hearing.4
presented and discussed.
A variety of alternatives were
These ranged from the purchase of
FDIC-issued interest-bearing obligations by SAIF-member
institutions to recapitalize the SAIF, to a one-time special
assessment on SAIF-member institutions, to use of interest on RTC
funds remaining at year-end to pay interest on the FI CO bonds, to
using the excess RTC funds in some form to meet future losses to
the SAIF, to merging the two insurance funds.
We intend to
consider the views of all of the witnesses, as well as the many
comment letters received, as we continue our analysis of the
proposed assessment rates.
One area in which I would like to believe that a consensus
was reached is a willingness by bank and thrift executives alike
Mto come to the table and talk."
To be sure, there was a
hesitancy on the part of many commercial bankers about bringing
their wallets with them, and also a suggestion that the table be
enlarged to include a broader range of financial institutions.
In fact, I think our witnesses were quite candid in expressing
4TheFDIC has a transcript of the hearing available to distribute
to all who are interested.
26
that competitive inter—industry rivalries continue to exist, that
there is a strong feeling among many banks that the SAIF "is not
our problem," and that this is a very emotionally charged issue*
It was even suggested that finding a solution that everyone can
live with may be akin to resolving the baseball strike.
We at
the FDIC certainly hope that is not the easel
Of particular interest was the testimony of individual
bankers about surviving the savings and loan crisis, the
agricultural bank crisis, and the demise of the Ohio Deposit
Guarantee Fund, to name a few.
will not be soon forgotten.
There were lessons learned that
The common thread was the effect on
financial institutions and their depositors when there is a
crisis of confidence.
Therefore, when gueried as to whether they
would be concerned if the SAIF failed, several bankers commented
that "FDIC insured" is like a prized brand name to customers —
the logo on the door of a financial institution represents
confidence —
and the integrity of that name must be preserved.
Clearly, there are no easy solutions to the problems of
capitalizing the SAIF and meeting the FI CO debt obligation, but I
am encouraged by the willingness expressed by so many of our
witnesses "to do the right thing" and to work together to find a
constructive resolution.
Several witnesses expressed their
belief that the FDIC has a "moral obligation" to bring these
problems to your attention and "the responsibility to articulate
27
a comprehensive solution to the Congress.M
I now would like to
turn to a discussion of possible legislative options.
A large number of proposals to address the SAIF problem have
been made.
In weighing the options, we must seek a real and
permanent solution, not one that simply defers the issue to a
later time while leaving in place the conditions that are the
source of the problem.
Standards
In that regard, any solution should be judged by how well it
accomplishes three goals.
First, it should reduce the premium
disparity between BIF— and SAIF—member institutions, and
to the extent possible the portion of the SAIF premium
attributable to the FI CO assessments.
This disparity encourages
SAIF members to engage in legal and regulatory maneuvering to
avoid SAIF assessments and in my view renders infeasible the
existing mechanism to fund the FICO.
This standard leaves open
the question of what level of premium disparity between BIF and
SAIF members would be small enough to eliminate the incentive for
SAIF members to flee the SAIF.
Second, it should result in the
SAIF being capitalized relatively quickly, perhaps no later than
1998.
The longer we allow the SAIF to be undercapitalized, the
greater the possibility that unanticipated losses will deplete
the fund.
Third, a solution should address the immediate problem
28
that on July 1, the SAIF will take over from the RTC the
responsibility of handling thrift failures.
Unfortunately, the
SAIF will assume this responsibility in a vulnerable and grossly
undercapitalized condition.
The progress towards capitalization,
in other words, should
be "front-loaded," with a substantial chunk of the capital coming
quickly.
We must also be concerned with the means used to achieve
these ends.
In that regard, we must consider the precedent that
is being set for the use of the deposit insurance funds.
To
ensure sufficient insurance reserves to meet future losses and to
protect the FDIC's independence, the deposit insurance funds
should be used only for deposit insurance purposes.
Ideally, the
converse should also be true that deposit insurance expenses
should not be paid out of public funds, although the savings and
loan crisis is evidence of an unfortunate breach of the latter
principle, and the diversions from the SAIF for other purposes
proves the rule about the former.
We also must carefully
consider the fairness of the solution to all concerned.
Finally,
to the extent that Congress may wish to consider options
involving the use of RTC money to address the problems outlined
here, there may be budgetary issues outside the purview of the
FDIC.
29
Options
A number of options for addressing these issues are
described below.
The options are grouped as follows: one, no
action; two, options using public funds; three, options involving
a special assessment on the SAIF assessment base; four, options
bhat would use investment income of the insurance funds to pay
bh® FXCO assessments; five, options using no public funds,
including merging the funds and sharing the FICO assessments
between BXF members and SAXF members; and six, options that
combine the above approaches.
Each option is described and
evaluated in terms of how well it achieves the three goals just
described.
Other relevant advantages and disadvantages also are
discussed.
Information about each option is presented in Table
2.
Ho Action
Without any legislative action, SAXF members would bear the
entir« $15.1 billion cost of bringing the BIF and the SAIF into
Parity (option 1 of Table 2).
Under a scenario that assumes no
major unanticipated losses, a gradual shrinkage of the SAIF
assessment base and a gradual increase in the portion of the base
ineligible for the FICO assessment, the SAIF would not reach the
designated reserve ratio until 2002.
The premium disparity would
be on the order of 19 basis points until the SAIF capitalizes.
30
After capitalization, and assuming equal expenses for the two
funds, the disparity would simply equal the basis-point
equivalent of the fixed $779-million-per-year FICO obligation.
Under the assumptions used regarding the shrinkage of the SAIF
assessment base, this would amount to 12 basis points at the time
of capitalization and would increase gradually until the FICO
bonds mature.5
Taking no action does not satisfy any of the three standards
stated above.
One, a premium disparity would continue to exist
for 24 years and would almost certainly render the existing FICO
funding mechanism obsolete.
Two, the SAIF would not capitalize
for at least seven years even assuming no major unanticipated
losses.
Three, there is no early injection of capital into the
SAIF to alleviate the immediate problem of significant
undercapitalization in the face of the requirement that the SAIF
take over from the RTC the responsibility of handling failures of
thrift institutions beginning July 1.
Approaches Using Excess RTC Fund«
It has been estimated that there will be between $10 billion
and $14 billion in RTC funds that have been appropriated but not
5The analysis in Table 2 assumes that the FDIC would set
assessments at the rate necessary to fund FICO interest payments
after the SAIF achieves its designated reserve ratio.
The law
leaves the decision to the discretion of the FDIC Board.
31
spent —
the so-called excess RTC funds.
It has been suggested
that these funds be used either to pay the FICO assessments or to
capitalize the SAIF, or some or all of both.
Two such approaches
are discussed below.
Use of Unspent RTC Funds to Pay the FICO Obligation.
Under
this approach, the FICO obligation would be paid out of excess
RTC funds.
This approach is presented in Table 2 as option 2.
The approximate cost to the Treasury of this option is $8.4
billion.
Under our proposed standards, one, there would be no premium
disparity arising from the FICO obligation and no chance of a
FICO shortfall.
Two, under this approach SAIF capitalization
would occur in 1998 assuming no large unanticipated losses,
significantly more quickly than currently expected.
Three, this
approach, however, would not address the immediate vulnerability
of the SAIF beginning July 1.
There are several other public-policy issues related to this
approach.
The Congress recognized in FIRREA that statutory draws
on the SAIF fund to support the FICO, the REFCORP, and the FRF
could result in an undercapitalized SAIF for an extended time.
Consequently the Congress authorized up to $32 billion in income
and net worth supplements for the SAIF —
appropriated.
monies that never were
In light of this legislative intent, it may be
32
appropriate for excess RTC funds to be used to pay the FI CO
obligation.
Another issue with this approach would relate to budgetary
scoring.
Under current law, deposit insurance outlays do not
trigger offsetting reductions in other federal spending or
require increased revenue; FICO assessments, however, are counted
as interest outlays rather than deposit insurance outlays.
In
this regard it should be noted that resolutions of failing banks
can often give rise to obligations that require the insurer to
make periodic payments.
Such periodic payments have been scored
as insurance outlays for budgetary purposes.
Congress may wish
to consider similarly classifying FICO assessments as insurance
outlays for budgetary purposes.
Use of Excess RTC Funds to Capitalize the SAIF.
Under this
approach, the excess RTC funds described above would be
contributed to the SAIF in the amount needed to allow the fund to
achieve its designated ratio of 1.25 percent of insured deposits
(option 3).
This would amount to $6.7 billion at year-end 1994.
Under our three proposed standards, one, this approach by
itself would do nothing to alleviate the 24-year premium
differential arising from the FICO assessments.
Without some
means to alleviate this differential, we could not rule out
further shrinkage in the SAIF assessment base, a resulting
33
increase in the premium disparity, and a deficiency in premium
income to service the FICO assessment base.
Two, the SAIF would
capitalize much much more quickly than under the status quo.
Three, the short-term vulnerability of the SAIF would be
eliminated.
As noted earlier, excess RTC funds are available to cover
insurance losses of the SAIF provided the FDIC certifies that an
increase in SAIF premiums would reasonably be expected to result
in greater loss to the Government, and that SAIF members are
unable to pay assessments to cover losses without adversely
affecting their ability to raise and maintain capital or maintain
the assessment base.
Congress required those certifications in
an effort to ensure that SAIF members pay the highest rates
possible before taxpayer funds are used to cover SAIF losses.
Of
course, this would have the effect of exacerbating the impending
premium differential.
In addition, it may be difficult for the
FDIC to certify that increasing SAIF assessments would result in
increased losses to the government prior to the SAIF being at or
near depletion.
Consequently, making RTC funds immediately
ava^la^ie to capitalize the SAIF would require modifying or
removing the existing certification requirements.
34
specified time period.
As discussed above, this would have to be
accompanied by modification or removal of the certification
requirements to provide meaningful relief from the possibility of
the SAIF being depleted.
This option for capitalizing the SAIF
is fundamentally different from others described in this
testimony in that it would involve contingent assistance rather
than upfront funded amounts.
There are substantial public-policy concerns with the
precedent set by using public funds to capitalize the SAIF.
Independence is vital to the effective functioning of the deposit
insurance system.
This does not mean freedom from accountability
but independence to constrain undue risk-taking and to protect
the insurance funds.
The exercise of safety-and-soundness
powers, pricing risk for insurance purposes, and closing and
disposing of insolvent institutions all are accomplished most
effectively when they are insulated from the political process.
Capitalization of the SAIF with appropriated money could create a
climate in which the FDIC's exercise of its insurance
responsibilities would be influenced by policy concerns outside
the scope of the FDIC's mission.
Approaches Involving a Special Assessment on the SAIF Base
Under this approach (option 4 of Table 2), a special one
time assessment that contributes to the capitalization of the
35
SAIF would be levied against the SAIF assessment base.
This
special assessment could amount to some or all of the $6.7
billion needed as of year-end 1994 to capitalize the SAIF.
In
order to collect the full $6.7 billion, a special assessment of
about 70 basis points would have to be levied over and above the
current average assessment of about 24 basis points.
The
question of how many additional thrift failures would be
triggered by such a special assessment is discussed below.
One, a special assessment would not eliminate the premium
disparity —
even if large enough to recapitalize the SAIF —
because of the continuing FICO obligation.
Two, it would
substantially reduce, or eliminate, the time needed to reach the
designated reserve ratio.
Three, it would inject funds quickly,
addressing the short-term vulnerability of the SAIF.
A special
assessment on SAIF members could act to short-circuit the types
of legal and regulatory assessment-avoidance tactics described
earlier.
To put it bluntly, a special assessment could tax SAIF
deposits before they can escape the fund.
In this regard,
Congress may wish to consider a cut-off date for a special
assessment that would ensure that institutions attempting to
avoid the assessment pay their fair share.
A special assessment
also would reduce to some extent the need for SAIF members to
engage in assessment—avoidance tactics by reducing the
capitalization component of the premium disparity.
36
If “the full $6.7 billion were not collected at once, the
SAIF would fall short of the 1.25 minimum reserve ratio.
current law this would mean that SAIF
p r e m ium s
Under
would have to
average at least 18 basis points until 1998, and at least 23
basis points thereafter, until the required reserve ratio is
achieved.
Thus, there would continue to be a premium disparity
on the order of 14 to 19 basis points until the SAIF is
capitalized, and possibly thereafter if FICO bonds remain a SAIF
obligation.
For a variety of reasons, however, if a special assessment
were levied against the SAIF assessment base, it may be
reasonable to eliminate the 18 basis—point statutory minimum
av©rage assessment rate required under current law.
Assuming
that the FICO—related premium disparity were eliminated by one of
the options described above, a premium disparity would exist
because of the need to complete the capitalization of the SAIF.
The greater the special assessment, the less would be the need
for additional assessment revenues to complete the capitalization
of the SAIF.
Table 3 shows how the size of the special
assessment (treated as an addition to the existing premiums) and
the time allowed to achieve capitalization affect the premium
necessary for the SAIF to capitalize in the desired time.
For example, under a special assessment of 30 basis points,
and assuming we wish the SAIF to reach the 1.25 reserve ratio in
37
1998, we would have to charge a SAIF premium of 15,5 basis points
and the resulting premium disparity would be approximately 11
basis points under the current proposal.
Alternatively, if we
were willing to impose a 40-basis point special assessment and
extend the deadline to capitalization to 1999, the necessary SAIF
premium would be about 9 basis points and the disparity would be
about 5 basis points.
These numbers assume that the minimum
assessment rate for BIF members would be 4 basis points, and that
there are no major unanticipated losses for either fund.
They
also assume that the FICO assessment and the current statutory
minimum assessment rates for SAIF could be eliminated.6
Depending on the size of the special assessment, a
disadvantage would be that there could be additional failures of
SAIF members as a result.
Under a one-time assessment on the
SAIF assessment base of 94 basis points, the full amount needed
to bring the SAIF to its designated ratio (70 basis point special
plus 24 basis point current assessment), three SAIF members with
total assets of $500 million would become critically
undercapitalized, based on year-end 1994 financial reports, and
another 103 SAIF members would be downgraded one notch from
current capital categories.
6If the FICO assessment were shared pro rata, both BIF and
SAIF premiums would be about 2.4 basis points higher than indicated
here.
38
Approaches Using Investment Income of the Insurance Funds to
Pav the Pico
There have been a number of proposals to use investment
Income of the Insurance funds to pay the FXCO assessments.
Two
such proposals are considered here as option 5 of Table 2.
One
proposal would inject RTC funds into the SAIF in the amount
needed to achieve the 1.25 reserve ratio.
The interest on the
SAIF's investment portfolio would then be used to pay a portion
th® FICO assessments.
With a fully invested fund at today's
interest rates, this would yield approximately $600 million
annually as compared with the $779 million required to meet FICO
debt service obligations.
Another option that has recently been proposed would allow
investment income equal to two basis points of the BIF assessment
base to be used to pay the FICO assessments.
Based on the
current BIF assessment base, about $500 million of the $779
million annual FICO assessment would be paid by the BIF under
this approach.
first option does not constitute a complete solution to
problems posed by the difference in the condition of the two
funds, but simply changes the form in which the FICO assessment
be paid by the SAIF industry.
Instead of being paid by the
SAIF members through assessments, the FICO would be serviced by
39
garnishing the SAIF7s income.
If the BIF and the SAIF started at
the same reserve ratio, had the same loss experience going
forward, and maintained their respective 1.25 ratios, SAIF
premiums would have to be higher than BIF premiums by a
sufficient amount to offset the drain in the SAIF's income caused
by the FI CO service.
Otherwise, if there were no premium
differential, the BIF reserve ratio would increase continuously
relative to the SAIF reserve ratio during the full 24-year period
in which the FICO bonds are outstanding, and SAIF members would
have to be assessed higher premiums to make up the difference if
losses to the SAIF dropped the balance below the 1.25 ratio.
The advantage of the approach is delaying the SAIF premium
increase until justified by losses.
On the other hand, over the
long term, this approach does not address the first standard set
out above, address the premium disparity arising from the FICO
assessment, as well as the incentive of SAIF members to avoid
these assessments, and the resulting difficulties in funding the
debt •
Our proposed standards two and three are m e t , because
the SAIF would be capitalized immediately.
Looking at the approach involving BIF investment income,
first, a premium differential arising from FICO assessments would
exist to the extent the SAIF's share of the remaining
Por“tion of the FICO assessment is greater than the investment
income of the SAIF.
Based on the current assessment bases of the
40
two funds, the SAIF would pay about two basis points more than
the BIF for its share of the FICO assessment.
This differential
could change over time if the BIF and SAIF assessment bases grew
at different rates.
The differential is not likely to be
substantial, but could increase somewhat over time.
Two, this
option would capitalize the SAIF in 1999 under current
conditions.
Three, it would do nothing to address the short-term
vulnerability of the SAIF.
Using investment income of the BIF to pay FICO assessments
would set a precedent for using BIF funds to pay expenses not
related to the BIF, although use of only investment income would
be a more limited precedent.
In addition, diverting investment
income of the BIF would increase the likelihood that assessment
for BIF members would have to be increased at some future
time to replace the contribution investment income would have
made to covering losses to the BIF from failed banks.
Use Of M o Pub l i c fnnaa
Options 6 and 7 in Table 2 present two approaches that rely
solely on FDIC-insured institutions to raise some or all of the
$15.1 billion needed to bring the SAIF into parity with the BIF.
These are sharing the FICO assessments between the BIF and the
SAIF without merging the funds (option 6) and merging the BIF and
the SAIF (option 7).
41
The BIF Share of the FICO Obligation Without a Merger.
Under this option, the BIF members would be assessed for a
portion of the FICO assessments.
For example, a pro rata sharing
of the FICO assessments between the BIF and the SAIF, based on
insured deposit levels in the two fluids, would cost BIF members
about $6.5 billion in present-value terms.
The BIF's share of
the annual $780 million obligation would be about $600 million,
or 2.4 basis points per year because 77 percent of the total
domestic deposits of FDIC-insured institutions are held by BIF
members, and 23 percent by SAIF members.
Under our proposed standards, this approach would, one,
eliminate any premium disparity arising from the FICO obligation,
currently about 11 basis points of the proposed 19 basis point
differential.
By making the entire assessment base of both funds
available to service the FICO debt, it would virtually rule out a
deficiency of premium income to service the FICO assessment.
Two, this approach would enable the SAIF to capitalize
significantly more quickly than currently anticipated by
eliminating most of the FICO drain on SAIF assessment revenue.
Assuming no large unanticipated losses, capitalization would
occur in 1999, three years earlier than currently projected.
Three, this approach would do nothing to address the concern
that the SAIF will begin resolving thrift failures on July 1 in a
significantly undercapitalized position and remain there for
42
several years.
This makes the SAIF very vulnerable to
unanticipated losses.
It thus leaves open the possibility that
the SAIF could be bankrupted and that both SAIF- and BIF-insured
institutions would suffer from the resulting negative publicity.
The other concern with this approach has already been discussed.
By using BIF funds for purposes other than paying for deposit
insurance costs, this approach sets a precedent that could erode
the effectiveness and independence of the deposit insurance
system.
Another alternative for this approach would be for the BIF
to contribute 50 percent of the cost of servicing the FI CO
obligation (option 6(b) of Table 2).
This currently would amount
to approximately 1.5 basis points annually for BIF members, or
about a $4.2 billion present-value cost.
Under our proposed standards, this approach, one, would not
eliminate the premium disparity.
Unlike the pro rata sharing
approach, this approach retains a 24—year premium disparity,
although at lower levels than some other options.
To illustrate,
with the 50 percent sharing described here, equal shares of the
annual FICO cost by the BIF and the SAIF of $390 billion would
amount to about 1.5 basis points for BIF members and 5.5 basis
points for SAIF members.
Thus, after the SAIF is capitalized,
there would remain a premium disparity of about four basis points
43
could grow larger if the SAIF assessment base were to
shrink.
Two, this approach would not achieve SAIF capitalization as
quickly as the alternative in which the BIF shares the FICO
assessments on a pro rata basis —
2000 rather than 1999 — , thus
leaving the SAIF undercapitalized for one more year.
Three, this
option also does not address the short-term vulnerability of the
SAIF.
In addition, this approach sets a precedent by using BIF
resources for other purposes•
BIF members probably would argue,
however, that equal dollar sharing is less unfair than
proportional sharing because it entails less use of BIF
resources.
Nftrging the BIF and the 8 AIF.
Under this option, the two
funds would be combined and the existing premium rates maintained
^ t i l the combined fund meets the designated reserve ratio.
assessments would continue to be paid by the thrifts.
FICO
The
designated reserve ratio for the combined fund could be expected
to be achieved in 1996.
The cost to the BIF of this approach is estimated at $5.5
billion, or the equivalent of a one-time charge of 22 basis
points on the BIF assessment base.
By our proposed standards,
44
one, there would be no premium disparity until capitalization of
the combined fund occurred.
At capitalization the disparity
would equal the size of the fixed $779 million FICO charge
relative to the SAIF assessment base.
This would be about 11
basis points in 1996, assuming no drastic change in the SAIF
assessment base during the next year.
This option meets standard two and three because there is an
immediate and substantial capital injection into the SAIF and the
combined fund recapitalizes quickly.
The resulting ll-basis
point disparity, based on the current SAIF assessment base, would
nevertheless appear large enough to provide an incentive for
further legal and regulatory maneuvering by SAIF members to avoid
assessments.
If successful, SAIF assessment revenue would prove
insufficient to fund the FICO earlier than otherwise.
Merging the funds would set an unfortunate precedent for the
use of the resources of the deposit insurance funds —
case the BIF.
in this
Existing law requires that BIF resources be used
to cover only BIF expenses; merging the funds would violate that
principle.
There is a danger in overriding the law governing the
use of insurance fund resources solely for the sake of
63fpediency•
If an insurance fund's resources can be used for
purposes other than protecting the depositors of that fund, where
should we draw the line about what charges to deposit insurance
reserves are appropriate?
Such "other uses" of deposit insurance
45
funds weaken 'the distinction between those funds and general
federal monies and pose a danger to the independence of the
deposit insurance system.
Moreover, there is a significant
question of fairness to BIF member banks, who have paid $22
billion during the last four years to recapitalize the BIF at the
level mandated by the Congress.
Finally, the current problem of
capitalizing the SAIF as a result of the diversions of SAIF
assessment revenue for other purposes illustrate the effect of
using deposit insurance funds for other purposes.
Comtek nation Options
This section presents some options that involve combinations
of the approaches outlined above.
8 in Table 2.
These are grouped under option
All of these options share a common theme: they
are designed to enhance some of the approaches above that did not
address the long-term premium disparity arising from the FICO
assessments•
The first such option involves merging the funds and having
BIF and SAIF share the FICO assessments proportionately.
The
most important shortcoming of merging the funds would be that,
taken by itself, it would do nothing to resolve the 2 4-year
premium disparity.
By providing that the FICO burden be shared
proportionately between current BIF and SAIF members this problem
could be mitigated.
The cost to the BIF would be $11.7 billion
46
or the equivalent of a one-time charge of 47 basis points on the
BIF assessment base.
This option would entail proportional
sharing between the BIF and the SAIF of the total $15.1 billion
cost of bringing the two funds into parity.
Under this approach, there would be no premium disparity,
and, because the SAIF would be capitalized quickly, there would
be an up-front substantial injection of funds.
therefore, meet our three standards.
It would,
On the other hand, as
emphasized above, there would be an unfortunate precedent set in
using the BIF for purposes other than BIF insurance costs.
The second option would be to combine RTC capitalization of
the SAIF with a pro rata sharing of the FICO assessments between
BIF and SAIF.
The drawback in using the excess RTC funds to
capitalize the SAIF is that such an approach by itself would not
alleviate the long-term premium disparity arising from the FICO
assessments.
This problem could be alleviated by combining this
approach with a pro rata sharing of the FICO assessments between
the BIF and the SAIF.
This approach would eliminate the premium
disparity and would result in an immediate capitalization of the
SAIF, thus meeting our proposed standards.
As emphasized above,
however, these advantages come at a cost: the use of public funds
and all that entails for the independence of the deposit
insurance system.
47
A special assessment on the SAIF assessment base, either in
combination with a BIF and SAIF sharing of the FI CO or with
excess RTC funds being used to pay the FICO assessment
constitutes the third and fourth options.
A special assessment
by itself does nothing to resolve the premium disparity arising
from the FICO assessments.
this problem.
Either two approaches could correct
Either of these two approaches are presented in
Table 2 under the assumption that the entire $6.7 billion needed
for the SAIF to achieve the reserve ratio is collected at once
through a special assessment.
Approaches involving smaller
special assessments were discussed above (see Table 3 and the
accompanying discussion).
Both approaches have advantages.
One,
there would be no long-term premium disparity; two and three, the
SAIF is capitalized immediately.
CONCLUSIONS
There is an urgent need for legislative action to reduce the
disparity in the financial condition of the BIF and the SAIF.
This immediate need arises from three sources.
First, on July 1
the SAIF will assume the responsibility for handling failures of
^ r^tt institutions.
It will not assume this responsibility in a
position of strength, because it is grossly undercapitalized.
This condition is directly attributable to the fact that until
1993, most assessment revenues from SAIF members were statutorily
diverted from the SAIF to pay for past losses related to the
48
thrift: crisis.
In addition, revenue and net worth supplements
totalling $32 billion that Congress had authorized for the SAIF
never were appropriated.
As a result of this history, the
existing SAIF balance simply does not provide an adequate margin
of comfort.
The resources of the SAIF are insufficient to absorb
the cost of the failure of one large or a few medium-sized
thrifts, or other substantial unanticipated losses.
Second, as a result of the SAIF's significant
undercapitalization, there can be no assurance that the Congress
will not again have to address these issues.
If there are no
major unanticipated losses, the SAIF balance should inch up to
its target over the next seven years.
Over this length of time,
it is difficult to take comfort that unanticipated losses will
not prevent the SAIF from reaching its target.
The longer the
time before the SAIF capitalizes, the greater the chance the SAIF
might fail to capitalize.
Third, the current structure for funding the FICO obligation
is not viable.
Requiring this fixed cost to be paid from deposit
insurance assessments on the SAIF creates enormous economic
incentives for the targeted group to engage in legal and
regulatory maneuvering to reduce their potential costs.
We are
already seeing such maneuvering in the current interest expressed
by some large thrifts in opening new banks and by applications
from thrifts to operate branches that would share bank and thrift
49
operations.
As stated earlier, the question is not whether there
will be insufficient premium income to service the FICO
obligations, but when the deficiency will occur.
Any solution to these problems should address all three
concerns.
It should eliminate the long-term premium differential
caused by the FICO assessments.
It should greatly reduce the
time needed to capitalize the SAIF.
The longer the SAIF is
allowed to remain undercapitalized, the greater the chance that
unanticipated losses will prevent us from reaching the target or
will force Congress to consider these issues again.
Finally, the
solution should include an immediate injection of funds into the
SAIF or a ready source of bac)cup funding for SAIF losses.
As
matters stand now, the SAIF will begin its responsibilities for
handling thrift failures after June 30 in a dangerously
vulnerable condition.
Madam Chairwoman, the FDIC is committed to finding solutions
that address these three concerns in a manner that is consistent
with good public policy;
We stand ready to assist the
Subcommittee in this effort in the weeks ahead.
I commend your
forsightedness in holding this hearing, and I look forward to
your questions and to questions from members of the Subcommittee.
TABLE 1
Proposed Assessment Rate Schedules
Second Sem iannual 1995 A ssessm ent Period
FD IC-lnsured Institutions
Proposed BIF Rate*
Capital
Category
Supervisory R isk Group
Group A
Group B
Group C
4
7
14
28
21
28
31
1. Well
2. Adequate
3. Under
7
14
Estimated Annual Assessment Revenue: $1.1 Billion
Average Annual Assessment Rate: 4.5 bp
Rate Spread: 27 bp
Proposed SAIF Rate*
Capital
Category
Supervisory R isk Subgroup
Group C
Group A Group B
1. Well
2. Adequate
3. Under
23
26
29
26
29
30
29
30
31
Estimated Annual Assessment Revenue: $1.7 Billion
Average Annual Assessment Rate: 24 bp
Rate Spread: 8 bp
Rates are in basis points
Resulting obligation to:
(basis points in parentheses)
nco*
BIF
SAIF
Treasury
SAIF*
Capitalization
Premium*
Disparity
Problem
a) Merge BIF/SAIF, BIF/SAIF pay FICO £rg rata
11.7 (47)
3.4 (47)
0
1996
none
none
b) RTC capitalizes SAIF, SAIF/BIF share FICO
pro rata
6.5 (26)
1.9 (27)
6.7
immediate
none
none
c) Special assessment on SAIF, BIF/SAIF share
FICO pro rata
6.5 (26)
8.6 (120)
0
immediate
none
none
d) Special assessment on SAIF, RTC funds pay
0
6.7 (94)
8.4
immediate
none
none
Option
8. Combination Options
nco
* Based on baseline assumptions as of 3-9-95.
Notes:
(i) Current estimated cost to defease FICO = $8.4 billion, discounting at 7.96%, the current (3/7/95) rate on Treasury IO strips due in 2018.
(ii) Amount needed at year-end 1994 to enable SAIF to meet 1.25% designated ratio = $6.7 billion (the amount neededto incurease the SAIF
from its year-end 1994 balance of $1,936 billion to 1.25% of year-end SAIF insured deposits of $693 billion, or $8.66 billion).
(iii) Pro rata shares are 77% BIF, 23% SAIF.
(iv) The one-time premium necessary to obtain $1 billion is about 4 basis points for BIF and 14 basis points for SAIF.
(v) Reducing SAIF premium from 23 bp to 18 bp would increase time to capitalization by about two years.
(vi) Results in right hand three columns rely on FDIC’s baseline assumptions.
Table 2
Options For Resolving Issues Related to the SAIF
($ billions)
Resulting obligation to:
(basis points in parentheses)
Option
1. No Action
BIF
SAIF
Treasury
SAIF*
Capitalization
Premium*
Disparity
FICO*
Problem
$0
$15.1 (211)
$0
2002
19 bp before, 12 bp at recap date
likely within 10 years
2. Use unspwit RTC appropriations to pay FICO
a) SAIF capitalizes itself
none
0
6.7 (94)
8.4
1998
19bp before, 0 after
none
0
8.4 (118)
6.7
immediate
11 bp at recap date
less likely but possible
0
15.1 (211)
0
immediate
11 bp at recap date
less likely but possible
0
8.4 (118)
6.7
immediate
uncertain
uncertain
5.3 (21)
9.8 (137)
0
1999
uncertain
uncertain
a) Erg rata shares
6.5 (26)
8.6 (120)
0
1999
16.5 bp before, 0 after
none
b) 50% shares
4.2 (17)
10.9 (153)
0
2000
17.5 bp before, 4 bp at recap date
less likely but possible
5.5 (22)
9.6 (134)
0
1996
0 bp before, 11 bp at recap date
less likely but possible
3. Use unspent RTC appropriations to capitalize
SAIF
a) SAIF pays FICO
4. Special assessment on SAIF to capitalize SAIF
a) SAIF members continue to pay FICO
5. Use BIF or SAIF investment income to pay FICO
a) RTC funds pay SAIF; SAIF investments pay
FICO
b) 2 bp of BIF investment income pays part of
FICO
6. BIF/SAIF pay FICO, no merger
7. Merge BIF and SAIF
a) SAIF pays FICO
• Based on baseline assumptions as of 3-9*95.
Table 3
Assessment Rate Necessary to Capitalize the SAIF by Given Year
Under Various 1995 Special Assessments
Special
Assessment
1998
1999
2000
2001
10 bp
23.0 bp
18.0 bp
14.5 bp
12.5 bp
20
19.0
15.0
H
to
o
1995
Assessment Rates
10.5
30
15.5
12.0
9.5
8.0
40
11.5
9.0
7.0
6.0
50
8.0
6.0
4.5
4.0
Resulting obligation to:
(basis points in parentheses)
BIF
SAIF
Treasury
SAIF*
Capitalization
Premium*
Disparity
FICO*
Problem
a) Merge BIF/SAIF. B1F/SAIF pay FICO £rg rats
11.7 (47)
3.4 (47)
0
1996
none
none
b) RTC capitalizes SAIF. SAIF/B1F share FICO
pro rata
6.5 (26)
1.9 (27)
6.7
immediate
none
none
c) Special assessment on SAIF. BIF/SAIF share
FICO pro rata
6.5 (26)
8.6 (120)
0
immediate
none
none
d) Special assessment on SAIF. RTC funds pay
0
6.7 (94)
8.4
immediate
none
none
Option
8. Combination Options
nco
* Based on baseline assumptions as of 3-9-95.
Notes:
(i) Current estimated cost to defease FICO = $8.4 billion, discounting at 7.96%, the current (3/7/95) rate on Treasury IO strips due in 2018.
(ii) Amount needed at year-end 1994 to enable SAIF to meet 1.25% designated ratio = $6.7 billion (the amount needed toincurease the SAIF
from its year-end 1994 balance of $1,936 billion to 1.25% of year-end SAIF insured deposits of $693 billion, or $8.66 billion).
(iii) Pro rg& shares are 77% BIF, 23% SAIF.
(iv) The one-time premium necessary to obtain $1 billion is about 4 basis points for BIF and 14 basis points for SAIF.
(y) Reducing SAIF premium from 23 bp to 18 bp would increase time to capitalization by about two years.
(vi) Results in right hand three columns rely on FDIC’s baseline assumptions.
ATTACHMENT A
LEGISLATIVE HISTORY OF SAIF FUNDING SCHEME
This legislative history reviews the primary statutes that
established the funding scheme intended by Congress to resolve
the thrift crisis of the 1980s and to provide capital to the
Savings Association Insurance Fund (SAIF) and its predecessor,
the Federal Savings and Loan Insurance Corporation (FSLIC).
Although these laws cover a broad range of issues with respect to
insured institutions, this review is limited to those provisions
that concern the funding of the FSLIC and the SAIF.
Background
From the inception of federal deposit insurance, insured
banks and thrifts were charged a flat rate for deposit insurance.
That flat-rate system generated sufficient revenue to cover the
costs of failures through the mid-1980s when bank and thrift
failures began to escalate rapidly.
In 1987, the Federal Home
Loan Bank Board (FHLBB) , as the agency with oversight
responsibility for the FSLIC, the thrift insurance fund,
announced that the FSLIC was insolvent.
Competitive Equality Banking Act of 1987
Congress passed the Competitive Equality Banking Act of 1987
(CEBA) against a backdrop of an increasing rate of thrift
failures.
One of the primary purposes of CEBA was to
recapitalize the FSLIC through a combination of capital market
borrowings and thrift industry contributions.
CEBA authorized
the FHLBB to charter the Financing Corporation (FICO) to issue
bonds in the capital markets, the net proceeds of which were used
to purchase redeemable nonvoting capital stock and nonredeemable
capital certificates of the FSLIC.
The FICO was authorized to
sell up to $10,825 billion in 30-year bonds to the public.
Of
that amount, $10 billion was to be used for FSLIC operations and
the remainder was to replace secondary reserve losses.
The FICO
issued 30-year non-callable bonds in a principal amount of
approximately $8.1 billion which mature in 2017 through 2019.
The principal amount of the FICO debt was to be paid by the
Federal Home Loan Banks (FHLBs). To cover interest costs, the
FICO was authorized to impose on each institution insured by the
FSLIC, both a regular assessment not to exceed 8.3 basis points
and, if required, a supplemental assessment not to exceed 12.5
basis points.
The FICO assessment was to be subtracted from the
insurance premium of 8.3 basis points charged by the FSLIC.
If
2
the full amount of the regular assessment authorized had been
assessed by the FICO, no funds would have remained to replenish
the FSLIC.
No institution could be required to pay more than the
maximum regular and supplemental assessment amounts, whether paid
to the FSLIC, the FICO or a combination of both.
The FICO's
assessment authority does not expire until 2019, the maturity
year of its last bond issuance.
A key element of the capitalization scheme was the
moratorium on changing insurance funds established in CEBA.
By
prohibiting thrifts from leaving the FSLIC, the moratorium
provided the FSLIC with a captive funding source so that the fund
could be built up.
In addition, it ensured that FSLIC members
would bear the burden of paying interest on the bonds issued by
the FICO, thereby contributing toward the payment of the fund's
past losses.
CEBA also provided the FICO with authority (with
FHLBB approval) to levy an exit fee on insured institutions that
terminated their FSLIC insurance.
Financial Institutions Reform. Recovery, and Bnforepwent; Agfc
In 1989, with losses from thrift failures continuing to
mount and the condition of the bank insurance fund beginning to
deteriorate, Congress enacted the Financial Institutions Reform,
Recovery, and Enforcement Act (FIRREA) for the purpose of
reforming, recapitalizing and consolidating the federal deposit
insurance system by 1) placing the deposit insurance funds on a
solid financial footing and 2) strengthening the supervisory and
enforcement authority of federal bank and thrift regulators.
FIRREA restructured the deposit insurance funds by
abolishing the FSLIC and establishing in its place the SAIF,
which was to be managed by the FDIC.
The FDIC's Permanent
Insurance Fund was renamed the Bank Insurance Fund (BIF) . FIRREA
established a designated reserve ratio (DRR) for each fund set at
1.25 percent of estimated insured deposits and directed the FDIC
to set rates and the DRR for the BIF and the SAIF independently.
FIRREA also departed from the previous flat-rate assessment
system by establishing a schedule of minimum annual assessment
rates for both BIF and SAIF members. The FDIC was authorized to
increase the minimum rates as necessary to achieve the DRR, but
the rate could not exceed 32.5 basis points, nor could it be
increased by more than 7.5 basis points in any one year.
Until
1998, the minimum assessment schedule set for SAIF members was
higher than that for BIF members, ranging from a difference of
approximately 12.5 basis points at enactment to 3 basis points
through 1997.
To continue to ensure a captive source of assessments to the
SAIF, FIRREA extehded for an additional five years the moratorium
3
on changing insurance funds with certain exceptions for troubled
institutions and for transfers of "an insubstantial portion of
total deposits," typically involving sales of branches by healthy
institutions.
FIRREA further established entrance and exit fees
to be paid by institutions that engaged in permissible transfers
between insurance funds.
Any institution that transfers deposits
from the SAIF to the BIF must pay an entrance fee to the BIF to
prevent dilution of the BIF reserve ratio and an exit fee to the
SAIF (currently 90 basis points). Exit fees received in
connection with transfers from the SAIF to the BIF are held in a
segregated account and may be made available to the FICO if the
FDIC and the Secretary of the Treasury determine that the FICO
has exhausted all other sources of funding for interest payments
on its bonds.
One of the exceptions to the moratorium authorized a bank
holding company that controlled a savings association to merge
the savings association with a subsidiary bank.
These so-called
"Oakar" banks pay premiums to the SAIF on deposits attributable
to the former savings association (the adjusted attributable
deposit amount). The moratorium did not affect the ability of
thrift institutions to convert to bank charters so long as the
resulting institution remained a member of the SAIF ("Sasser"
banks).
The funding framework established in FIRREA to pay for the
escalating cost of thrift resolutions created three new entities,
the FSLIC Resolution Fund (FRF), the Resolution Trust Corporation
(RTC) and the Resolution Funding Corporation (REFCORP). The FRF
was created to liquidate the assets and liabilities of the FSLIC.
The FRF paid to the SAIF all amounts needed for administrative
and supervisory expenses from creation of the SAIF through
September 30, 1992.
The FRF received funds from amounts assessed
against SAIF members by the FDIC that were not required for
principal payments on bonds issued by the REFCORP or interest
payments on bonds issued by the FICO.
FIRREA established the RTC to manage and resolve all
troubled thrift institutions previously insured by the FSLIC as
well as future thrift Resolutions through August 9, 1992.
This
date was subsequently extended to June 30, 1995.
Since enactment
of FIRREA, the SAIF's resolution responsibility has been limited
to the SAIF-insured portion of BIF-member Oakar banks and thrifts
chartered since 1989.
The SAIF will assume resolution
responsibility for thrifts on July 1, 1995.
Finally, pursuant to FIRREA, the REFCORP was created to
provide funding for the RTC by issuing bonds.
The principal of
REFCORP bonds was to be paid by the FHLBs, up to a maximum annual
amount of $300 million or 20 percent of net earnings per FHLB.
To the extent that monies from the FHLBs were insufficient to pay
the principal amount, with the approval of the Board of Directors
4
of the FDIC, the REFCORP was authorized to assess SAIF members.
The amount of REFCORP's assessment could not exceed the amount
authorized to be assessed by the FDIC, less any FICO assessment.
Under the funding scheme established in FIRREA, the FICO
continued to retain first priority on SAIF assessments followed
by the REFCORP and the FRF, limited by the maximum amount
authorized to be assessed by the FDIC.
If the FICO, REFCORP and
FRF assessments exhausted the amount of the FDIC's authorized
assessment, then no funds were available to deposit in the SAIF.
Congress recognized in FIRREA that the diversion of SAIF
assessments to the FICO, REFCORP and FRF would necessarily delay
the capitalization of the SAIF.
Therefore, in addition to
assessment revenue, Congress authorized the appropriation of
funds to the SAIF in an aggregate amount of up to $32 billion to
supplement SAIF revenue and to maintain a statutory minimum net
worth.
Congress authorized an annual appropriation to SAIF to
supplement assessment revenue by ensuring an income stream of $2
billion (after subtracting the amounts diverted to the FICO,
REFCORP and FRF) each year through 1999, not to exceed $16
billion in the aggregate, and to meet statutorily mandated
minimum net worth targets through 1999, not to exceed $16 billion
in the aggregate.
Subsequent legislation extended the date for
receipt of appropriated funds to 2000.
Federal Deposit Insurance Corporat-•»rm Tmprovew»gn^ *ct of 1991
In December 1991, faced with continuing bank and thrift
failures and the impending bankruptcy of the BIF, Congress passed
the Federal Deposit Insurance Corporation Improvement Act of 1991
(FDICIA) . In FDICIA, Congress focused its efforts on preventive
actions to protect the insurance funds by 1) requiring a variety
of regulatory and supervisory measures intended to limit the risk
of loss to the insurance funds and 2) restructuring the deposit
insurance assessments system.
FDICIA restructured completely the basis upon which deposit
insurance assessments are determined by replacing the flat-rate
assessment system with a risk-related assessment system in which
an institution's insurance premium is a function of the risk
posed to the applicable fund by that institution.
Congress
intended the system to serve as an incentive to curtail
activities that posed a greater risk to the funds.
In addition
to the implementation of a risk-related system, Congress
authorized the FDIC to set assessments to maintain the reserve
ratio at the DRR once that level is achieved.
However, until
that time, the FDIC is required to set rates not lower than the
statutory minimum assessments.
Currently, SAIF members are
5
assessed risk-related rates ranging from 23 basis points to 31
basis points, which is higher than the statutory minimum
assessment of a weighted average of 18 basis points.
If the SAIF
is not recapitalized by January 1, 1998, or if the SAIF has
outstanding Treasury borrowings on that date, the FDIC must
promulgate a recapitalization schedule for the SAIF and the
statutory minimum assessment will increase to a weighted average
rate of 23 basis points.
Finally, FDICIA reaffirmed that FICO
assessments must be subtracted first from the assessments
established by the FDIC for SAIF members.
In early 1992, because of the continuing weak position of
the SAIF, the FDIC asked the Treasury Department and the Office
of Management and Budget to request funding for the revenue and
net worth supplements authorized under FIRREA.
Despite these
requests, no funds were ever requested or appropriated for these
purposes.
Finally, to provide additional avenues for resolution of
troubled institutions, Congress broadened the "Oakar" exception
to the moratorium on conversions to permit acquisitions by banks
not in a holding company structure and to enable SAIF-insured
institutions to acquire BIF-insured institutions.
The resulting
SAIF-insured institution would pay assessments to the BIF for the
deposits attributable to the former BIF member.
In 1992, the FDIC Legal Division determined that as a matter
of law assessments paid by BIF-member Oakar banks on deposits
acquired from SAIF members must remain in the SAIF and may not be
allocated among the FICO, REFCORP, or FRF. The FDIC General
Counsel recently reaffirmed this opinion and further stated the
Legal D i v i s i o n s position that assessments paid by any former
savings association that has converted to a bank and remains a
SAIF member (Sasser banks) are not available to the FICO.
(See
Notice of FDIC General Counsel's Opinion No. 7, 6 0 FR 7055
(Feb. 6, 1995)).
Resolution Trust Corporation Completion Act
The Resolution Trust Corporation Completion Act (RTCCA) was
enacted in 1993 to "provide for the remaining funds needed to
assure that the United States fulfills its obligation for the
protection of depositors at savings and loan institutions. . ."
and to provide the final funding for the RTC.
The RTCCA extended
the moratorium on transfers between insurance funds to such time
as the SAIF first attains the DRR and authorized the FDIC to
extend any SAIF recapitalization schedule beyond the 15-year time
limit specified in FDICIA to a date that will maximize the amount
of semiannual assessments received by SAIF.
The RTCCA also
replaced the revenue and net worth supplements authorized in
FIRREA with an authorization to use up to $8 billion of
6
appropriated funds for losses incurred by the SAIF in fiscal
years 1994 through 1998.
In addition, the RTCCA authorized the
use by SAIF of unexpended RTC funds for losses incurred or
reasonably expected to be incurred.
In both cases, these funds
can be received only if the FDIC certifies to Congress that 1)
assessments on SAIF members cannot be increased further without
causing additional losses to the Government and 2) SAIF members
cannot pay higher assessments to cover losses to the SAIF without
adversely affecting their ability to raise and maintain capital
or to maintain the assessment base.
ATTACHMENT B
FLOWS IN AND OUT OF THE SAVINGS ASSOCIATION INSURANCE FUND
Summary
The Savings Association Insurance Fund (SAIF) was created by the Financial Institutions
Reform, Recovery, and Enforcement Act of 1989 (F1RREA) as the successor to the Federal
Savings and Loan Insurance Corporation (FSUC). The SAIF will have resolution authority for
all failed thrifts as of no earlier than January 1, 1995, nor later than July 1, 1995.
Consequently, there have been limited demands on the SAIF for insurance losses since its
inception. Through 1992, the majority of SAIF-member assessment revenue was diverted to
thrift resolution funding needs other than the SAIF. Because uses of funds have nearly equalled
sources of funds since 1989, the SAIF began Fiscal Year 1993 with a minimal balance of
approximately $200 million.
The SAIF received nearly $0.17 billion in net assessment revenue in Fiscal Year 1993. For
many reasons, it is difficult to project the SAIF’s exposure. If insurance losses exceed
assessment revenue, the SAIF may be required to access its other sources of funds. Under the
Resolution Trust Corporation Completion Act of 1993 (RTCCA), up to $8 billion total is
authorized to be appropriated for Fiscal Years 1994 through 1998 to the SAIF for loss purposes
only, subject to certifications by the FDIC’s Board of Directors. Absent such funds, the SAIF
would be authorized to borrow from the FDIC’s $30 billion credit line with the Treasury, with
borrowings to be repaid over time with SAIF-member assessments.
Background
FIRREA abolished the FSLIC and the Federal Home Loan Bank Board (FHLBB). Their
functions were transferred to the FDIC, the Office of Thrift Supervision (OTS), the Federal
Housing Finance Board (FHFB), and the RTC. Under FIRREA, the FDIC became the
administrator of two separate and distinct insurance funds: the Bank Insurance Fund (BIF),
formerly the Deposit Insurance Fund, and the SAIF, the successor to the FSUC Fund. Both
insurance funds are maintained separately to carry out their respective legislative mandates, with
no commingling of assets or liabilities.
A third separate fund under FDIC management is the FSUC Resolution Fund (FRF). The
FRF is funded through assessment revenue from SAIF-member institutions (through calendar
year 1992) Congressional appropriations and asset sales. The RTC will resolve all troubled
thrift cases that occur from January 1, 1989 through at least December 31, 1994, but not later
than June 30, 1995, after which the SAIF will resolve all new thrift cases. The FRF will
complete the resolution of all thrifts that failed or were assisted before January 1, 1989, and also
will complete the resolution of any RTC conservatorships that are unresolved as of the RTC’s
termination in December 1995. Resolution responsibility is summarized in Table 1.
2
Prim ary Sources and Uses of Funds
The primary source of funds for the SAIF is assessment revenue from SAIF-member
institutions. Since the creation of the SAIF and through the end of calendar year 1992,
however, almost all assessments from SAIF-member institutions were diverted to other needs,
as stipulated by FERREA and as described below. Only SAIF assessment revenue generated
from BIF-member institutions that acquired thrifts under Section 5(d)(3) of the Federal Deposit
Insurance Act (FDI Act), Le., Oakar banks, was deposited in the SAIF throughout this period.
Through 1992, assessment revenue from SAIF-member institutions was diverted to the
Financing Corporation (FICO)1, the Resolution Funding Corporation (REFCORP)2, and the
FRF. Under Section 21 of the Federal Home Loan Bank (FHLB) Act, the FICO has an ongoing
first claim on SAIF assessment income through the year 2019 to fund the interest payments on
bonds issued by the FICO. Section 21 of the FHLB Act also requires that SAIF assessment
income be used, if necessary, to provide funding for REFCORP.3 Because REFCORP’s
principal fund is fully funded, SAIF assessment income was not required for REFCORP
purposes in 1992. During the period beginning on the date of enactment of FIRREA and ending
on December 31,1992, Section 11 of the FDI Act requires that "to the extent funds are needed"
the sources of funds for the FRF shall include amounts assessed against SAIF members by the
FDIC pursuant to Section 7 that are not required by FICO or REFCORP. Table 2 summarizes
the flow of assessment revenue from 1989 through 1994.
Because most of SAIF’s assessment revenue has been diverted since its inception, net
revenue to the SAIF has been limited. However, there have been only limited demands on the
SAIF, as losses were small and it was reimbursed by the FRF for administrative and supervisory
expenses through September 30, 1992. SAIF’s balance as of Decembrer 31, 1994, was $1.9
billion and the fund is not expected to reach the 1.25 percent reserve ratio until 2002.
As noted above, assessment revenue net of the FICO obligation began flowing into the SAIF
on January 1,1993. SAIF now is obligated to fund its administrative and supervisory expenses,
although die draw is relatively minor. By not sooner than January 1, 1995, and not later than
'The FICO was created by the Competitive Equality Banking Act of 1987 (CEBA) as a
mixed-ownership government corporation to recapitalize the FSLIC; its funds were used by the
FRF after the enactment of FIRREA. The FICO’s authority to issue obligations was terminated
on December 12,1991, by the Resolution Trust Corporation Thrift Depositor Protection Reform
Act of 1991.
2The REFCORP was created by FIRREA as a mixed-ownership government corporation to
provide funding for the RTC.
3SAIF assessment revenue was used to purchase zero-coupon bonds to repay the REFCORP
obligations at maturity.
3
July 1, 1995, the SAIF will have responsibility for all new thrift resolutions.
Table 1
RESOLUTION RESPONSIBILITY:
SAIF-MEMBER INSTITUTIONS
FRF
RTC
SAIF
Thrifts previously insured by FSLIC:
Fail prior to 1/89
Fail 1/89 - at least 12/31/94 but not
later than 6/30/95
Fail after not sooner than 1/1/95
and not later than 7/1/95
X
X
X
Conversion banks: Oakars and Sassers4
Fail after 8/89
X
Thrifts chartered post-FIRREA
Fail after 8/89
X
Thrifts in RTC conservatorship, and all remaining
RTC assets and liabilities as of December 1995
X
4Under the "Oakar" Amendment, insured depository institutions are allowed to merge
without changing insurance coverage for the acquired deposits. Oakar bank SAIF deposits are
deposits insured by the SAIF but held by BIF-member banks. There were approximately $154
billion in Oakar bank SAIF deposits as of March 31, 1994. Sasser deposits are insured by the
SAIF but belong to banks that previously had been thrift institutions.
4
Table 2
Application of SAIF-Insured Institution Assessments
Dollars in Millions
Calender
Year
Cash
Assessment
Revenue
HCO
REFCORP
1989
$ 394
$295
$0
99
$ 394
$0
1990
1,828
738
1,090
0
1,828
0
1991
1,883
757
(29)
1,155
1,883
0
1992
1,777
772
0
740
1,512
265
1993
1,690
779
0
0
779
911
1994
1,729
596
0
0
596
1,133
Total
9,302
3,937
1,061
1,994
6,993
2,309
Assessments Diverted to:
FRF
Total
Assessments
Diverted
Cash
Assessments
Retained by
SAIF
O ther Sources
There are several potential sources of funds for the SAIF apart from assessment revenue.5
First, a total of up to $8 billion is authorized to cover insurance and losses under RTCCA for
Fiscal Years 1994 through 1998. Funds may be appropriated for covering incurred losses only,
subject to the following certifications to the Congress by the FDIC Board ("certification funds"):
(1) SAIF members are unable to pay additional assessments at rates required to cover
losses OR meet a repayment schedule for Treasury borrowings without adversely
affecting the ability of SAIF members to raise and maintain capital or to maintain
the assessment base; AND
5Although these provisions were replaced by those of RTCCA, 1) as revenue supplements,
if needed to supplement net assessment revenue to reach $2 billion annually for each of the
Fiscal Years 1993 through 2000; and 2) as net worth supplements, to pay for any amounts that
may be necessary to ensure that the SAIF meets the statutory specified minimum net worth for
each of the Fiscal Years 1992 through 2000. No funds were ever appropriated under the
FDRREA authorization.
5
(2) An increase in assessment rates needed to cover losses OR repay Treasury borrowings
could reasonably be expected to result in greater losses to the Government.
One problem with this language is the ambiguity concerning the definition of "unable to
pay." Moreover, the two-pronged nature of the test makes it difficult for the FDIC Board to
make the certification except under extreme conditions.
Any unexpended RTC funds before the end of the two-year period beginning at the date of
the RTC’s termination may be provided to the SAIF to cover losses, provided that the above
certifications to the Congress are made by the Board.
FIRREA also authorizes the SAIF to obtain working capital by borrowing funds from
Federal Financing Bank (FFB). For loss funds, the SAIF may borrow from the Treasury
part of the FDIC’s $30 billion line of credit). Additionally, the SAIF may borrow from
Federal Home Loan Banks. Finally, FIRREA allows for discretionary payments to be
the SAIF by the RTC.6
the
(as
the
to
Outlook
There is tremendous uncertainty regarding the future caseload of SAIF. If the economy
falters, it is possible that the SAIF would need to borrow from the $30 billion credit line if it
is responsible for resolving a large number of thrifts. Alternatively, the need to borrow could
be avoided if certification funds were appropriated upon meeting the required certifications.
The attached diagrams illustrate the sources and uses of SAIF funds between 1991 and June
30, 1994.
These sources are shown in the diagram illustrating sources of funds for the SAIF for 1993,
but not earlier, because these sources have not been accessed yet.
SAIF: Sources of Funds
1991
All dollar amounts In millions.
* Net of $20 million provision for losses.
SAIF: Sources of Funds
Assessments From
SAIF Members
1992
Assessments From
Conversion Banks
$172
$772
$845
$0
All dollar amounts in millions.
* Estimated year-end fund balance.
$0
SAIF: Sources of Funds
Additional Sources:
• Borrowings from FFB, Treasury,
and FHL Banks
• RTC discretionary payments
All dollar amounts in millions
SAIF: Sources of Funds
A dditional Sources:
• Borrowings from FFB, Treasury,
and FHL Banks
• RTC discretionary payments
All dollar amounts in millions
vo
A t t a c h me n t C
Aptd^sis g©£ Issues
Confronting tfrei:«.
Savings A ssociation
Insuranbe Fund
FDI€
Federal D eposit Insurance Corporation
Division of Research and Statistics
March 1995
Executive Summary
In its recent proposals on deposit insurance assessment rates, the FDIC Board of
Directors (Board) proposed to lower assessment rates for all but the riskiest BIF-insured
institutions upon recapitalization of the BIF. However, as the SAIF is much farther away from
capitalizing, the Board proposed to retain the existing assessment rates for the SAIF. If adopted
as proposed, there would be a rate differential between the average BIF assessment rate of 4.5
basis points and the average SAIF assessment rate of 24 basis points. This so-called "SAIF
differential" would be approximately 19.5 basis points. This premium differential arises for two
reasons discussed below—the need for SAIF members to build their fund to the designated
reserve ratio and the draw on SAIF revenues from assessments levied by the Financing
Corporation (FICO).
The SAIF currently is substantially undercapitalized; its year-end 1994 unaudited fund
balance of about $1.9 billion is $6.7 billion shy of die amount needed to achieve the designated
reserve ratio. The SAIF would have capitalized by year-end 1994 if assessment revenue had not
been diverted for other purposes. These diversions began with the inception of the SAIF in 1989
and totaled $7 billion through 1994: $3.9 billion for the Financing Corporation (FICO), $2
billion for the FSLIC Resolution Fund and $1.1 billion for the Refinancing Corporation. As a
result of this history, SAIF resources are inadequate to handle the failure of a large thrift or
several medium-sized thrifts. The longer the undercapitalization is allowed to persist, the greater
the chance that unanticipated losses will prevent the SAIF from meeting its target. This is a
particular concern because the analysis shows that while under a relatively optimistic baseline
assumption the SAIF capitalizes in 2002, this date is extremely sensitive to assumptions about
the volume of assets in failing thrifts.
The FICO assessment is currendy the primary obstacle to capitalizing the SAIF as well
as the primary source of the premium differential. The FICO assessment, which pays interest
on 30-year FICO bonds issued between 1987 and 1989, amounts to approximately $780 million
per year, or 45 percent of current SAIF assessment revenue. The FICO has a first claim on
SAIF-member assessments that will continue until the year 2019. The premium disparity arising
from the FICO assessment thus will last for 24 years and currendy amounts to 11 basis points
paid by SAIF members; this figure is likely to increase given the probable shrinkage of the SAIF
assessment base. The SAIF assessments that are available to FICO, however, are limited by law
to those assessments paid by institutions that are both SAIF members and savings associations.
Two types of institutions, so-called "Oakar" and "Sasser" institutions, do not meet both criteria.
As a result, FICO payments depend on revenues raised from approximately 67 percent of the
SAIF assessment base.
There are two potential effects of the premium disparity that are of concern. First and
most immediate is the potential for a substantial shrinkage or change in composition of the SAIF
assessment base that could imperil the ability of the FICO to service its obligations. This ran
occur in two ways. One is through Oakar acquisitions or Sasser conversions, in which case the
deposits stay in the SAIF but are not available for FICO payments. The second way is for
deposits to migrate from the SAIF to the BIF. This can come about as thrifts lose deposits to
bank competitors who pass on the differential to customers or through legal, regulatory, or other
maneuvering by thrift holding companies that attempt to migrate deposits into new or existing
banking subsidiaries. Even assuming minimal shrinkage of two percent per year in the FICOavailable assessment base and a moderate increase in Oakar acquisitions, FICO interest payments
cannot be serviced at current assessment rates by the year 2005. Rapid shrinkage of 10 percent
per year creates a FICO problem within two years. Such a scenario is not unrealistic in light
of recent announcements by thrift institutions attempting to establish new banking charters, and
the existence of other methods of transferring SAIF deposits to the BIF that do not require
regulatory approval.
The second concern is that the premium disparity could adversely affect the health of the
thrift industry and could result in increased losses to the SAIF. A premium differential could
adversely impact SAIF-insured institutions by increasing the cost of remaining competitive with
BIF-member institutions. Of particular concern to the FDIC is the impact a differential could
have on weaker SAIF-insured institutions and on failure rates for these institutions. An analysis
using a thrift model based on 1994 performance shows that under a variety of interest-rate and
asset-quality assumptions a premium differential of 20 basis points appears unlikely to increase
failures beyond a level manageable by the SAIF. The analysis shows that the possible effects
of rising interest rates and/or deteriorating asset quality may have greater effects on failure rates
- and therefore pose greater risks to the SAIF - than would a differential. Such potential
effects have led the FDIC to express concern about the undercapitalization of the SAIF since its
creation.
An Analysis of the Issues Confronting the
Sayings Association Insurance Fund
I. The Problems Facing the SAIF
The SAIF Is Undercapitalized
The FDIC Board of Directors (Board) recently issued for public comment separate
proposals on assessment rates for the Bank Insurance Fund (BIF) and the Savings Association
Insurance Fund (SAIF). The BIF is rapidly approaching recapitalization; the reserve ratio of
the BIF to estimated insured deposits is expected to reach the statutory m inim um Designated
Reserve Ratio (DRR) of 1.25 percent between May 1 and July 31, 1995. As of December 31,
1994, the BIF had a fund balance of $21.8 billion (unaudited) and an estimated reserve ratio of
1.15 percent. Upon recapitalization, the fund balance is expected to be almost $25 billion. The
BIF has reached this goal much more rapidly than originally projected; as a result, an average
BIF assessment rate of 23 basis points, or 23 cents for every $100 of insured deposits, will no
longer be required by law.1 The law requires that BIF assessment rates be set to m aintain the
DRR after that ratio has been achieved. There is currently no factual basis for raising the DRR
above 1.25 percent because at present there is no indication of significant risk of substantial
future losses to the fund. Accordingly, in order to m aintain the DRR at the statutory target of
1.25 percent, the Board proposed to lower assessment rates for all but the riskiest BIF-insured
institutions, while maintaining a risk-based assessment rate structure.2
However, the SAIF is much farther from achieving the DRR of 1.25 percent of estimated
insured deposits mandated by Congress and is not expected to become fully capitalized until the
year 2002. As of year-end 1994, the fund balance stood at $1.9 billion (unaudited), while the
target is approximately $8.7 billion. Thus, the SAIF currently re m ains undercapitalized. It has
been widely recognized for some time that this is the fundamental problem facing the SAIF.3
1The legal requirement for a weighted average assessment of 23 basis points will become
operative if the reserve ratio remains below the DRR for at least a year.
2In addition to a new assessment rate schedule, the Board proposed to widen the rate spread
of the current risk-based assessment rate structure applicable to BIF-insured institutions. The
assessment rate for institutions in the best risk classification would be reduced from 23 to 4 basis
points; the weakest institutions would continue to pay 31 basis points. The resulting rate spread
from best-rated to weakest would be 27 basis points. The average assessment rate under the
proposed schedule would be 4.5 basis points. Assessment rates for all nine risk categories are
shown in the proposed BIF assessment rate schedule (Attachment 1). See Federal Register 60
(February 16, 1995): 9270-79.
3This issue has been recognized by the FDIC since the creation of the SAIF. It was raised
on January 10, 1992, in a letter from William Taylor, Chairman of the FDIC, to Richard
Damian, Director, U.S. Office of Management and Budget, and it was raised again in a letter,
2
Beginning July 1, 1995, the SAIF will assume responsibility for resolution of failures of SAIF
members from the Resolution Trust Corporation (RTC). As the insurer, the FDIC, in particular,
is concerned about the ability of the SAIF to handle a large failure or several mid-sized failures
without additional capitalization.
The Board has the authority to reduce SAIF assessment rates to 18 basis points, or 18
cents for every $100 of insured deposits, until January 1, 1998, after which the average rate
must remain at 23 basis points or higher until the SAIF is capitalized. However, reduction of
the average rate to 18 basis points is projected to delay capitalization of the SAIF by three years,
until 2005. Moreover, if assessment rates were lowered to 18 basis points as allowed, it is
projected that available SAIF assessment revenues would not be sufficient to cover fully the
interest payment on FICO bonds as early as 1996.4 Given that the SAIF remains
undercapitalized and that the SAIF soon will begin resolving failures of SAIF members, the
Board chose to retain the existing assessment rates for the SAIF. The existing SAIF assessment
rate schedule yields an average assessment rate of 24 basis points, or 24 cents for every $100
of insured deposits.3 The details of the FDIC’s projections for SAIF capitalization are discussed
in the following section of this report.
Why the SAIF Is U ndercapitalized
The SAIF is behind in meeting its target because for the first three years of its existence,
1989 to 1992, SAIF-member assessment revenue did not flow to the SAIF; instead it was used
to pay for Federal Savings and Loan Insurance Corporation (FSLIC) losses incurred before the
enactment of the Financial Institutions Reform, Recovery, and Enforcement Act of 1989
(FIRREA). From 1989 through 1992, approximately 95 percent of total SAIF assessment
revenue was diverted to the FSLIC Resolution Fund (FRF), the Resolution F u n ding Corporation
dated February 20, 1992, from Stanley J. Poling, Director, FDIC Division of Accounting and
Corporate Services, to Jerome H. Powell, Assistant Secretary for Domestic Finance, U.S.
Treasury. More recently, the issue was addressed in a letter dated September 23, 1993, from
Andrew C. Hove, Jr., FDIC Acting Chairman, to the House and Senate Banking Committee
Chairmen and Ranking Minority Members. (See Attachment 2.) See, for example, the
Testimony of Andrew C. Hove, Jr., Acting Chairman of the FDIC, on "The Condition of the
Banking and Thrift Industries," before the United States Senate Committee on Banking, Housing
and Urban Affairs, September 22, 1994.
4FICO bonds and FICO’s assessment authority on SAIF assessment revenues are discussed
in the following section.
^ e proposed SAIF assessment rate schedule is shown in Attachment 1. See Federal
Register. 60 (February 16, 1995): 9266-70.
3
(REFCORP) and the Financing Corporation (FICO).6 As detailed in Figures 1 and 2, these
diversions totaled $7 billion through 1994: $3.9 billion for the Financing Corporation (FICO),
$2 billion for the FSLIC Resolution Fund and $1.1 billion for the Refinancing Corporation.
Without these diversions, the SAIF would have capitalized last year. Importantly, a significant
portion of SAIF assessment revenue continues to be diverted to pay the interest on bonds issued
by FICO from 1987 to 1989, referred to as FICO bonds.
The FICO assessment on SAIF members that are savings associations, referred to as the
FICO assessment, is the major current obstacle to the capitalization of the SAIF. Interest on
FICO bonds of approximately $780 million per year is paid from SAIF assessments. FICO
bonds are scheduled to mature between the years 2017 and 2019. This FICO assessment
effectively amounts to a tax on the thrift industry. FICO has the first draw on current SAIF
assessment revenue, draining revenue that otherwise would belong to the fund and contribute to
SAIF’s capitalization. The FICO draw currently represents approximately 45 percent of SAIF
assessment revenue, or 11 basis points out of the average assessment rate of 24 basis points.
In the absence of the FICO assessment going forward, the SAIF could capitalize in 1998, four
years earlier than currently projected.
6The remaining 5 percent consists primarily of assessment revenue from BIF-member banks
that owned SAIF-insured deposits. Until July 1, 1995, the SAIF’s total resolution responsibility
is limited to the SAIF-insured portion of these BIF-member institutions
4
FIGURE 1
No Premiums W ent To SAIF Before 1992
% of SAIF
Premiums
100
SAIF
Premiums
Flowing into
SAIF
Diversions
of SAIF
Premiums
1989
1990
1991
1992
1993
1994
1995
(P rojected)
Dollars in Millions)
diversion of S A IF
A ssessm en ts, Total
$394
$1,828
$1,883
$1,512
$779
FICO
295
738
757
772
779
FRF
99
0
1,155
740
0
0
0
0
1,090
(29)
0
0
0
0
0
0
0
265
911
1,133
903
$394
$1,828
$1,883
$1,777
$1,690
$ 1 ,729
$ 1,682
REFCO RP
Sash A sse ssm e n ts
:lowing into S A IF
otal S A IF
A ssessm ent
Revenue
$596
596*
$779
779
*The 1994 FIC O payment reflects a one-tim e $185 million refund of excess cash by FICO .
5
FIGURE 2
SAIF Is Undercapitalized
$8.9 Billion
$8.7 Billion
Assessments
Diverted From
SAIF, 1989-94
($7.0 Billion)
SAIF
(Fully
Capitalized)
SAIF $1
* Unaudited balance, net of reserves, 12/31/94.
6
The Ability of the SAIF to Fund FICO
The SAIF assessments that are available to FICO are limited by law to those assessments
paid by institutions that are both SAIF members and savings associations.7 Two types of
institutions, so-called "Oakar" and ’’Sasser" institutions, do not meet both criteria. Oakar
institutions, which are created from the purchase of SAIF-insured deposits by a BIF member,
pay assessments to both the BIF and the SAIF based on the proportion of BIF- and SAIF-insured
deposits held by the institution at the time of purchase. They are BIF members. Oakar
institutions held 25.2 percent of the SAIF assessment base as of year-end 1994. Sasser
institutions are SAIF members that have switched charter type and primary federal supervisor
without changing insurance fund membership; that is, they are either commercial banks (stateor federally chartered) or FDIC-supervised state savings banks. They are not savings
associations. Sasser deposits as of year-end 1994 comprised 7.4 percent of the SAIF assessment
base.
Since 1989, Oakar and Sasser institutions have increased their combined share of the
SAIF assessment base to approximately 33 percent as of the fourth quarter of 1994. FICO
payments depend on revenues raised from the remaining 67 percent of the assessment base. If
the Oakar and Sasser portion of the SAIF assessment base continues to increase, it will become
increasingly difficult to make FICO interest payments from current SAIF assessment revenues.
A legislative change to make Oakar and Sasser assessment revenue available to FICO would
reduce the likelihood of a near-term FICO shortfall, but would not address the fundamental
implications of the drain from the SAIF represented by the FICO draw on SAIF assessments.
In the absence of further movement of the SAIF deposit base into Oakar and Sasser
institutions, the ability of the SAIF to fund FICO will be affected by continued overall shrinkage
of SAIF deposits. The issues relating to such shrinkage of deposits are discussed below.
The SAIF Differential
One important effect of the FICO assessment is to exacerbate any differential that may
exist between BIF and SAIF assessment rates. A "SAIF differential,” that is, a difference
between the average BIF assessment rate and the average SAIF assessment rate, will be created
whenever the BIF recapitalizes and BIF assessment rates are lowered. The FDIC’s proposed
7A 1992 FDIC legal opinion determined that FICO assessments can be made only on savings
associations that are SAIF members. This opinion was described as "reasonable" by the
Comptroller General in a letter to the FDIC Board of Directors, dated May 11, 1992 and
recently reconfirmed by the FDIC. See Federal Register 60 (February 6, 1995): 7055-58.
7
change in BIF rates, if ultimately adopted by the Board, would create a SAIF differential of
approximately 19.5 basis points (24 basis points minus 4.5 basis points).8
The presence of a SAIF differential likely would create an incentive for SAIF members
to avoid assessments. However, there is currently a moratorium on fund conversions t h a t
generally prohibits institutions from converting their fund membership from the SAIF to the BIF.
The moratorium on conversions will continue until the SAIF reaches the DRR of 1.25 percent.
At that time, a SAIF differential would create an incentive for SAIF members to convert, thus
further reducing the SAIF assessment base. Nonetheless, conversions from the SAIF to the BIF
will not be costless: SAIF members will be required to pay an exit fee to the SAIF and an
entrance fee to the BIF.9 SAIF members choosing to convert also will face costs related to the
tax treatment of their cumulative loss reserve deductions. These costs would limit the extent to
which conversions from the SAIF to the BIF will occur after the SAIF has capitalized, absent
alternatives for shifting deposits from the SAIF to the BIF.
As part of their efforts to minimize the impact of a differential, thrifts could reduce
premium costs by shrinking their assessable deposits. Nonassessable liabilities, such as Federal
Home Loan Bank advances, could be substituted for assessable deposits, or funding needs could
be reduced through securitization. Because the FICO assessment is a fixed annual amount,
further shrinkage in the SAIF assessment base could increase the FICO "tax" from the current
11 basis points, which would create an additional incentive to reduce the use of SAIF deposits.
The Great ^Western ..Proposal. A SAIF differential also creates an incentive to migrate
deposits from the SAIF to the BIF. For example, deposit migration between SAIF- and BIFmember institutions within a holding company structure could occur. On March 1, 1995, Great
Western Financial Corporation, the parent company of a SAIF-member federal savings hank
with offices in California and Florida, announced that it had submitted applications for two
national bank charters. These commercial banks would share Great Western’s existing branch
locations. Presumably, with higher deposit interest rates being offered by the BIF subsidiary,
customers would be enticed to move their deposits from the SAIF subsidiary to the BIF
subsidiary, and these transfers would not be subject to exit and entrance fees. By mid-March,
five other SAIF-insured institutions had indicated that they are considering similar actions. If
these efforts are successful, certainly others will follow, and there is a potential for dramatic
shrinkage in the SAIF assessment base. These first six institutions have about $80 billion in
SAIF deposits, or nearly 12 percent of the SAIF assessment base. Removal of these deposits
*An analysis of the impact of a SAIF differential on troubled SAIF-insured institutions is
presented in Section III of this report.
9The SAIF exit fee is 90 basis points applied to the amount of insured deposits that are
transferred from the SAIF to the BIF. The BIF entrance fee is the BIF reserve ratio applied
against the amount of insured deposits transferred.
8
from the SAIF would result in a significantly smaller base from which to generate the fixed
FICO assessment.
Such a large shift in deposits would also have ramifications for the BIF. An additional
$80 billion in BIF-insured deposits would require an additional $1 billion in BIF reserves -1 .2 5
percent of $80 billion. While these announcements are unlikely to result in a large enough shift
in insured deposits from the SAIF to the BIF by midyear to delay recapitalization of the BIF,
such a shift could ultimately push the reserve ratio below 1.25 percent. If this were to occur,
premiums paid by banks would have to be increased in order to again reach and maintain the
1.25 target ratio. The six new BIF members would begin contributing assessments to the BIF,
but other BIF members would pay the preponderance of the needed $1 billion addition to
reserves. It is estimated that many more thrift institutions are considering ways of shifting
deposits to the BIF.
While the announced proposals require various approvals associated with chartering new
institutions, there are other means to achieve the same ends that do not require such approvals,
and are likely to lead to a further shrinkage in the SAIF assessment base. For example, existing
affiliations between BIF and SAIF members enable deposit-shifting without the need for new
charters or approvals by regulators. Markets respond to cost differences; those who suggest that
regulators can prevent the movement of deposits out of the SAIF appear to underestimate the
market's ability to innovate around constraints. If the rate of shrinkage in the SAIF assessment
base increases to 4 percent per year as a result of all available techniques, then the ability of
SAIF to fund FICO is threatened as early as 2001. Rapid shrinkage of 10 percent per year
creates a FICO problem within two years. Such a scenario is not unrealistic in light of recent
announcements by thrift institutions attempting to establish new banking charters, and the
existence of other methods of transferring SAIF deposits to the BIF that do not require
regulatory approval.
Condusioiis
Lower BIF premiums are not the fundamental problem, and an overcapitalized BIF is not
the solution. If BIF premiums were not reduced until the SAIF reserve ratio reaches 1.25
percent of insured deposits, as mandated by the Congress, the BIF would grow under reasonable
assumptions regarding bank failures to approximately $70 billion, or 3.2 percent of insured
deposits and $45 billion more than the $25 billion the BIF is expected to have upon
recapitalization. Overcapitalization of the BIF does not facilitate the capitalization of the SAIF,
which is the fundamental issue.
The existence of a differential is likely to initiate actions by thrifts to lessen or even
eliminate its effects and also may cause the rate of failures to increase as the profitability of the
thrift industry declines. As subsequent analysis will show, however, the premium differential
by itself is not likely to cause a substantial increase in failures. Nevertheless, the SAIF remains
vulnerable to unanticipated increases in losses. As illustrated in Figure 3, if thrift failures rise
minimally to one-half the level that banks have experienced over the past twenty years, that is,
9
22 basis points or about $2 billion per year, the SAIF would capitalize by 2002. If thrift failure
rates are slightly more than double the rate experienced by banks over the past twenty years,
SAIF will not capitalize and the fund will become insolvent early next century.
It is difficult to anticipate how thrifts will react to the differential, but it is certain that
there is a potential for rapid shrinkage of the SAIF assessment base. This can come about in
two ways. One is through Oakar acquisitions or Sasser conversions, in which case the deposits
stay in the SAIF but are not available for FICO payments. The second way is for deposits to
migrate from the SAIF to the BIF. This can come about as thrifts lose deposits to bank
competitors that pass on the differential to customers or through defensive maneuvering by thrift
holding companies who attempt to migrate deposits into new or existing banking subsidiaries.
Under a baseline assumption incorporating minimal shrinkage of 2 percent per year in the FICOeligible SAIF deposits and a moderate increase in Oakar purchases, FICO interest payments
cannot be serviced at current assessment rates by the year 2005. Rapid shrinkage of 10 percent
per year creates a FICO problem within two years, a scenario that is not unrealistic in light of
recent announcements referred to above. Such scenarios are considered in Figure 4.
10
FIGURE 3
Higher Failure Rates Prevent SAIF Capitalization,
And Threaten SAIF's Solvency
S A IF R eserve Ratio (% )
11
FIGURE 4
More Rapid Shrinkage Of SAIF Deposits
Means An Earlier FICO Shortfall
$ Billions
12
n . FDIC "Baseline” Projections for the SAIF
The ability of the SAIF to capitalize and to meet the FICO assessment will be affected
by a variety of factors. The growth or shrinkage of thrift deposits, the number of thrift failures
and the dollar amount of failed assets going forward will affect the SAIF’s fund balance. Other
factors, such as the percentage of thrift industry deposits held by Oakar and Sasser institutions,
in light of statutory constraints on the use of those institutions’ assessments for FICO payments,
also will have an influence.
Assuming modest insurance losses, moderating growth of Oakar institutions, and a slight
decline in thrift deposits over the next few years, the FDIC’s "baseline" projection shows that
the SAIF is expected to capitalize by reaching the DRR of 1.25 percent of insured deposits in
the year 2002. This result is unchanged from previous projections made in September 1994 and
January 1995. Under these assumptions, it also is expected that there would be sufficient
assessment revenue to cover the FICO interest payment through the year 2004, but a shortfall
will occur in the year 2005.
It must be emphasized that these assumptions are for analytical purposes, and while the
projections cover a period of 20 years or more, their fundamental purpose is to support the
setting of assessment rates for a six-month period, in this case the second semiannual assessment
period of 1995. A significant variation in any one of the assumptions could substantially affect
the ability to fund FICO or capitalize the SAIF, or both. The sensitivity of these factors to
changing assumptions is discussed in Section IV. A discussion of the assumptions used in the
baseline projection follows:
•
Failed-institution assets for 1995 and 1996 are based on estimates made by the FDIC’s
interdivisional Bank and Thrift Failure Working Group10. In November 1994, the
Working Group estimated failed SAIF-insured institution assets at $3 billion for 1995 and
$2 billion for 1996. The 1995 estimate of S3 billion is based on the Division of
Supervision’s projected failure of specific institutions that could occur in the second half
of the year, when the SAIF assumes resolution responsibility from the RTC. Beyond
1996, the assumed failed-asset rate for SAIF will be 22 basis points, or about $2 billion
per year.11
In the FDIC’s projections, banks and thrifts were assumed to face similar longer-run loss
experience. The BIF’s historical average failed-asset rate from 1974 to 1994 was about
10The Working Group’s membership is comprised of representatives of the Divisions of
Research and Statistics, Supervision, Finance, and Resolutions.
nThe failed-asset rate is based on the total assets of SAIF members, adjusted for Oakar
deposits.
13
45 basis points. However, a lower failure rate than the recent historical experience of
the BIF was assumed because the thrift industry is relatively sound following the RTC’s
removal of failing institutions from the system, and the health and performance of the
remaining SAIF members has improved markedly. As of year-end 1994, 86 percent of
all SAIF-member institutions were in the best risk classification of the FDIC’s riskrelated premium matrix.
One of the purposes of the Federal Deposit Insurance Corporation Improvement Act of
1991 (FDICIA) was to minimize losses to the insurance funds. FDICIA increased
regulatory oversight and emphasized capital. Specifically, FDICIA requires the closing
of failing institutions prior to the full depletion of their capital, limits riskier activities
by institutions that are less than adequately capitalized, and establishes audit standards
and statutory time frames for examinations. The law also requires the implementation
of risk-related assessments, which have provided effective incentives for institutions to
achieve and maintain the highest capital and supervisory standards. In light of these
provisions, the high levels of thrift failures and insurance losses experienced over the
past decade are not an appropriate baseline for the industry’s future performance.
•
The nominal loss rate on failed-thrift assets will be 13 percent. The expected loss rate
rises to 15 percent when the present value of the interest cost over the life of a
receivership is included. This loss rate approximates the loss experience of the BIF since
1986.
•
The asset growth rate for SAIF members will be zero, based on the industry’s recent
experience reflected in Table 1, which shows a slowing in asset shrinkage as fewer
institutions are placed into RTC conservatorship. Since the beginning of 1993, the total
assets of those SAIF members not in conservatorship have been quite stable, even
increasing slightly in each of the last three quarters of 1994. During this period, SAIFmember failures declined to nine in 1993 and two in 1994.
14
Table 1
Total Assets of SAIF-M ember Institutions
($ Millions)
Not In Conservatorship
Year:
Qtr
Assets
4-Qtr
Change
94:4
772,342
94:3
In Conservatorship
Total
Assets
4-Qtr
Change
Assets
4-Qtr
Change
2.0%
1,993
-90.9%
774,335
-0.6%
764,121
0.6
3,574
-87.2
767,705
-2.5
94:2
756,385
-1.1
11,999
-62.9
768,384
-3.6
94:1
752,522
-2.4
19,744
-39.8
772,266
-4.0
93:4
757,358
-8.1
21,901
-41.3
779,259
-9.6
93:3
759,745
-9.0
28,010
-12.0
787,755
-9.2
93:2
764,429
-10.6
32,361
48.1
796,790
-9.1
93:1
771,236
-11.5
32,816
28.1
804,052
-10.4
92:4
824,266
-6.7
37,289
-15.5
861,555
-7.1
91:4
883,187
-11.8
44,150
-43.9
927,337
-14.2
90:4
1,001,804
-12.7
78,658
-14.3
1,080,462
-12.8
89:4
1,147,611
—
91,768
—
1,239,379
—
•
The SAIF assessment base will continue to shrink, at 2 percent per year. Deposit
shrinkage since 1989 is shown in Figure 5 and Table 2. Although the emergence of a
SAIF differential may encourage less reliance on SAIF-assessable liabilities, the higher
overall shrinkage rates of recent years have slowed d ram a tic ally , from around 7 percent
per year in the years 1990 through 1992 to 1.2 percent in 1994.
As can be seen in Figure 5, a significant portion of the shrinkage is attributable to the
decline in RTC conservatorships. Since 1989, the cumulative reduction in deposits from
the time when institutions were placed in conservatorship to when they were resolved
was $82 billion. Although some portion of these deposits were transferred to other
15
FIGURE 5
SAIF Domestic Deposits
March 31,1989 to December 31,1994
$ Billions
16
Table 2
SAIF Assessment Base: Domestic Deposits ($ Millions) and
Percentage Change from P rio r Year-End
Year
Oakars *
Sassers *
Conserva
torships **
1994
180,118
28.8%
52,848
21.4%
1,629
-90.9%
486,228
-5.6%
720,823
-1.2%
1993
139,795
80.6%
43,520
51.2%
17,913
-43.1%
528,211
-15.2%
729,429
-4.1%
1992
77,395
9.9%
28,788
139.5%
31,480
-15.4%
622,813
-11.1%
760,475
-7.3%
1991
70,434
107.3%
12,018
333.2%
37,202
-45.1%
700,574
-9.4%
820,228
-6.5%
1990
33,971
1,494.1%
2,774
NM%
67,767
-24.4%
773,151
-9.9.%
877,663
-7.6%
1989
2,131
0
89,687
858,457
950,275
Other **
SAIF
Total
* Not available for FICO assessment
** Available for FICO assessment
healthy SAIF-insured institutions, the shrinkage is characteristic of weakened and failed
institutions, and because the number of such institutions has been greatly reduced, related
shrinkage can be expected to slow. Other evidence indicates that shrinkage was more
prevalent at weaker thrifts during periods when some better-managed thrifts experienced
deposit growth.12
Brokered deposits were another factor in the shrinkage of SAIF deposits, falling from
$64 billion at the end of 1989 to $9.8 billion at year-end 1994. This decline is due in
part to continuing legislative and regulatory constraints placed on their use by insured
institutions.
Another factor accounting for SAIF deposit shrinkage was depositor flight from the
declining or low interest rates which prevailed from 1990 to the latter part of 1994, as
shown in Figure 6.
12Larry Cordell et. a l., Deposit Flows at SAIF- a n d BIF-Insured Institutions: D ecem ber 1988
to September 1992 (Washington, D .C .: Office of Thrift Supervision, January 1993).
17
Figure 6
Quarterly Average Interest Rates
3-Month Treasury Bills
8.00
7.00-i;
6.00 - 1
5.00 -fi
4.00
3.00
2.00
89:4
902
90:4
912
91:4
922
92:4
932
93:4
942
94:4
Source: Derived from monthly average rates in the Federal
Reserve’s H.15 Statistical Release.
In seeking higher returns, many customers of depository institutions moved their
investments out of depository institutions and into mutual finds. Figure 7 shows that
household ownership of mutual funds more than doubled after short-term interest rates
began falling in early 1990.
It is recognized that the proposals by Great Western and others discussed in Section I
pose a potential for substantially faster shrinkage of the SAIF assessment base.
However, because the proposals have not been acted upon, this potential shrinkage has
not been factored into the baseline projection but rather is discussed in the sensitivity
analysis in Section IV.
18
Figure 7
•
Oakar deposits will grow at 2 percent per year, the estimated growth rate for BIFmember deposits. The purchase rate for Oakar deposits, while still positive, will decline.
The purchase rate of Oakar deposits will be 4 percent and 2 percent for the years 1995
and 1996, respectively, and will decline to 1 percent per year beginning in 1997.
Under FDICIA, Oakar deposits are adjusted annually by the acquiring institution's
overall domestic deposit growth rate (net of acquisitions). BIF-member domestic deposits
grew more than 9 percent per year in 1985 and 1986, but since then the growth rate has
slowed considerably. Since 1990, these deposits have increased, on average, 0.6 percent
per year, including a 0.3 percent rise in 1994. This reflects a greater reliance on
foreign-office deposits and other nonassessable liabilities. However, BIF-member
domestic deposits increased 1.9 percent during the fourth quarter of 1994, and with the
proposed reduction in BIF assessment rates, BIF-insured deposits will become more
attractive. For these reasons, BIF-member deposits in these projections were assumed
to grow by 2 percent per year, which - according to FDICIA - becomes the growth rate
for their Oakar deposits.
As shown in Figure 5 and Table 2, Oakar deposits have grown rapidly in recent years,
in part because a significant portion of those deposits were acquired from failed
19
institutions through the RTC. However, as the RTC completes the clean-up of the thrift
industry, these opportunities have all but disappeared.
Another incentive that prompted banking companies to acquire SAIF deposits was the use
of failed or failing thrifts as entry vehicles to states otherwise closed to them. However,
with the enactment of the Riegle-Neal Interstate Banking and Branching Efficiency Act
of 1994, banking companies may have other options available to them.
A premium differential may make SAIF deposits less attractive for acquisition by BIF
members, although the likelihood of an eventual differential has been known and should
have been a consideration in purchase decisions.
In 1994, Oakar deposit growth for the year ($40 billion) represented 6.8 percent of the
pool of SAIF deposits available for Oakar acquisition, that is, non-Oakar deposits, at the
beginning of the year. For the reasons stated above, this purchase rate is expected to
slow, which is reflected in the baseline assumptions.
•
The average assessment rate will remain at 24 basis points until the SAIF is capitalized.
As mentioned previously, the Board has the option of temporarily lowering the SAIF
assessment rate to 18 basis points until January 1, 1998, but the need to capitalize the
SAIF as soon as possible was given priority in the Board's proposal.
20
m . The Impact of a SAIF Differential on Troubled Institutions
The SAIF Differential
In the second half of 1995, the average assessment rate paid by BIF-insured institutions
would fall to 4.5 cents per $100 dollars of domestic deposits under the Board's proposal, while
the rate for the SAIF will remain at an average of 24 cents. If assessment savings for BIF
members were to be passed on to depositors or borrowers, SAIF members would incur higher
costs to remain competitive in the pricing of deposits and loans. An analysis was undertaken
to estimate the impact of this differential on the failure rate of troubled SAIF-insured institutions
and the implications for the SAIF.
Summary of Methodology
The analysis was based on the 272 SAIF-insured institutions that had FDIC supervisory
ratings of 3, 4 or 5 as of December 31, 1994.13 Five-year projections were run under a variety
of interest-rate and asset-quality scenarios. The model was relatively simple, with a basic
premise that institutions would continue to perform as they did during 1994, with some
adjustments for high levels of problem assets and restructuring charges. A nnual net interest
income was reduced by an amount equal to the differential multiplied by an institution's SAIFassessable deposits. Differentials between 5 and 20 basis points (0.05 percent and 0.20 percent)
were tested. Below are the results of die tests for 5 and 20 basis points. (Analysis revealed that
the results for 10- and 15-basis point differentials were distributed proportionally between those
of 5- and 20-basis points.)
Summary of Results
Within the framework and assumptions of this model, it appears unlikely that a 20-basis
point differential by itself would cause failures to increase beyond a level manageable by the
SAIF, within the five-year period considered in this analysis. The incremental failures indicated
by die model were generally smaller institutions. Unfavorable economic conditions that
adversely impact interest-rate spreads and asset quality generally have a greater effect on failure
rates, according to our study, than does a 20-basis point premium differential.
The projections indicated that a 5-basis point differential would have a minimal effect on
failures, while a differential of 20 basis points would increase the number of failures and failedinstitution assets by as much as one-third, depending on the assumptions in a particular scenario.
Of all the scenarios considered, the highest amount of failed-institution assets attributable to the
differential was $4.7 billion (over five years). Although actual losses would vary from year to
^Supervisory ratings range from 1, for the strongest institutions, to 5, for the weakest
institutions. The group of 272 institutions included 22 BIF-member Oakar banks which held
some SAIF-insured deposits.
21
year, on average this equates to a loss to the SAIF of $140 million per year, based on a loss rate
of 15 percent of failed-institution assets.14 This level of potential loss would be manageable for
the SAIF provided that losses from other causes, such as adverse economic conditions, are not
unexpectedly high. Moreover, the model’s failure projections are probably at the high end of
the range of what would be likely to occur because the model included some pessimistic
assumptions on the earnings impact of the differential.15
Methodology and Assumptions
As stated, five-year projections were run for the 272 SAIF-insured institutions with FDIC
supervisory ratings of 3, 4 or 5. These institutions had total assets of $207 billion at the end
of 1994.
•
•
The model assumed institutions would continue to perform as they did in 1994, with the
following exceptions:
►
Institutions with above-average expected losses in their loan
portfolios that were not covered by existing reserves were forced
to set aside sufficient reserves over the first two years of the
projection to cover their "excess" loss. An institution’s expected
loss was estimated as the greater of (a) 15 percent of its past due
and nonaccruing loans (the industry’s recent loss experience on
these assets) or (b) the industry’s 1994 average loss of 0.50
percent of average loans.
►
A few institutions booked substantial restructuring charges in 1994.
It was assumed that these were one-time charges that would not
recur in subsequent periods.
►
Consistent with the model’s basic assumption of holding 1994
performance constant, the model assumed no asset growth and no
asset or liability repositioning.
The differential was defined as the cost that SAIF members would incur in order to
remain competitive if BIF members pass their assessment reductions on to depositors or
borrowers. This cost to SAIF members was based on their SAIF-assessable deposits and
14As noted earlier, the FDIC’s failed-asset recovery rate has been 85 percent since 1986.
15The model assumed BIF members would pass on their entire assessment savings to their
customers and that SAIF-insured institutions would absorb the entire cost of this competitive
disadvantage. The likelihood of these and other qualifying factors are discussed at the end of
Section III.
22
was included as an added interest expense or reduced interest income, both of which
result in reduced net interest income.
•
The income tax rate of 36 percent was based on thrift industry results for 1994.
Institutions with positive earnings for a given year paid income tax, and those losing
money did not.
•
For the purpose of the model, annual loan-loss provisions were assumed to equal annual
net charge-offs. For any given year, provisions may exceed loan losses, or vice versa,
but over the longer term (such as the five-year period used here), these fluctuations
would be expected to level out.
•
Institutions that paid dividends in 1994 continued to pay the same amount. Some
institutions that paid dividends despite being unprofitable were assumed to continue to
pay dividends to enable parent companies to service obligations such as preferred stock
and subordinated debt.
•
Institutions were considered to have failed when their tangible equity ratio fell to 2
percent or less, the regulatory standard for "critically undercapitalized."
The Scenarios
Scenarios were designed to test the effects of 5- and 20-basis point differentials under
stressful interest-rate and asset-quality conditions. The following sets of tables show (1) a
"baseline" scenario, which shows the effects of a differential with no interest-rate or asset-quality
factors, (2) a "moderate" scenario and (3) a "severe" scenario. Results also are shown for
differentials at three different levels: zero, 5 basis points and 20 basis points. The interest-rate
and asset-quality scenarios were run separately and then in combination.
Interest-Rate Scenarios
Rising short-term interest rates experienced in 1994 and early 1995 have reduced net
interest margins16 for many banks and thrifts. Historically, deposit rates have adjusted more
quickly to changes in market interest rates than have asset yields, so in a rising interest-rate
environment net interest margins can be expected to narrow. This analysis was based on the
performance of these thrifts for 1994, and because short-term interest rates have increased
further since then, additional deterioration in some institutions’ net interest margins can be
expected.
16The net interest margin is the ratio of interest income minus interest expense, as a
percentage of average earning assets.
23
The most recent period of prolonged rising interest rates occurred in the late 1980s, but
changes in thrifts’ net interest margins during that period may also have been affected by the
industry’s severe credit-quality problems and other turmoil attributable to the savings-and-loan
(S&L) crisis. With no comparable recent precedent, changes in net interest margins were
examined from the fourth quarter of 1993 to the fourth quarter of 1994. For SAIF-member
institutions, the weighted-average net interest margin fell 13 basis points during this period to
3.12 percent, but one out of every eight of these institutions incurred a decline of 50 basis points
or more.17 This decline is attributable, at least in part, to rising market interest rates. The
distribution of changes in net interest margins is shown in Table 3.
Table 3
Changes in SAIF-Member Institutions9 Net Interest M argins
Fourth Q uarter 1993 to Fourth Q uarter 1994
Gains or Losses
(basis points)
Percent of
Institutions
Up 50 or More
6.0
Up 25 to 49
10.1
Up 1 to 24
21.1
Unchanged
1.2
Down 1 to 24
28.4
Down 25 to 49
20.2
Down 50 or More
13.1
In order to test the impact of a SAIF differential in a more stressful interest-rate
environment, the effects of rising interest rates were incorporated as percentage decreases in the
net interest margin. For example, the average decline from 3.25 to 3.12 mentioned above is
about 4 percent. A 10 percent reduction in the margin equates to about 33 basis points and a
15 percent decline is about 50 basis points: In’the interest-rate cycle used in this model, it was
assumed that interest rates would climb for two years, the same length of time as the recent rate
decline, from 1990 to 1992 (see Figure 6). Net interest margins would worsen during this
17The average commercial bank net interest margin was 4.42 percent for the fourth
quarter of 1994. Bank margins, on average, are somewhat higher than those of thrifts, in part
because banks have larger proportions of lower-cost demand deposits and higher-yielding
commercial and industrial loans.
24
period and then begin to recover as interest rates stabilize or decline and asset repricing catches
up with increases in deposit costs. In the model, institutions’ net interest margins were reduced
from their 1994 levels by the percentages shown in Table 4.
Table 4
Interest-Rate Assumptions:
Percentage Change in Net Interest M argins
Year
Scenario
1
2
3
4
5
Baseline
0
0
0
0
0
Moderate
-5
-10
-5
0
0
Severe
-10
-15
-10
-5
0
The results for these scenarios are shown in Table 5. The table shows the number of
failures and failed-institution assets over five years attributable to the differential (the incremental
failures). Thus, in the "baseline" scenario, which included no interest-rate factors, a differential
of 5 basis points would cause no additional failures compared to a differential of zero, and a 20basis point differential would cause 11 additional failures compared to a differential of zero.
Under the "moderate" scenario, a differential of 5 basis points would cause six additional
failures compared to a scenario with "moderate" interest-rate assumptions and no differential.
Table 5
Interest-Rate Scenarios:
Increm ental Failures Caused by the SAIF D ifferential
(Assets in Millions)
5 Basis Points
20 Basis Points
Scenario
Number
Assets
Number
Assets
Baseline
0
0
11
$1,282
Moderate
6
$816
17
$3,811
Severe
4
$336
15
$3,071
25
Under the 20-basis point differential, there were fewer failures attributable to the
differential in the "severe" scenario (15) than in the "moderate" scenario (17). This is because
the "severe" interest-rate factors caused a greater proportion of the failures than did the
differential, when compared to the "moderate" scenario. This phenomenon also occurs in other
tables presented in this section.
The estimated loss per year to the SAIF can be estimated using the FDIC’s recovery rate
on failed-institution assets since 1986 of 85 percent. For example, failed assets of S3.8 billion
over five years (from the table above) represent an average of $762 million per year, and the
expected loss per year would be 15 percent of $762 million, or $114 million.
More detailed results are presented on the following page.
T able 6
Results of Interest—Rate Scenarios
Fo r Institutions Rated 3y 4 and 5
B a se C a se : No SA IF Differential
$ Millions
F a ile d In situ tio n s
Number
A ssets
SA IF Deposits
R em ain in g In stitu tio n s
Number
A ssets
Number L e s s than
Adequately Capitalized
B aselin e
M oderate
S e ve re
47
13,771
10,341
52
14,215
10,695
65
17,197
12,985
225
193,865
220
193,421
207
190,439
32
35
32
SA IF Differential of 5 B a sis Points
$ Millions
F a ile d In situ tio n s
Number
A ssets
SA IF D eposits
In c re a s e fro m N o D ifferen tial
Number of Failu res
A ssets
R e m ain in g In stitu tio n s
Number
A ssets
Number L e s s than
A dequately Capitalized
B aselin e
M oderate
S e vere
47
13,771
10,341
58
15,031
11,362
69
17,533
13,280
0
0
6
816
4
336
225
193,865
214
192,605
203
190,104
36
30
29
SA IF Differential of 20 B a sis Points
$ Millions
F a ile d In situ tio n s
Number
A ssets
S A IF D eposits
In c re a s e fro m N o D ifferen tial
Num ber of Failu res
A ssets
R em ain in g In stitu tio n s
Number
A ssets
Num ber L e s s than
A dequately Capitalized
B aselin e
M oderate
S e ve re
58
15,053
11,378
69
18,026
13,739
80
20,268
15,536
11
1,282
17
3,811
15
3,071
214
192,584
203
189,610
192
187,368
30
31
31
27
Asset-Quality Scenarios
In the model, deteriorating asset quality is characterized by rising loan losses. For 1994,
the thrift industry’s loan-loss rate was 0.50 percent of average loans. For recent full years, the
industry’s loan-loss rates were as follows:
Year
Loss Rate
1994
1993
1992
1991
1990
0.50 %
0.65
0.59
0.65
0.61
The industry’s condition at the end of 1994 showed substantial improvement over recent
years, and because of the reduction in problem loans, loan losses for the near term can be
expected to remain near their recent low level. The thrift industry’s noncurrent loans were 1.48
percent of total loans on December 31, 1994, down from 2.10 percent at the end of 1993 and
2.58 percent at year-end 1992.18 A variety of problems can contribute to asset-quality
deterioration, either individually or in combination. National or regional economic downturns
or poor credit-underwriting judgments would be contributors, but other possible factors include
fluctuations in interest rates, competition and changes in the regulatory environment. A
premium differential could contribute to asset-quality problems for SAIF-insured institutions if
they take on additional risk in attempting to increase asset yields to offset the cost of a
differential.
Table 7 shows the loan-loss rates used in the asset-quality scenarios. In the "moderate"
scenario, the loss rate returns to its highest level of recent years before recovering, while in the
"severe" scenario the loss rate rises steadily to 0.90 percent. While the thrift industry
experienced substantially higher loss rates in the mid- to late 1980s, it seems highly improbable
that the industry could deteriorate to that level within the five-year time horizon used for this
analysis given the industry’s current condition, the vast amount of problem assets removed by
the RTC and by the industry’s own clean-up effort, and the increased emphasis on capital levels
and prudential supervision. Currently, 86 percent of SAIF members are in the best risk category
for deposit insurance premiums.
18Noncurrent loans include loans past due 90 days or more and those in nonaccrual status.
28
Table 7
Asset-Quality Assumptions:
Loan-Loss Rates (Percent of Average Loans)
Year
Scenario
1
2
3
4
5
Baseline
0.50
0.50
0.50
0.50
0.50
Moderate
0.50
0.60
0.65
0.65
0.60
Severe
0.50
0.60
0.70
0.80
0.90
The summary results of the asset-quality scenarios are presented in Table 8.
Table 8
Asset-Quality Scenarios:
Increm ental Failures Caused by the SAIF D ifferential
(Assets in Millions)
5 Basis Points
20 Basis Points
Scenario
Number
Assets
Number
Assets
Baseline
0
0
11
$1,282
Moderate
1
$92
12
$2,021
Severe
3
$452
17
$2,816
As can be seen in this table and on the following page in greater detail in Table 9, the
asset-quality factors caused somewhat fewer failures when compared to the interest-rate factors
(see Tables 5 and 6). The premium differential had less of a marginal effect on failures in the
asset-quality scenarios than in the interest-rate scenarios.
T able 9
Results of Asset—Quality Scenarios
Fo r Institutions Rated 3, 4 and 5
B ase C a se : No SA IF Differential
$ Millions
F a ile d In situ tio n s
Number
A ssets
SA IF D eposits
R e m ain in g In stitu tio n s
Number
A ssets
Num ber L e ss than
Adequately Capitalized
Baseline
M oderate
Severe
47
13,771
10,341
50
14,114
10,626
54
14,595
11,002
225
193,865
222
193,522
218
193,041
32
35
34
SA IF Differential of 5 B a sis Points
$ Millions
F a ile d In situ tio n s
Number
A ssets
SA IF D eposits
In c re a s e from N o D ifferen tial
Num ber of Failu res
A ssets
R em ain in g In stitu tio n s
Number
A ssets
Num ber L e ss than
A dequately Capitalized
B aseline
M oderate
Severe
47
13,771
10,341
51
14,206
10,691
57
15,047
11,374
0
0
1
92
3
452
225
193,865
221
193,430
215
192.589
36
35
34
SA IF Differential of 20 B a sis Points
$ Millions
F a ile d In situ tio n s
Num ber
A ssets
SA IF Deposits
In c re a s e from N o D ifferen tial
Num ber of Failu res
A ssets
R e m a in in g In stitu tio n s
Num ber
A ssets
Num ber L e s s than
Adequately Capitalized
B aseline
M oderate
Severe
58
15,053
11,378
62
16,135
12,173
71
17,411
13,252
11
1,282
12
2,021
17
2,816
214
192,584
210
191,501
201
190.225
30
33
31
30
Combination Scenarios
The interest-rate and asset-quality scenarios were run independently in order to make the
effects easier to interpret. However, in a higher interest-rate environment, credit quality is
likely to suffer eventually as lenders take additional risks in seeking higher returns to offset
shrinking net interest margins and borrowers encounter repayment difficulties.
These scenarios combined the "moderate” interest-rate parameters with the "moderate"
asset-quality parameters, and the "severe" interest-rate parameters with the "severe" asset-quality
parameters (see Tables 4 and 7). The summary results are shown in Table 10.
Table 10
Combination Scenarios:
Increm ental Failures Caused by SAIF Differential
(Assets in Millions)
5 Basis Points
20 Basis Points
Scenario
Number
Assets
Number
Assets
Baseline
0
0
11
$1,282
Moderate
4
$823
17
$4,661
Severe
3
$363
9
$1,770
As noted earlier, in some instances the differential had less of a marginal effect on
failures in the "severe" scenario than in the "moderate" scenario because the interest-rate and
asset-quality factors caused a greater proportion of the failures. Under the 20-basis point
differential in the table above, the "moderate" economic factors pushed 17 institutions (with
$4.66 billion in assets) to near-failure, and the addition of the differential caused them to fail.
The "severe" economic factors cause some of these 17 institutions to fail and left nine (with
assets of $1.77 billion) on the brink of failure that were caused to fail by the differential. Table
11 presents these results in greater detail.
T able 11
Results of Combination Scenarios
Fo r Institutions Rated 3, 4 and 5
B ase C a se : No SA IF Differential
$ Millions
F a ile d In situ tio n s
Number
A ssets
S A IF D eposits
R e m ain in g In stitu tio n s
Num ber
A ssets
Num ber L e s s than
Adequately Capitalized
B aselin e
M oderate
Severe
47
13,771
10,341
59
15,085
11,404
78
20,007
15,298
225
193,865
213
192,551
194
187.629
32
33
30
SA IF Differential of 5 B a sis Points
$ Millions
F a ile d In situ tio n s
Num ber
A ssets
SA IF D eposits
In c re a s e fro m N o D ifferen tial
Num ber of Failu res
A ssets
R em ain in g In stitu tio n s
Num ber
A ssets
Num ber L e s s than
Adequately Capitalized
B aselin e
M oderate
Severe
47
13,771
10,341
63
15,908
12,065
81
20,370
15,620
0
0
4
823
3
363
225
193,865
209
191,728
191
187.267
36
33
29
S A IF Differential of 20 B a sis Points
$M illions
F a ile d In situ tio n s
Num ber
A ssets
S A IF D eposits
In c re a s e fro m N o D ifferen tial
Num ber of Failu res
A ssets
R e m ain in g In stitu tio n s
Num ber
A sse ts
Num ber L e s s than
Adequately Capitalized
B aselin e
M oderate
Severe
58
15,053
11,378
76
19,746
15,218
87
21,777
16,676
11
1,282
17
4,661
9
1,770
214
192,584
196
187,890
185
185.859
30
28
34
32
Conclusions
This analysis indicates that failed-institution assets attributable to a premium differential
could range from zero to $4.7 billion over five years, depending on the effective size of the
differential and contributing economic factors. The higher failed-asset figure would amount to
an average annual loss to the SAIF of about $140 million attributable to the differential, but
losses of this magnitude should be manageable for the SAIF over the next five years, provided
there is no unexpected spiking of losses attributable to other factors.
The model’s interest-rate factors had more of an impact than the asset-quality factors, but
with the availability of hedging instruments, interest-rate fluctuations are likely to have fewer
adverse effects than they have had historically.
However, both interest rates and asset quality had a greater effect on failure rates than
did a premium differential, even at the 20-basis point level. Therefore, to the extent these
results are actually realized, it can be concluded that these economic factors pose greater risks
to the SAIF than does the differential.
Caveats with Respect to the Methodology and Assumptions
The model assumes BIF-insured institutions would pass on their entire assessment
reduction to depositors or borrowers. While some institutions may do this, others will pass
along some or none of their savings to depositors or borrowers, electing instead to enhance
shareholder value. Decisions on deposit pricing are based on funding needs, funding alternatives
and competition, while decisions on loan pricing are a function of risk, investment alternatives,
funding costs and competition.
The model assumes thrifts would absorb the entire cost of the differential. In reality,
they could lessen the impact by raising revenues, reducing other expenses or substituting
liabilities that are not SAIF-assessable, such as FHLB borrowings and reverse repurchase
agreements. Also, a number of the thrifts included in this analysis have been paying more than
the minimum assessment rate of 23 basis points.19 Therefore, since 1992 they have already
been operating with a differential of up to 8 basis points compared with many of their bank and
thrift competitors. Moreover, in earlier years -- 1984 to 1989 -- the premium differential
between banks and thrifts was about 12.5 basis points. Section V discusses historical
differentials in greater detail.
Also, the model does not allow for management actions that could result in turnarounds.
Institutions losing money in 1994 are projected to continue to do so, whereas in reality one
19Within the risk-related assessment rate matrix which has been in effect since 1992, rates
vary from 23 basis points to 31 basis points, based on an institution’s capital and supervisory
categories.
33
would expect to see portfolio restructurings, asset sales and recapitalizations, among other
things, in an effort to improve results.
Failed-asset figures are somewhat overstated to the extent they include the total assets of
failed BIF-member Oakar banks. Costs to resolve the assets of failed Oakars would be allocated
to the BIF or the SAIF based on the proportion of the institution’s deposits each fund insures.
In the scenario that resulted in the greatest amount of failed assets, about 2 percent of the total
would be resolved by the BIF, not by the SAIF.
Some parameters were determined by industry averages, but significant differences may
exist among institutions according to portfolio composition and institution size and location. For
example, average loss rates on multifamily residential real-estate loans (1.30 percent of average
loans) are greater than loss rates on l-to-4 family loans (0.25 percent), and the use of these more
detailed loss rates could yield somewhat different results than the average loss rate (0.50 percent)
used in the model.
The model was intended to focus attention on the incremental failures attributable to a
premium differential. The numbers and assets of projected total failures in Tables 6, 9 and 11
are probably less accurate in successive years because of the model’s relatively simple design
and limited focus. A comprehensive thrift performance model would take a more dynamic
approach to future performance. This approach would require m ak in g numerous assumptions
as to how the industry would react to the differential and to other regulatory, competitive and
economic factors.
34
IV. Sensitivity Analysis
Although the preceding analysis concludes that the SAIF differential by itself does not
create significant failures, the differential will create incentives for thrift institutions to shrink
their assessable base. Although the FDIC’s baseline projection calls for the SAIF to capitalize
in the year 2002, changes in the underlying assumptions could alter the projected date.
Similarly, the ability to fund FICO could be affected. This section examines the circumstances
under which problems for SAIF capitalization and the SAIF’s ability to support FICO
assessments could arise. In each case the current assessment-rate structure for the SAIF is
assumed to remain in place.
SAIF Capitalization. Factors including the growth or shrinkage of thrift deposits and the
assessment base, and the amount of failed assets going forward will affect the SAIF’s fund
balance. Of these, the primary factor affecting SAIF capitalization is the failed-asset rate, that
is, the amount of failed-thrift assets in a given year as a percent of total thrift assets. As
discussed in Section II above, the baseline failed-asset rate is assumed to be 22 basis points of
SAIF assets, or approximately $2 billion per year. This rate is reflective of the industry’s
current sound condition. Of interest to this analysis, then, is the extent to which SAIF
capitalization could be affected by alternative assumptions for the failed-asset rate.
Deposit or assessment-base shrinkage does not have a large impact on the year in which
the SAIF is expected to capitalize, as long as failed-asset rates are reasonably low. As
illustrated in Table 12, given the baseline assumption for failed assets of 22 basis points, the
projected SAIF capitalization in year 2002 generally is not affected by changes in the deposit
shrinkage rate. This primarily is due to the fact that changes in the base are "mirrored” in the
reserve ratio; increases or decreases in the base lead, respectively, to decreases or increases in
the ratio.20
Table 12 presents the results of an analysis in which the sensitivity of SAIF capitalization
to failed-asset rates and deposit-growth rates was examined. The year in which the SAIF was
projected to capitalize is shown under varying combinations of failed-asset rates and depositgrowth rates. The FDIC's baseline projection, discussed above in Section II, projected SAIF
capitalization in year 2002. This is denoted by superscript "a" in Table 12. The example noted
above can be found by comparing the projected capitalization dates when the failed-asset rate
is assumed to be 22 basis points. Even with a shrinkage rate of 15 percent, which could result
from the proposals by Great Western and others to migrate deposits from SAIF to BIF,
capitalization of the SAIF would actually occur one year earlier, in the year 2001, provided
20The reserve ratio is defined as the ratio of the SAIF fund balance to SAIF-insured deposits.
For a given fund balance, decreases in SAIF-insured deposits cause the SAIF reserve ratio to
increase. When deposit-shrinkage rates are sufficiently high, 10 percent to 20 percent in this
example, the reserve ratio increases lead to an earlier projected SAIF capitalization date.
35
failed-asset rates remain moderate. The impact of such a high rate of shrinkage on the ability
to fund FICO is discussed later.
Table 12
Sensitivity of SAIF Capitalization to
Failure Rates and Deposit Growth Rates
(SAIF Capitalization Dates)
Failed-Asset Rate
(Basis Points of SAIF Assets)
FICO-Eligible
SAIF DepositGrowth Rate
11
22
44
66
110
+ 2%
2001
2002
2005
(2004)
-2 %
2001
2002*
2007b
2010
*
(2001)
-4 %
2001
2002
2007e
*
(2000)
-6 %
2001
2002
2006
*
(2000)
-8 %
2001
2002
2006
*
(2000)
- 10 %
2000
2001
2005
*
(2000)
- 15 %
2000
2001
2004
*
(1999)
-2 0 %
1999
2000
2003
(2011)"
(1999)
* The SAIF does not capitalize by 2019.
Figures in parentheses represent the year of SAIF insolvency.
The following scenarios illustrate the sensitivity of the projected SAIF capitalization year
to alternate assumptions for the failed-asset rate and the deposit-growth rate. The first example,
denoted by superscript "b" in Table 12, combines the baseline assumption of 2 percent SAIF
deposit shrinkage21 with a failed-asset rate of 44 basis points of SAIF assets, or approximately
S4 billion per year. This rate approximates the BIF historical average failed-asset rate from
1974 to 1994 of 45 basis points. Under this higher failed-asset rate, SAIF capitalization would
be delayed until year 2007. A second example, denoted "c," shows that if the baseline
21The 2 percent deposit-shrinkage rate applies only to the non-Oakar or FICO-eligible portion
of the SAIF assessment base. The assumptions regarding Oakar deposit-growth and purchase
rates were discussed in Section II.
36
assumption of 2 percent deposit shrinkage is doubled to 4 percent, and a failed-asset rate of 44
basis points is assumed again, the expected SAIF capitalization date is unchanged at year 2007.
When the failed-asset rate is sufficiently high the SAIF may not be able to capitalize at
all. If the failed-asset rate is tripled to 66 basis points, or approximately $6 billion in failed
assets per year, which is about one and one-half times the BIF average failed-asset rate from
1974 to 1994, the SAIF generally does not capitalize by 2019. As denoted by "d," when
combined with a deposit-shrinkage rate of 20 percent, the SAIF becomes insolvent in 2011.
Under an even more pessimistic failed-asset rate of 110 basis points, the SAIF becomes insolvent
by the turn of the century.
The FICO Assessment. The primary factors that affect the SAIF’s ability to fund FICO
are the growth or shrinkage rates for FICO-eligible SAIF deposits and the percentage of the
SAIF assessment base that is held by Oakar and Sasser institutions. This analysis explores the
conditions under which FICO payment problems could arise.22 In particular, the analysis
examines the extent to which changes in these factors could affect the ability of the SAIF to fund
FICO.
Unlike the baseline projection discussed in Section II, this analysis is based on simplified
assumptions about the size of the FICO-eligible SAIF assessment base and the rate at which
FICO-eligible SAIF deposits shrink. While the baseline projection assumes moderate growth
in Oakar institutions going forward, this analysis holds the proportion of the assessment base
constant while the deposit-shrinkage rate is varied. The impact of alternate deposit-shrinkage
rates on the ability of the SAIF fund FICO under these simplified assumptions is shown in Table
13.
Currently, the percentage of the SAIF assessment base that is held by Oakar and Sasser
institutions is approximately 33 percent, leaving 67 percent of the SAIF assessment base
available for FICO payment purposes. In addition to the current FICO-eligible SAIF assessment
base of 67 percent, smaller FICO-eligible assessment bases of 60 and 50 percent are examined.
These reflect the growth of Oakar and Sasser institutions to 40 and 50 percent of the total SAIF
assessment base, respectively. For each of these FICO-eligible assessment bases, the depositshrinkage rate for FICO-eligible SAIF deposits is varied from 2 percent to 20 percent. The
following examples are illustrative of the results.
22The Board has the discretion to consider FICO’s debt-service needs in setting assessments
for SAIF members.
Table 13
Ability to Fund FICO from the
FICO-Eligible SAIF Assessment Base
(FICO Problem Dates)
FICO-Eligible SAIF Assessment Base
As a Percent of the Total SAIF Base*
FICO-Eligible SAIF
Deposit-Growth
Rate
67 %
60 %
50%
-2 %
2014*
2008e
\999t
-4 %
2004
2001
1997
-6 %
200lb
1999
1996*
-8 %
1999
1998
1996
- 10 %
1998c
1997
1995
-15 %
1997*
1996
1995
-2 0 %
1996
1996
1995
* This analysis holds the proportion of the assessment base constant while the
deposit-shrinkage rate is varied.
In the first example, denoted by superscript "a" in Table 13, the percentage of the SAIF
assessment base that is available for FICO payment purposes is m aintained at the current level
of 67 percent. If FICO-eligible SAIF deposits are assumed to shrink at a rate of 2 percent,
which, again, is the deposit-shrinkage assumption used in the FDIC’s baseline projection, full
FICO payments likely would be made only through the year 2013. In other words, a "FICOshortfair could occur in year 2014.
FICO problems will be encountered earlier if the deposit-shrinkage rate for FICO-eligible
SAIF deposits increases. For the next example, denoted by "b," assume that the percentage of
the SAIF assessment base available for FICO payment purposes remains at 67 percent. Assume
that FICO-eligible SAIF deposits shrink at a rate of 6 percent, a rate that is slightly h ig h e r than
the rate experienced, on average, since 1989 and is reflective of a period that included numerous
thrift failures. This combination would result in a FICO shortfall in year 2001; that is, full
FICO payments would be expected to be made only through year 2000. If a higher depositshrinkage rate of 10 percent is assumed, again keeping the FICO-eligible SAIF assessment base
at 67 percent, the increased rate would be expected to lead to a FICO shortfall in the year 1998.
This example is denoted by "c" in Table 13.
38
Without further shifting of SAIF deposits into Oakar and Sasser institutions, severe
shrinkage of FICO-eligible SAIF deposits, such as that suggested by the Great Western proposal,
would lead to an imminent FICO shortfall. Denoted by "d" in Table 13, severe depositshrinkage --15 percent per year —against the current FICO-eligible SAIF assessment base yields
an expected FICO shortfall in year 1997.
The ability of SAIF to fund FICO also will be affected if the percentage of the
assessment base held by Oakar and Sasser institutions continues to increase, thereby shrinking
the available FICO-eligible SAIF assessment base. These examples are denoted by superscripts
"e" and "f," respectively, in Table 13. First, given a deposit-shrinkage rate of 2 percent, a
decrease in the FICO-eligible SAIF assessment base from 67 percent to 60 percent leads to an
expected FICO problem in 2008. Next, a decline of the FICO-eligible SAIF assessment base
to 50 percent leads to an expected FICO problem in the year 1999.
In combination, changes in the deposit-shrinkage rate for FICO-eligible SAIF deposits
and the percentage of the SAIF assessment base available for FICO payments can be expected
to lead to the earlier onset of FICO problems. For example, as denoted by "g, 1 if FICO-eligible
SAIF deposits shrink at a rate of 6 percent, while the percentage of the SAIF assessment base
available for FICO payment purposes shrinks to 50 percent, the expected year in which FICO
payments cannot be made from available assessment revenue is 1996.
39
V. Competitive Issues
There is likely to be a negative impact on the competitiveness of SAIF-insured institutions
from a significant premium differential with BIF members. This effect is difficult to quantify.
It is probable that SAIF members will experience more difficulty raising capital in external
markets and increasing capital internally. However, as discussed below, there are other factors
that must be considered in evaluating the competitiveness of the industry.
R am in p s Impact of a Premium Differential. Twenty-five percent of SAIF members had
a return on assets (ROA) of 1.13 percent or higher for the year 1994. Under the rather
pessimistic assumption that pretax earnings are reduced by the full amount of the differential,
for this group of institutions with high ROAs, a premium differential of 20 basis points would
reduce pretax operating earnings by 6.8 percent. For institutions with ROAs at the median value
of 0.86 percent, the differential represents about 12 percent of pretax earnings. However, the
actual impact on earnings is likely to be less than these figures indicate because BIF members,
in aggregate, are likely to pass along less than the filli amount of their assessment savings to
customers, and the impact of any related cost increase for SAIF members can be mitigated to
the extent they can raise revenues or reduce other expenses.
Historical Evidence on Differentials. Savings associations historically have paid
somewhat higher deposit insurance premium rates than banks. From 1935 through 1980, the
effective premium rates (net of credits and other reductions) paid by savings associations were
4 to 5 basis points higher than bank rates. Since 1980, the average premium differential has
varied from zero (1992) to 12.5 basis points (1984 through 1989). Since 1992, when riskrelated assessment rate schedules went into effect for BIF and SAIF members, SAIF members
have paid, on average, 1 to 2 basis points more than BIF members. However, both banks and
thrift institutions in the highest rate category (31 basis points) have paid a differential of 8 basis
points as compared with their healthiest competitors.
Another form of differential relates to the different interest-rate ceilings that were applied
to banks and thrifts. Beginning in 1966, savings associations and savings banks were allowed
to pay higher interest rates on deposits than were commercial banks, creating a differential which
remained in effect until 1984. The interest-rate differential, which was as high as 100 basis
points but most frequently was set at 25 to 50 basis points, was intended to assure a flow of
funds to thrifts to finance the nation’s housing needs. This interest-rate differential was further
justified by the advantage commercial banks had by being able to accept demand deposits
(checking accounts) and engage in commercial lending. However, to the extent this advantage
existed, it was eroded during the 1970s and early 1980s by innovation, market forces and,
finally, legislation.
While it is important to note that there have been differences in the treatment of the two
industries historically, it is difficult to draw any conclusions based on this information regarding
the competitive effects of a premium differential over the next few years. First, the likely
magnitude of the future premium differential is larger than the premium differential that existed
40
in the past. Second, the effects of the differing price ceilings such as those in effect from 1966
through 1983 are conceptually different than the effect of differing tax rates that will result from
a premium differential. Finally, the economic, competitive and regulatory environment is much
different today.
Longer-Term Implications. The thrift industry also may face longer-term structural
problems. The industry may not be able to earn long-run competitive returns, in part, because
the business of mortgage lending has become more competitive. The growth of the secondary
mortgage market and government-sponsored enterprises such as the Federal National Mortgage
Association and the Federal Home Loan Mortgage Corporation have reduced the profitability
of holding mortgage loans to maturity. In addition, there are asset restrictions stemming from
the Qualified Thrift Lender test that must be met to realize the tax benefits available under a
thrift charter.23
^Under the Qualified Thrift Lender test, first enacted in FIRREA and subsequently
amended, savings associations must maintain 65 percent or more of their tangible assets in
"qualified thrift investments," which are predominantly loans and investments related to domestic
real estate. Failure to meet the test can result in, among other things, having to recapture the
bad debt reserve into taxable earnings.
41
VI. Conclusions
The SAIF began 1995 with a balance of $1.9 billion, barely one-fifth of its statutorily
required level. The primary current obstacle to capitalizing the SAIF is the FICO assessment.
If SAIF assessment revenues had not been diverted to FICO, the SAIF would have been
expected to capitalize in 1996; if other diversions of SAIF assessments totaling $7 billion to date
had not occurred, the fund would have capitalized in 1994. Similarly, if the FICO assessment
were removed from the SAIF today, the SAIF would be expected to capitalize in 1998. While
the thrift industry is in relatively healthy condition and failures projected for the near term
appear manageable, the fund remains vulnerable to a single large-institution failure or several
mid-sized failures that could result from adverse economic conditions or from management or
other problems affecting the asset quality or earnings of individual institutions. The SAIF has
little or no cushion for such adversities as it assumes responsibility for losses from failed
institutions beginning July 1, 1995.
A premium differential between BIF- and SAIF-insured institutions could create a
competitive disadvantage for SAIF members that would result in an increase in failures of SAIFinsured institutions. The fund should be able to absorb the expected losses such failures would
cause in the next five years, assuming other larger losses do not threaten the fund’s solvency.
Indeed, other factors - reduced net interest margins and asset-quality problems - could result
in a greater increase in failures of SAIF-insured institutions than are likely to result from the
proposed premium differential, according to our analysis.
Under certain baseline assumptions, the SAIF is projected to capitalize in 2002. The
capitalization date is sensitive to increases in failed-asset rates, from whatever cause. The
baseline projection also indicates that there would be sufficient assessment revenue to cover the
FICO interest payment through 2004, leaving a shortfall in subsequent years. However, this
date is sensitive to increases in the rate of assessment-base shrinkage or in the proportion of
Oakar or Sasser assessments. Hfforts by SAIF-insured institutions to lessen or avoid a premium
differential could significantly accelerate assessment-base shrinkage and hasten the date at which
there is a FICO shortfall.
The overall conclusion is that SAIF is assuming the full responsibility for resolving thrift
failures in a severely undercapitalized condition. Moreover, the impending premium differential
undoubtedly will spark sufficient entrepreneurial efforts to avoid the differential, thus all but
ensuring that FICO interest payments will not be met absent a significant and potentially
counterproductive increase in SAIF premium rates.
Attachment 1
Proposed Assessment Rate Schedules
Second Sem iannual 1995 A ssessm en t Period
FD IC-lnsured Institutions
Proposed BIF Rate*
Capital
Category
Supervisory R isk Group
Group A Group B Group C
1. Well
Adequate
3. Under
2.
4
7
21
7
14
28
14
28
31
Estimated Annual Assessment Revenue: $1.1 Billion
Average Annual Assessment Rate: 4.5 bp
Rate Spread: 27 bp
Proposed SAIF Rate*
Capital
Category
1. Well
2. Adequate
3. Under
Supervisory R isk Subgroup
Group A Group B G roupC
23
26
29
26
29
30
29
30
31
Estimated Annual Assessment Revenue: $1.7 Billion
Average Annual Assessment Rate: 24 bp
Rate Spread: 8 bp
Rates are in basis points
6-01 SUN 2U22
Attachm ent 2
®
FED ER A L D EP O SIT IN SU RA N CE CO RPO RA TIO N .
O F F IC E O F T H E C H A IR M A N
npton. oczocs
C c\ f?. C_L<
January 10, 1992
Honorable Richard D a m a n
Director
Office of Management and Budget
Washington, D.C. 20503
Dear Director Daman:
We have been advised by your staff that FDlC's appropriation
request for the Savings Association Insurance Fund vill not be
submitted in the President's budget for fiscal year 1993. our
submission included a request for $1,285 billion in Treasury
funding as provided for by FXKREA to bring the SAIF's revenue to
the designated level of $2 billion*
We strongly urge you to reconsider FDIC's appropriation
request for SAIF. The Congress and Administration outlined in
FIRREA an extensive funding plan from a combination of S&L
industry and taxpayer resources to put the SAIF on a sound
financial footing* According to legislation, modified slightly
by the recent Recap Bill, Treasury is to pay SAIF sufficient
funds to maintain its income at $2 billion during fiscal years
1993 through 2000. The FY93 Treasury payments for revenues, if
shipped, would not be available to the SAIF in the future*
The General Accounting Office in its 1990 financial audit of
the SAIF, expressed concern regarding the adequacy of funding
sources for SAIF to meet future resolution demands and achieve
net worth goals set by FIRREA. Clearly, not funding SAIF at this
point vill not allow the fund to build the resources, as
envisioned by Congress, to meet its future obligations.
Ftd e rA 'S « D o s1 t In su ran ce Corporation
W tsfctesan. i?C 204»
.
,
W fict cl the Dimeter
Dmtion e l Accounting tns Corpomt« Semiets
T ciru crj 20, 1992
Hr. Jerome H. Povell
Assistant Secretary, Domestic T inane«
Department of the Treasury
1500 Pennsylvania Avenue, K W
Washington, DC 20220
Pear Hr. Povell:
A s of December 31, 1991, the Savings Association Insurance Fund
(SAIF) is reporting a negative net vorth of $20,713,000 due to
losses recognised from the failure of a financial institution for
which the SAIF had a share of financial responsibility.
Section 11(a) (6) (F) of the Federal Deposit Insurance Act (12 U.6.C.
1S21 (a) (6) (F)) states that the S e c r e t a r y of the Treasury shall
p a y to the Savings Association Insurance Fund, for each fiscal year
... (beginning October 1 of 1991 through 1999), any additional
amount which nay be necessary, as determined by the Corporation and
the Secretary of the Treasury to ensure that such Fund has the
minimum net vorth referred to in [the table included as part of
this provision) throughout each ... fiscal year [noted above}....91
The minimum net vorth prescribed in this section of the TD1 Act for
the fiscal year beginning October 1, 1991 (l.e., fiscal year 1992}
i s aero.
Section 11(a) (6) (J) of the FDl'* Act (12 TJ.S.C. 1B21 (a)(€)(J)>
states that 91[t)here are authorized to be appropriated to the
Secretary of the Treasury such sums as nay be necessary to carry
out the provisions
of this
paragraph
(subject to certain
limitations) •. • •19
Accordingly,
ve reguest that t h e Treasury
acknowledge its
©hlioation to fund the aforementioned SAIF net vorth deficiency by
signing the concurrence line b e l o w and returning this letter to me
as soon as possible. Based on your concurrence, ve will establish
a receivable from the 0.8. T r e a s u r y for $20,713,000, which vill
bring the SAIF's fund balance t o zero as reguired b y lav.
Concurs
Jerom e H. Powell
Assistant
Secretary, Domestic Finance
F E D E R A L D E P O S IT IN S U R A N C E C O R P O R A T IO N . w«sn.ngton. oc 20429
O FFIC E O F TH E CHAIRMAN
September 23, 1993
Honorable Jim Leach
Ranking Minority Member
Committee on Banking, Finance
and Urban Affairs
House of Representatives
Washington, D.C. 20515
Dear Congressman Leach:
I would like to express my appreciation for your leadership with respect to the RTC/S AIF
funding legislation. As the House and Senate begin to resolve their differences in the legislation,
I would like to take this opportunity to raise a concern regarding the Savings Association
Insurance Fund. Although the current versions of RTC funding legislation are an improvement
over the status quo, both bills leave unresolved issues regarding the viability and the future of
the thrift industry and the SAIF.
The House adopted a provision providing the Resolution Trust Corporation with an
additional 18 months (until April 1, 1995) to resolve failing SAIF-insured institutions. I would
support the Conferees in adopting the 18-month extension.
Prior to the SAIF accepting failed institutions for resolution, it is my hope that the
Congress will examine what the viability and future is for the thrift industry and the SAIF. The
SAIF has three major obligations: to fund insurance losses associated with failures of SAIF
members, to recapitalize the insurance fund to an amount equal to 1.25 percent of insured
deposits, and to provide approximately $800 million per year of FICO bond interest payments
through the year 2019. In 1989, FIRREA authorized the Treasury, under certain conditions,
to provide appropriated funds to SAIF. These funds could have been used to meet these
obligations through the year 2000. Thus, while Congress envisioned a healthy, growing thrift
industry, FIRREA was crafted so that a backstop would be available in the event that
unfavorable industry conditions persisted.
The proposed legislation focuses on the obligation to fund insurance losses over the next
several years. This is accomplished in the House bill by extending the RTC’s responsibility for
failed institutions by eighteen months, thus providing an opportunity for the SAIF to capture the
net premium income during this time without incurring any insurance losses. In both the House
and Senate proposals, Treasury funds are available to SAIF only to cover losses subject to
certain certifications.
By focusing on insurance losses, the proposed legislation leaves recapitalization and the
FICO obligation as the responsibilities of SAIF members. While Treasury funding for
recapitalization was contemplated by FIRREA, Congress subsequently has determined it is more
appropriate to hold SAIF members responsible for recapitalizing their insurance fund. This
means that, with respect to recapitalization, SAIF members will be held to the same .standard
as BIF members. However, the FICO obligation creates a troubling disparity between BIF and
SAIF members. Given the current assessment base, FICO interest payments add 10 basis points
to SAIF premiums. Even if the insurance losses of the two funds are comparable in the future,
a differential premium rate will exist for most of the next 25 years as a result of the FICO
obligation.
The specter of a continuing premium differential creates a powerful incentive for SAIFinsured institutions to minimize premium costs by shrinking the base against which assessments
are levied (currently domestic deposits). This can be accomplished in a variety of ways even
if Congress enacts a moratorium on conversions of SAIF- to BIF-insured institutions and if the
definition of the assessment base is expanded to include other direct funding sources.
Furthermore, shrinkage may be hastened by the thrifts* awareness that their share of the FICO
burden will increase as the assessment base dwindles. The net result could be a dramatic
shrinkage of the assessment base, and therefore assessment revenue, that outpaces any increase
in premium rates. This could ultimately frustrate any attempt to recapitalize the SAIF and could
threaten the ability of the industry to fund FICO payments.
If you or your staff wish to discuss this further, please do not hesitate to contact me.
Acting Chairman
Oral Statement
Ricki Heifer
Chairman
Federal Deposit Insurance Corporation
on
The Conditions of the Bank and the Savings Association Insurance Funds
and Related Issures
Before the
Subcommittee on Financial Institutions
and Consumer Credit
Committee on Banking and Financial Services
U.S. House of Representatives
March 23, 1995
Madam Chairwoman and Members of the Subcommittee, I am here
today to present the views and analyses of the Federal Deposit
Insurance Corporation (FDIC) concerning the condition of the Bank
Insurance Fund (BIF) and the Savings Association Insurance Fund
(SAIF) .
We face a compelling problem -- and one that has grown more
compelling this year.
As my written statement discusses in
detail, the BIF is in good condition and its prospects appear
favorable.
In contrast -- despite the general good health of the thrift
industry -- the SAIF is not in good condition and its prospects
are not favorable.
Any solution to the SAIF problem requires action by the
Congress -- and, in fact, the need for Congressional action is
more urgent today than ever before.
Beginning later this year, Madam Chairwoman, a substantial
disparity between the deposit insurance premiums paid by BIF
members and SAIF members is likely to occur.
The disparity is
mandated by current statutory provisions.
The FDIC cannot avoid
bringing the disparity into being.
Only Congress can change the
laws that will soon require the FDIC to promulgate significantly
1
different assessments for the two deposit insurance funds.
Like
the tip of an iceberg, the premium disparity is only the visible
manifestation of a larger difficulty, most of which lies beneath
the surface.
This difficulty -- which most recently has been described in
depth in a report by the General Accounting Office -- has three
dimensions.
One, the SAIF is significantly underfunded. At year-end
1994, the SAIF had a balance of $1.9 billion -- or 28 cents in
reserves for every $100 in insured deposits.
This amounts to six
percent of the assets of SAIF-insured "problem” institutions.
The $21.8 billion BIF, in contrast, amounts to 52 percent of the
assets of BIF-insured problem institutions.
Two, an ongoing fixed draw of $779 million on SAIF revenue
arises from an obligation to pay interest on bonds issued by the
Financing Corporation (FICO) in the 1980s.
If you have ever
tried to fill a bucket with a hole in it, you understand what I
mean.
This draw alone creates a premium differential between BIF
members and SAIF members that likely will persist for 24 years
until the bonds are repaid.
This differential, at least 11 basis
points, could provoke further shrinkage in the SAIF assessment
base and a shortfall of assessment revenue to pay the FICO
obligation, which would lead to default on the bonds.
Although
FICO bonds are not obligations of the FDIC, interest on the bonds
is a significant drain on the SAIF.
Three, for the first time, the SAIF will assume
responsibility for resolving failed thrifts after June 30 of this
year.
Given the underfunding of the SAIF, significant insurance
losses in the near-term could render the SAIF insolvent and put
the taxpayer at risk.
To establish parity between the BIF and the SAIF today would
require about $15.1 billion, or about 25 percent of the total
equity capital of SAIF members.
Of this total, $6.7 billion
would be needed to increase the SAIF from its year-end 1994
balance of approximately $1.9 billion to $8.7 billion, the amount
that currently would achieve the designated reserve ratio
required by Congress of 1.25.
The remaining $8.4 billion of the
$15.1 billion is the amount that would be necessary at current
interest rates to defease the FICO obligation.
That is to say,
it is the amount that would have to be invested today to generate
an income stream sufficient to service the FICO bonds until
maturity between the years 2017 and 2019 because the bonds are
not callable.
Requiring these amounts to be collected entirely through
SAIF insurance premiums raises difficult questions.
What will be
the effect on the ability of SAIF members to raise new capital,
to prosper, and to compete effectively? Will erosion of the SAIF
assessment base and changes in its composition jeopardize the
a
ability of the FICO to meet its obligations?
burden be shared? And by whom?
Should some of the
There is no magic answer to these questions.
No matter how
the $15.1 billion cost is borne, there will be an outcry by at
least one constituency that a great injustice is being done.
There is no way for the FDIC to resolve this issue through the
exercise of its regulatory authority.
For two reasons the need to find solutions to the problems
grows more urgent.
One, as mentioned earlier, starting July 1,
1995, the cost of all new thrift failures must be paid out of the
SAIF.
Two, recently announced efforts by some SAIF-insured
institutions to transfer deposits into BIF-insured institutions
raises the specter that the insured deposit base of the SAIF
could shrink so rapidly that, under current assessment rates,
debt service on the FICO bonds would quickly run into trouble.
Six institutions have declared their intent to be " b o m again" as
BIF institutions.
Together, they total about $80 billion in
SAIF-insured deposits.
Although the need for immediate Congressional action
concerning the SAIF is evident, there is considerable
disagreement over precisely what action should be taken and
whether it should be taken this year or later.
The most
frequently mentioned sources of money to address SAIF's needs
include the thrift industry, the banking industry, and the U.S.
Treasury.
Others have been mentioned, too, as having an interest
in resolving the problems. None of the possible sources of
funding is happy about the prospect of footing the bill for
capitalizing the SAIF and funding the FICO interest payments.
As I noted earlier, the SAIF is significantly
undercapitalized -- it is constantly being drained to meet
obligations from savings-and-loan failures in the 1980s -- and it
must begin paying for thrift failures that occur after mid-1995.
I will discuss each of these three issues in turn.
First -- as chart number one shows -- the SAIF is
significantly undercapitalized.
As noted earlier, the SAIF had a
balance of $1.9 billion, or only 0.28 percent of insured deposits
at year-end 1994.
At the current pace, and under reasonably
optimistic assumptions, the SAIF would not reach the minimum
reserve ratio of 1.25 percent until at least the year 2002.
Consequently, it would be impossible to lower SAIF premiums to
the proposed levels for the BIF for at least seven years, and
because of the continuing need to fund interest payments on the
FICO bonds, probably much longer.
Second, SAIF assessments have been -- and continue to be -diverted to purposes other than the fund.
This problem was
described in detail in the recent General Accounting Office
3
report.
In short - - a s chart number two shows — from 1989 to
1994/ $7 billion -- approximately 95 percent of SAIF assessments
during that time -- was diverted from the SAIF to pay off
obligations from thrift failures in the 1980s through the
Resolution Funding Corporation (REFCORP), the Federal Savings and
Loan Insurance Resolution Fund (FRF), and the Financing
Corporation (FICO). Of the $9.3 billion in SAIF assessment
revenue received from 1989 to 1994, a total of $7 billion was
diverted: $1.1 billion was diverted to REFCORP; $2 billion was
diverted to FRF, and $3.9 billion was diverted to FICO.
By far
the largest of the drains on SAIF assessment income, the FICO was
established by Congress in 1987 in an attempt to recapitalize the
defunct Federal Savings and Loan Insurance Corporation.
From
1987 to 1989, the FICO issued approximately $8.1 billion in
bonds.
SAIF assessment revenue currently amounts to just over
$1.7 billion a year and FICO interest payments run $779 million a
year, or about 45 percent of all SAIF assessments.
Without these
diversions, the SAIF would have reached its designated reserve
ratio -- and would have been fully capitalized -- in 1994.
The
REFCORP and FRF no longer have claims on SAIF assessments, but -as things now stand -- the FICO claim will remain as an
impediment to SAIF funding for 24 years to co m e .
Third, the SAIF will be under stress beginning on July 1,
1995, when it takes over responsibility for resolving the
failures of SAIF-insured savings associations from the Resolution
Trust Corporation (RTC). One large or several sizable thrift
failures could bankrupt the fund.
The outlook for the SAIF is further complicated by the fact
that the law limits SAIF assessments that can be used for FICO
payments to assessments on insured institutions that are both
savings associations and SAIF members. As chart number three
shows, because assessment revenue from these institutions cannot
be used to meet debt service on FICO bonds, over 32 percent of
SAIF-insured deposits were unavailable to meet FICO payments in
1994.
This portion was up from 25 percent at the end of 1993.
This shift contributed significantly to a 7.9 percent decline in
1994 in the SAIF assessment base available to service FICO, even
though the overall insured deposit base of the SAIF declined by
only 1.1 percent in 1994.
At current assessment rates, an
assessment base of $325 billion is required to generate revenue
sufficient to service the FICO interest payments.
As chart number four shows, the FICO-available base at yearend 1994 stood at $486 billion.
The difference of $161 billion
can be thought of as a cushion which protects against a default
on the FICO bonds.
If there is minimal shrinkage in the FICO
assessment base -- 2 percent -- a FICO shortfall occurs in 2005.
4
Chart number five shows, however, that -- if shrinkage
increases -- for whatever reason -- the shortfall occurs earlier
-- as early as 1997 or even 1996 under some assumptions.
On March 1, 1995, Great Western Financial Corporation, the parent
company of a SAIF-member federal savings bank with offices in
California and Florida, announced that it had submitted
applications for two national bank charters.
Under the
applications these commercial banks would share Great Western's
existing branch locations.
As I noted before, by mid-March, five
other SAIF-insured institutions announced that they were
considering similar actions to shift deposits from the SAIF to
the BIF.
If these efforts in converting SAIF-insured deposits to BIFinsured deposits are successful, others are likely to follow.
These six institutions have approximately $80 billion in SAIF
deposits -- and that represents 50 percent of the FICO-cushion
mentioned earlier.
There are also other methods that do not
require applications or approvals to shift deposits from SAIF to
BIF.
For these reasons, the SAIF assessment base could shrink
significantly -- and quickly.
Removal of substantial deposits
from the SAIF would result in a significantly smaller base from
which to generate the fixed FICO assessment.
On Friday, March 17, the FDIC Board of Directors held an
unprecedented public hearing on the agency's proposals to reduce
deposit insurance premiums for most banks while keeping insurance
rates unchanged for savings associations.
Although written
comments-are not due until April 17, we have received almost 800
comment letters -- more than 100 in the 24 hours since we
completed our written testimony.
One message came through loud and clear from the majority of
the witnesses at the hearing:
In weighing proposals to address
the SAIF problem -- and many proposals have been made -- we must
seek a real and permanent solution, not one that simply defers
the issue to a later time while leaving in place the conditions
that are the source of the problem.
In that regard, any solution should be judged by how well it
accomplishes three goals.
First, it should reduce the premium disparity between BIF
and SAIF member institutions, and eliminate to the extent
possible the portion of the SAIF premium attributable to the FICO
assessments.
This disparity encourages SAIF members to engage in
legal and regulatory maneuvering to avoid SAIF assessments and in
my view renders infeasible the existing mechanism to fund the
FICO.
This standard leaves open the question of what level of
premium disparity between BIF and SAIF members would be small
enough to eliminate the incentive for SAIF members to flee the
SAIF.
5
Second, it should result in the SAIF being capitalized
relatively quickly, perhaps no later than 1998.
The longer we
allow the SAIF to be undercapitalized, the greater the
possibility that unanticipated losses will deplete the fund.
As
chart number six shows, under moderate failure assumptions, the
SAIF capitalizes in 2002.
Chart number seven, however, shows
that, if failures climb dramatically, they can prevent SAIF
capitalization altogether, and even threaten that insurance
fund's solvency.
Third, a solution should address the immediate problem that
on July 1, the SAIF will take over from the RTC the
responsibility of handling thrift failures.
Unfortunately, the
SAIF will assume this responsibility in a vulnerable and grossly
undercapitalized condition.
The progress towards capitalization, in other words, should
be "front-loaded," with a substantial chunk of the capital coming
quickly.
In addition, we need to be concerned about the means to
achieve these ends.
In that regard, we must consider the
precedent that is being set for the use of deposit insurance
funds.
To ensure sufficient insurance reserves to meet future
losses and to protect the FDIC's independence, deposit insurance
funds should be used for deposit insurance purposes.
Ideally,
the converse should also be true that deposit insurance expenses
should not be paid out of public funds, although the savings and
loan crisis is evidence of an unfortunate breach of the latter
principle, and the diversions from the SAIF for other purposes
prove the rule about the former.
We also must carefully consider
the fairness of the solution to all concerned.
Finally, to the
extent that Congress may wish to consider options involving the
use of RTC money to address the problems outlined here, there may
be budgetary issues outside the purview of the FDIC.
My written statement analyzes a number of options for
addressing these issues.
Madam Chairwoman, I take to heart Yogi Berra's observation
that "All predictions are dangerous, especially ones about the
future."
I do not try to foretell the future.
As a bank
regulator and a deposit insurer, however, it is a part of my job
to think about what could happen.
The resources of the SAIF are insufficient to absorb the
cost of the failure of one large or a few medium-sized thrifts,
or other substantial unanticipated losses.
If there are no major unanticipated losses, the SAIF balance
would inch up to its target over the next seven years.
Over this
6
length
losses
longer
chance
is too
of time,
will not
the time
the SAIF
thin.
however, it is difficult to take comfort that
prevent the SAIF from reaching its target.
The
before the SAIF capitalizes, the greater the
might fail to capitalize.
The margin of comfort
Therefore, there is a compelling need for legislative action
to reduce the disparity in the financial condition of the BIF and
the SAIF.
Again, I want to stress that any solution to the SAIF
problem should eliminate the long-term premium differential
caused by the FICO assessments. It should greatly reduce the
time needed to capitalize the SAIF.
It should include an
immediate injection of funds into the SAIF or a ready source of
backup funding for SAIF losses.
Madam Chairwoman, the FDIC is committed to finding solutions
that address these three concerns in a manner that is consistent
with good public policy.
We stand ready to assist the
Subcommittee in this effort in the weeks ahead.
I commend your
farsightedness in holding this hearing and I look forward to your
questions and to questions from members of the Subcommittee.
Thank you.
t
-t
3
7
@FedFRASER