[Federal Register: September 18, 2007 (Volume 72, Number 180)]
[Proposed Rules]
[Page 53181-53196]
From the Federal Register Online via GPO Access [wais.access.gpo.gov]
[DOCID:fr18se07-23]
=======================================================================
-----------------------------------------------------------------------
FEDERAL DEPOSIT INSURANCE CORPORATION
12 CFR Part 327
RIN 3064-AD19
Assessment Dividends
AGENCY: Federal Deposit Insurance Corporation (FDIC).
ACTION: Advance notice of proposed rulemaking (ANPR).
-----------------------------------------------------------------------
SUMMARY: The FDIC is seeking comments on alternative methods for
allocating dividends as part of a permanent final rule to implement the
dividend requirements of the Federal Deposit Insurance Reform Act of
2005 (Reform Act) and the Federal Deposit Insurance Reform Conforming
Amendments Act of 2005 (Amendments Act). The existing FDIC regulations
on assessment dividends will expire on December 31, 2008.
DATES: Comments must be submitted on or before November 19, 2007.
ADDRESSES: You may submit comments by any of the following methods:
Agency Web Site: http://www.fdic.gov/regulations/laws/federal.
Follow instructions for submitting comments on the Agency Web
Site.
E-mail: Comments@FDIC.gov. Include ``ANPR on Assessment
Dividends'' in the subject line of the message.
Mail: Robert E. Feldman, Executive Secretary, Attention:
Comments, Federal Deposit Insurance Corporation, 550 17th Street, NW.,
Washington, DC 20429.
Hand Delivery/Courier: Guard station at the rear of the
550 17th Street Building (located on F Street) on business days between
7 a.m. and 5 p.m. (EST).
Federal eRulemaking Portal: http://www.regulations.gov.
Follow the instructions for submitting comments.
Public Inspection: All comments received will be posted without
change to http://www.fdic.gov/regulations/laws/federal including any
personal information provided. Comments may be inspected and
photocopied in the FDIC Public Information Center, 3501 North Fairfax
Drive, Room E-1002, Arlington, VA 22226, between 9 a.m. and 5 p.m.
(EST) on business days. Paper copies of public comments may be ordered
from the Public Information Center by telephone at (877) 275-3342 or
(703) 562-2200.
FOR FURTHER INFORMATION CONTACT: Munsell W. St. Clair, Senior Policy
Analyst, Division of Insurance and Research, (202) 898-8967 or
mstclair@fdic.gov; Missy Craig, Senior Program Analyst, Division of
Insurance and Research, (202) 898-8724 or mcraig@fdic.gov; or Joseph A.
DiNuzzo, Counsel, Legal Division, (202) 898-7349 or jdinuzzo@fdic.gov.
SUPPLEMENTARY INFORMATION:
I. Background
In October 2006, the FDIC issued a temporary final rule to
implement the dividend requirements of the Reform
[[Page 53182]]
Act.\1\ At the time, the FDIC stated its intention to initiate a
second, more comprehensive notice-and-comment rulemaking on dividends
beginning with an advance notice of proposed rulemaking to explore
alternative methods for distributing future dividends after the
temporary dividend rules expire on December 31, 2008.
---------------------------------------------------------------------------
\1\ 71 FR 61385 (October 18, 2006).
---------------------------------------------------------------------------
The possibility of a dividend before the temporary rule expires
appears remote. In fact, because the FDIC has the ability to lower
assessment rates below the base assessment rate schedule (2 to 4 basis
points for institutions in Risk Category I), the FDIC can, if it
chooses, reduce the probability of a dividend occurring thereafter.
Reform Act Requirements
The Federal Deposit Insurance Act (FDI Act), as amended by the
Reform Act,\2\ requires that the FDIC, under most circumstances,
declare dividends from the Deposit Insurance Fund (DIF or fund) when
the reserve ratio at the end of a calendar year exceeds 1.35 percent,
but is no greater than 1.5 percent.\3\ In that event, the FDIC
generally must declare one-half of the amount in the DIF in excess of
the amount required to maintain the reserve ratio at 1.35 percent as
dividends to be paid to insured depository institutions. However, the
FDIC's Board of Directors (Board) may suspend or limit dividends to be
paid, if the Board determines in writing, after taking a number of
statutory factors into account, that:
---------------------------------------------------------------------------
\2\ The Reform Act was included as Title II, Subtitle B, of the
Deficit Reduction Act of 2005, Public Law 109-171, 120 Stat. 9,
which was signed into law by the President on February 8, 2006.
\3\ 12 U.S.C. 1817(e)(2).
---------------------------------------------------------------------------
1. The DIF faces a significant risk of losses over the next year;
and
2. It is likely that such losses will be sufficiently high as to
justify a finding by the Board that the reserve ratio should
temporarily be allowed to grow without requiring dividends when the
reserve ratio is between 1.35 and 1.5 percent or exceeds 1.5
percent.\4\
---------------------------------------------------------------------------
\4\ This provision would allow the FDIC's Board to suspend or
limit dividends in circumstances where the reserve ratio has
exceeded 1.5 percent, if the Board made a determination to continue
a suspension or limitation that it had imposed initially when the
reserve ratio was between 1.35 and 1.5 percent.
---------------------------------------------------------------------------
In addition, the statute requires that the FDIC, except in certain
limited circumstances, declare a dividend from the DIF when the reserve
ratio at the end of a calendar year exceeds 1.5 percent. In that event,
the FDIC generally must declare the amount in the DIF in excess of the
amount required to maintain the reserve ratio at 1.5 percent as
dividends to be paid to insured depository institutions.
The FDI Act directs the FDIC to consider each insured depository
institution's relative contribution to the DIF (or any predecessor
deposit insurance fund) when calculating an institution's share of any
dividend. More specifically, when allocating dividends, the Board must
consider:
1. The ratio of the assessment base of an insured depository
institution (including any predecessor) on December 31, 1996, to the
assessment base of all eligible insured depository institutions on that
date (the 1996 assessment base ratio);
2. The total amount of assessments paid on or after January 1,
1997, by an insured depository institution (including any predecessor)
to the DIF (and any predecessor fund);
3. That portion of assessments paid by an insured depository
institution (including any predecessor) that reflects higher levels of
risk assumed by the institution; and
4. Such other factors as the Board deems appropriate.
The statute does not define the term ``predecessor'' (of a
depository institution) for purposes of distributing dividends.
Predecessor deposit insurance funds are the Bank Insurance Fund (BIF)
and the Savings Association Insurance Fund (SAIF), as those were the
deposit insurance funds that existed after 1996 until their merger into
the DIF pursuant to the Reform Act. The merger was effective March 31,
2006.
Among other things, the statute expressly requires the FDIC to
prescribe by regulation the method for calculating, declaring, and
paying dividends.\5\ In May 2006 the FDIC issued a proposed rule to
implement the dividend requirements of the Reform Act.\6\ After
considering the comments received on the proposed rule, the FDIC, as
noted above, issued a temporary final rule on assessment dividends,
with a sunset date of December 31, 2008.
---------------------------------------------------------------------------
\5\ The dividend regulation must also include provisions
allowing a bank or thrift a reasonable opportunity to challenge
administratively the amount of dividends it is awarded. Any review
by the FDIC pursuant to these administrative procedures is final and
not subject to judicial review.
\6\ 71 FR 28804 (May 18, 2006).
---------------------------------------------------------------------------
The Temporary Final Rule
The temporary final rule mirrors the dividend provisions of the
Reform Act, provides definitions (including the definition of a
``predecessor'' depository institution) to implement the statute and
details how an institution may request the FDIC's Division of Finance
(DOF) to review the FDIC's determination of the institution's dividend
amount and how an institution may appeal DOF's response to that
request. In the temporary final rule, the FDIC adopted a simple system
for allocating any dividends that might be declared during the two-year
duration of the regulation. Any dividends awarded before January 1,
2009, will be distributed in proportion to an institution's 1996
assessment base ratio, as determined pursuant to the one-time
assessment credit rule.\7\
---------------------------------------------------------------------------
\7\ 12 CFR 327.53.
---------------------------------------------------------------------------
The sole focus of this ANPR is on the type of assessment dividend
allocation method that the FDIC should adopt. Whether and how the FDIC
should retain or revise the other aspects of the temporary final rule
(such as the timetable for determining and paying dividends and
institutions' requests for review) will be addressed in the notice of
proposed rulemaking that will follow the ANPR.
II. Alternative Methods
The ANPR presents two general approaches to allocating dividends--
the fund balance method and the payments method. These methods are
described below.\8\
---------------------------------------------------------------------------
\8\ Appendix A describes the two methods in more detail, using
formulas.
---------------------------------------------------------------------------
The allocation methods potentially differ most significantly in the
way they balance two of the statutory factors that the FDIC must
consider when allocating dividends--institutions' relative 1996
assessment bases and assessments paid after 1996--and, thus, in the way
each method treats older versus newer institutions. The fund balance
method implicitly balances the two factors; the payments method
requires explicit decision making.
``Older'' and ``Newer'' Institutions
In this context, the terms ``older'' and ``newer'' do not simply
refer to age. For purposes of this ANPR, the smaller an institution's
1996 assessment base is compared to its current assessment base, the
``newer'' it is. Thus, an institution that was chartered after 1996 and
had no 1996 assessment base is a newer institution. An institution
chartered before 1996 that has since grown greatly--and whose 1996
assessment base is, therefore, small compared to its current assessment
base--is also a newer institution. Conversely, the larger an
institution's 1996 assessment base is compared to its current
assessment base, the ``older'' it is.
[[Page 53183]]
Relative Dividend Shares
For purposes of analyzing the effects of each allocation method on
older and newer institutions, the notion of an institution's relative
dividend share is useful. An institution's relative dividend share at a
given time is the ratio of its share of any potential dividend to its
share of the current aggregate assessment base. A high relative
dividend share means that an institution would receive more than its
proportional share of a dividend given its current assessment base; a
low relative dividend share means that an institution would receive
less than its proportional share of a dividend given its current
assessment base.
The notion of a relative dividend share allows comparison of
dividend allocation methods by eliminating the effect of size. A newer
institution would initially have a zero or low relative dividend share,
whatever its size, while an older institution (as that term is used in
this ANPR) would initially have a high relative dividend share, again
regardless of size.
Some of the most important potential differences between the
dividend allocation methods are how quickly and under what
circumstances the relative dividend share of a newer institution would
equal the relative dividend share of an older institution. Equal shares
imply that what an institution paid prior to 1997 (using the 1996
assessment base as a proxy) no longer affects its dividend share. Under
most variations of the dividend allocation methods, the relative
dividend shares of older and newer institutions may never be exactly
equal, but they may become approximately equal; that is, over time, for
both older and newer institutions, shares of any potential dividend may
approximately equal shares of the current aggregate assessment base.
For purposes of the analysis in this ANPR, relative dividends shares
will be deemed to be approximately equal (or be said to have converged)
when the average relative dividend share of the group of institutions
that have the highest relative dividend shares as of January 1, 2007,
are no more than 15 percent greater (or less) than the average relative
dividend shares of newer institutions that initially have no dividend
shares.\9\ Under both allocation methods, the average relative dividend
share of the group of institutions that would have the highest relative
dividend shares as of January 1, 2007, would be 2.2; that is, in this
group, on average, an institution's share of any potential dividend
would be 2.2 times its share of the current assessment base.
---------------------------------------------------------------------------
\9\ This group is determined by dividing all institutions into 1
of 10 unequally sized groups, based on the size of their relative
dividend shares as of January 1, 2007. Because this date is the
beginning of the new risk-based assessment system, initial dividend
shares are proportional to shares of the 1996 assessment base.
---------------------------------------------------------------------------
The Fund Balance Method
Description
Under the fund balance method, every quarter, each institution
would be assigned a dollar portion of the fund balance (its fund
allocation), solely for purposes of determining the institution's
dividend share. Each institution's most recent fund allocation (as a
percentage of the fund balance) would determine its share of any
dividend. The fund allocation would increase or decrease each quarter
depending upon fund performance and assessments paid by each
institution. Specifically:
Initially, the December 31, 2006 fund balance would be
divided up among institutions in proportion to 1996 assessment bases.
Thus, initially, each institution's fund allocation would equal its
1996 ratio times the December 31, 2006 fund balance.
A variant on this method would divide only a portion of
the December 31, 2006 fund balance among institutions. The remainder of
the fund balance would be unallocated.
Thereafter, from quarter to quarter, fund allocations
would grow or shrink depending upon the performance of the fund.
Fund losses, FDIC operating expenses and dividends from
the fund would diminish an institution's fund allocation, all else
equal.
Fund gains (for example, from investment income or
``ineligible'' premium income, which is discussed immediately below)
would increase an institution's fund allocation, all else equal.
In addition, each ``eligible'' premium would increase an
institution's fund allocation, dollar for dollar. An ``eligible''
premium (which would need to be defined) would be the portion of an
institution's premium that would count toward increasing its share of
dividends.
Possible definitions for an eligible premium include: (1)
All premiums charged; (2) premiums charged up to the lowest rate
charged a Risk Category I institution; or (3) something in between, for
example, premiums charged up to the maximum rate for a Risk Category I
institution, in all cases minus any credit use.\10\ Ineligible premiums
would be those paid through the use of credits or those paid in cash at
rates in excess of the eligible premium rate.
---------------------------------------------------------------------------
\10\ However, an eligible premium would never be negative.
---------------------------------------------------------------------------
Eligible premiums would include surcharges in a
restoration plan.\11\
---------------------------------------------------------------------------
\11\ The Reform Act requires that the FDIC adopt a restoration
plan whenever the DIF reserve ratio is below 1.15 percent or is
expected to be below 1.15 percent within 6 months. The plan must
provide that the reserve ratio of the DIF will return to 1.15
percent, ordinarily within 5 years. 12 U.S.C. 1817(b)(3)(E).
---------------------------------------------------------------------------
Risk Reduction Incentives
As set forth above, when allocating dividends the FDIC is required
to take into account the portion of assessments paid by an insured
depository institution that reflects higher levels of risk assumed by
that institution. Consequently, in defining eligible premiums, an
important consideration (which applies to any approach) is the degree
to which dividend allocation should reinforce the risk incentives of
the risk-based premium system. Would an institution in the riskiest
category, for example, get credit for dividend purposes for the full
premium it paid or just for some smaller portion? If an eligible
premium were defined as a premium paid at the lowest (least-risky)
rate, an institution paying the highest assessment rate and an
institution paying the lowest assessment rate would increase their
dividend shares at the same rate, all else equal. Thus, the institution
paying the lower assessment rate on this base would benefit more,
thereby increasing the incentives for an institution to lower the risk
it poses. On the other hand, if the FDIC defined an eligible premium as
any cash premium, dividend awards, per se, would not provide an
institution with an incentive to reduce the risk it poses. If the FDIC
defined an eligible premium as something in between (for example, cash
premiums up to the maximum rate charged to an institution in Risk
Category I), the dividend system would give those institutions paying
higher rates than the eligible premium rate some incentive to lower
risk.
The Treatment of Older Versus Newer Institutions
Fund performance and assessment rates. Under the basic form of the
fund balance method, in which the entire fund would be allocated among
institutions, low to moderate fund losses would lead to older
institutions retaining a relatively large share of any dividends for
decades, while newer institutions would take decades to obtain a
relatively similar share of dividends. In other words, the assessments
paid by an institution prior
[[Page 53184]]
to 1997 (using the 1996 assessment base as a proxy) would affect an
institution's potential dividend for a very long time. On the other
hand, large fund losses would quickly diminish the relative shares of
older institutions compared to newer institutions.\12\
---------------------------------------------------------------------------
\12\ The results in the text, charts and tables that follow: (1)
Assume that the entire fund balance is allocated among institutions;
(2) assume that an eligible premium is a premium paid at the minimum
rate applicable to a Risk Category I institution; and (3) are based
upon a model that divides all institutions into 1 of 10 unequally
sized groups, based on the size of their relative dividend shares as
of January 1, 2007. The model assumes that all institutions grow at
the same rate. It makes many other assumptions, as well, including
levels of assessment rates, investment income, and corporate
expenses. These assumptions are set out in more detail in Appendix
B.
---------------------------------------------------------------------------
Chart 1 illustrates the relative dividend shares of two groups of
institutions--those that initially have no dividend shares (the newest
group) and those with the highest relative dividend shares (the oldest
group)--under a low loss scenario; Chart 2 illustrates the relative
dividend shares of these two groups under a high loss scenario similar
to the banking crisis of the late 1980s and early 1990s for the third
through tenth years, preceded and followed by low losses in earlier and
subsequent years. Assuming high fund losses similar to the banking
crisis of the late 1980s and early 1990s, the relative dividend share
of the newest group could take only 9 years to become approximately
equal to that of the oldest group (i.e., the relative dividend shares
of each group would be nearly equal to one).
[GRAPHIC] [TIFF OMITTED] TP18SE07.000
[[Page 53185]]
[GRAPHIC] [TIFF OMITTED] TP18SE07.001
Using the low loss scenario used in Chart 1, Table 1 compares
projected dividend share and dividends received for three institutions,
each with $500 million in deposits on December 31, 2006; one initially
has no dividend share (or credits) because it is new; one initially has
the median relative dividend share of those institutions that have any
initial dividend share (or credits); and one initially has a very large
relative dividend share because it is in the oldest group shown in the
charts above. Table 2 makes the comparison under the high loss scenario
used in Chart 2. The institutions are assumed to pay the lowest rate
applicable in any period. Like Charts 1 and 2, the dividend share
amounts in Tables 1 and 2 illustrate that older institutions will
benefit for many years from this method absent a repeat of the banking
crisis era.
The low loss scenario in Chart 1 and Table 1 (and in subsequent
charts in tables) assumes annual insurance losses that are
significantly lower than the average annual losses for the past 10
years and that the Board would not lower rates below the base
assessment rate schedule (2 to 4 basis points for institutions in Risk
Category I). In fact, if the Board did lower assessment rates
sufficiently below the base rate schedule, the dividends shown in Chart
1 would not occur.
BILLING CODE 6714-01-P
[[Page 53186]]
[GRAPHIC] [TIFF OMITTED] TP18SE07.002
[[Page 53187]]
All else equal, higher assessment rates (whether to cover rapid
insured deposit growth or from other causes) would shorten the time to
convergence of relative dividend shares of older and newer
institutions. However, the effect of higher rates would likely be less
marked than the effect of high fund losses similar to those during the
banking crisis of the late 1980s and early 1990s.
Institutions chartered in the future. Absent significant insurance
fund losses, the fund balance will tend to increase over time. Under
the fund balance method, all else equal, the larger the fund grows, the
longer it would take an institution chartered in the future to obtain a
share of potential dividends that was roughly equal to its share of the
assessment base; that is, for its relative dividend share to
approximately equal that of older institutions. Thus, an institution
chartered 30 years from now could take many decades to obtain a share
of potential dividends that was roughly equal to its share of the
assessment base.
Simplicity
The fund balance method relies on more data than the payments
method described below and is more complex, which may reduce
transparency. Both methods of fund allocation discussed in this ANPR
are operationally feasible, however.
Remaining Decision-Making Requirements
Both methods require the FDIC to define eligible premiums. Once the
definition of an eligible premium is chosen, however, the fund balance
method allocates dividends among older and newer institutions
automatically, without the need for explicit FDIC decision making about
the relative importance to assign the 1996 assessment base compared to
post-1996 eligible premiums.\13\ Only if the FDIC adopted the variant
of this method in which something less than the December 31, 2006 fund
balance was allocated among older institutions would it make explicit
decisions about how to allocate dividends between older and newer
institutions.
---------------------------------------------------------------------------
\13\ The FDIC's definition of an ``eligible'' premium would have
some effect on the way the fund balance method allocates dividends
between newer and older institutions, considered as a group. The
lower the eligible premium rate, the longer older institutions, as a
group, would retain a relatively larger share of dividends, all else
equal.
---------------------------------------------------------------------------
The Payments Method
Description
In its basic form, under most probable scenarios, the fund balance
method would most likely benefit older institutions. The payments
method, on the other hand, offers considerably more options for
allocating dividends between older and newer institutions. The payments
method could be constructed so as to benefit older institutions for
many years, or it could be constructed to accelerate convergence
between older and newer institutions.
Under the payments method, unlike the fund balance method, neither
fund performance nor dividends paid would affect dividend shares
directly. Rather than hinging on its assigned portion of the fund
balance, an institution's share of any dividend would depend upon its
(and its predecessors') 1996 assessment base (or, equivalently, its
1996 ratio), weighted in some manner, and its quarterly assessments
under the new assessment system. Specifically:
Initially, each institution's dividend share would depend
upon its 1996 assessment base compared to all other institutions. For
example, initially, each institution's dividend share could equal:
1. Its 1996 ratio times the fund balance on December 31, 2006;
2. Its 1996 ratio times the fund balance at some other time; or
3. Its 1996 ratio times insurance fund assessment income over some
period of time leading up to December 31, 1996, in each case as a
percentage of the total for all institutions.
The resulting value assigned to each institution based on
its 1996 ratio could either remain unchanged or be assigned a declining
weight over time.
The possible definitions of an eligible (and an
ineligible) premium are the same as those under the fund balance
method. However, under certain variations of this method discussed
below, assessments offset through credit use could increase an
institution's dividend share.
Cumulative eligible premiums paid into the fund since 1996
would add to an institution's share.
Alternatively, the FDIC could count only eligible premiums
paid over some recent period, for example, the most recent 3, 5, 10 or
15 years. In contrast, the fund balance method would necessarily take
into account all assessment payments made under the new assessment
system.
Another variation would allow the FDIC to subtract
dividends paid to an institution from its eligible premiums.
The Board would explicitly determine the relative importance to assign
to each institution's 1996 assessment base and to its eligible premiums
paid under the new system. The rate at which the relative importance of
eligible premiums paid under the new system increased (and the relative
importance of the 1996 assessment base decreased) could be slow or
fast. Alternatively, the FDIC could, at the outset of the system,
reserve the right to change the balance in the future.\14\
---------------------------------------------------------------------------
\14\ A simplified version of the payments method would
substitute assessment bases as proxies for eligible premiums. Each
institution's share of any dividend would depend on its portion of
the 1996 assessment base, weighted in some fashion, and its
cumulative quarterly assessment bases under the new system. In this
version, an institution would automatically have an added incentive
to be charged the lowest possible rate, since, given identical
assessment bases, an institution paying the lowest assessment rate
would increase its dividend share at the same rate as an institution
paying the highest assessment rate, all else equal.
---------------------------------------------------------------------------
Risk Reduction Incentives
As under the fund balance method, the degree to which dividend
allocation would reinforce the risk incentives of the risk-based
premium system would depend upon the FDIC's definition of an eligible
premium.
The Treatment of Older Versus Newer Institutions
Relative weight of the 1996 assessment base. The relative weight to
be accorded the 1996 assessment base could have a great influence on
how quickly the relative dividend shares of newer and older
institutions would converge.
How the payments method would affect the dividend shares of older
and newer institutions would depend on the weight that the Board
assigned the 1996 assessment base (initially and over time) compared to
the weight it assigned eligible premiums paid each year after 1996. Two
illustrative variations of the payments method are described below.
Variation 1. The Board could, as under the fund balance method,
initially divide the 2006 fund balance based on each institution's
share of the December 1996 assessment base. Eligible premiums after
1996 would be added to that amount. As illustrated in Chart 3 and Table
3, this method of implementation would result in older institutions
retaining relatively large dividend shares for many years--similar to
the fund balance method--given low losses. (Compare with Chart 1 and
Table 1.) \15\
---------------------------------------------------------------------------
\15\ The low loss scenario in Chart 3 and Table 3 again assumes
annual insurance losses that are significantly lower than the
average annual losses for the past 10 years and that the Board would
not lower rates below the base assessment rate schedule (2 to 4
basis points for institutions in Risk Category I). In fact, if the
Board did lower assessment rates below the base rate schedule, the
dividends shown in Chart 3 and Table 3 would not occur. See also
footnote 13.
---------------------------------------------------------------------------
[[Page 53188]]
[GRAPHIC] [TIFF OMITTED] TP18SE07.003
[[Page 53189]]
[GRAPHIC] [TIFF OMITTED] TP18SE07.004
[[Page 53190]]
Under the payments method--unlike the fund balance method--fund
gains and losses would not directly affect an institution's relative
dividend share. However, higher insurance fund losses could lead to
higher assessment rates, which would affect relative dividend shares.
All else equal, higher assessment rates (either resulting from fund
losses or rapid insured deposit growth) would tend to make the relative
dividend shares of older and newer institutions converge more quickly.
However, as illustrated in Chart 4 and Table 4, the effect of an
increase in higher assessment rates on relative dividend shares would
not be as large as the direct effect of large insurance losses under
the fund balance method. (Compare with Table 2 and Chart 2.) \16\
---------------------------------------------------------------------------
\16\ Chart 4 and Table 4 assume that an institution's dividend
share is initially determined by multiplying its 1996 ratio times
the fund balance at the end of 2006 and adding eligible premiums
over time. See also footnote 13.
[GRAPHIC] [TIFF OMITTED] TP18SE07.005
[[Page 53191]]
[GRAPHIC] [TIFF OMITTED] TP18SE07.006
[[Page 53192]]
Variation 2. Another way to implement the payments method would be
to consider only premiums paid over some prior period (such as the
previous 15 years). When the prior period covered any year before 2007,
the years 1997 through 2006 would be skipped, since the great majority
of institutions paid no deposit insurance premiums then. Thus, for
example, to determine dividend shares at the end of 2009, the method
would consider premiums paid from 1985 through 1996 and from 2007
through 2009. Premiums paid during 2007, 2008 and 2009 would include
only eligible premiums. However, because the weight accorded the 1996
ratio would effectively decline to zero over time, eligible premiums
after 2006 would include eligible premiums offset with credits. An
eligible premium paid in 1996 or any earlier year would be calculated
as an institution's share of the 1996 assessment base times total
deposit insurance fund assessment income in that year.\17\
---------------------------------------------------------------------------
\17\ For years prior to 1990, deposit insurance fund assessment
income used to produce Chart 5 and Table 5 includes such income for
both the FDIC and the Federal Savings and Loan Insurance
Corporation.
---------------------------------------------------------------------------
As illustrated in Chart 5 and Table 5, newer and older institutions
would have equal relative dividend shares after 15 years.\18\ \19\ \20\
---------------------------------------------------------------------------
\18\ The low loss scenario in Chart 5 and Table 5 again assumes
annual losses that are significantly lower than the average annual
losses for the past 10 years and that the Board would not lower
rates below the base assessment rate schedule (2 to 4 basis points
for institutions in Risk Category I). In fact, if the Board did
lower assessment rates below the base rate schedule, the dividends
shown in Chart 5 and Table 5 would not occur. See also footnote 13.
\19\ If eligible premiums did not include eligible premiums
offset with credits, newer institutions would actually have higher
relative dividend shares than older ones after 15 years (because
older institutions would use credits in early years, which would
reduce their eligible premiums). Thereafter, however, the dividend
shares of older and newer institutions would tend to converge again.
\20\ A high loss scenario would lead to a more rapid
convergence.
[GRAPHIC] [TIFF OMITTED] TP18SE07.007
[[Page 53193]]
[GRAPHIC] [TIFF OMITTED] TP18SE07.008
BILLING CODE 6714-01-C
[[Page 53194]]
The relative dividend shares of older and newer institutions would
converge similarly if an institution's dividend share were initially
determined by multiplying its 1996 ratio by the fund balance at the end
of 2006 and adding eligible premiums over time, where the weight
accorded the 1996 ratio diminished linearly and steadily to zero over
15 years (again allowing eligible premiums to include eligible premiums
offset with credits). However, institutions chartered in the future
would be at a greater disadvantage than if only recent payments (e.g.,
those made within the previous 15 years) were considered.
In general, the length of time it would take an institution
chartered in the future to obtain a share of potential dividends that
was roughly equal to its share of the assessment base would depend to a
great extent upon the relative weight to be accorded the 1996 ratio. If
the 1996 ratio (or 1996 assessment base) were heavily weighted and
payments accumulated indefinitely, it could take an institution
chartered in the future many years to obtain an equal share of
potential dividends. However, if the 1996 ratio received a small weight
and only very recent assessments (rather than cumulative payments) were
considered, it would take an institution chartered in the future only a
short time to obtain an equal share of potential dividends.
Simplicity
The payments method would require less data than the fund balance
method and would be relatively easy to administer. If the payments
method considered only recent payments (e.g., 3 or 5 years), data needs
and record retention requirements for the industry and the FDIC would
be particularly simple.\21\
---------------------------------------------------------------------------
\21\ The simplification of the method in which assessment bases
are used as a proxy for actual payments requires only that
institutions and the FDIC retain data on assessment bases.
---------------------------------------------------------------------------
Decision-making
Like the fund balance method, the payments method would require
that the FDIC define eligible premiums. Under the payments method the
FDIC would have considerably more options regarding the allocation of
dividends between older and newer institutions than it would under the
fund balance method. The FDIC would decide:
How much weight to accord the 1996 assessment base
compared to premiums paid under the new system;
Whether that weight should change over time and whether
the FDIC should reserve the right to change the weight in the future;
and
Whether all payments under the new system should be
considered or only more recent payments.
III. Request for Comments
The FDIC requests comment on all aspects of the fund balance method
and the payments method, and on any alternative approach not presented
in this ANPR that a commenter chooses to discuss. In particular, the
FDIC invites comment on the following:
1. Which method is preferable and why?
2. Is a method not presented in this ANPR preferable? If so, why?
3. Is there a variation or way of implementing any method that is
preferable or less preferable? If so, why?
4. How should an eligible premium be defined and why should it be
so defined?
5. If the payments method were selected:
(a) Are any of the two illustrative variations more or less
preferable?
(b) Should eligible premiums be considered only over some limited
prior period, such as 3, 5 or 10 years?
(c) Should premiums paid with credits count toward dividend share,
as described in the second illustrative variation?
(d) Should premiums paid over some very recent period (e.g., the
previous year) be excluded to avoid creating an incentive for
institutions to increase their assessment base and assessments in hope
of obtaining a larger dividend?
(e) Should dividends paid to an institution be subtracted from its
eligible premiums?
(f) How should the 1996 assessment base be taken into account or
weighted? How quickly should its relative importance decrease over
time? Should the FDIC reserve the right to change its relative
importance in the future?
6. Is any method particularly burdensome or not burdensome?
7. Any other aspects of either of the two methods or of a method
not presented in this ANPR.
Appendix A--Definition and Description of the Fund Balance Method
An institution's dividend share would equal the dollar portion
of the fund balance assigned to it (its fund allocation) as a
percent of the total adjusted fund balance. An institution's
dividend share would be defined recursively. Its initial dividend
share (DSi,0), on January 1, 2007, would be:
[GRAPHIC] [TIFF OMITTED] TP18SE07.009
where ai,0 is institution i's fund allocation on January
1, 2007, and F0 is the fund balance as of December 31,
2006.
For quarters ending after December 31, 2006, adjusted fund
balances are used. An adjusted fund balance differs from the actual
fund balance by excluding estimated premium income for the quarter.
Premiums earned for each quarter would be estimated because they
would not be determined for, and collected from, each institution
until the following quarter.
An institution's fund allocation at time 0 would be derived from
its share of the 1996 aggregate assessment base. Therefore, equation
(1) can be restated as:
[GRAPHIC] [TIFF OMITTED] TP18SE07.010
In the equation above, fi is the share of the 1996
aggregate base for institution i and is calculated as:
[[Page 53195]]
[GRAPHIC] [TIFF OMITTED] TP18SE07.011
where ab96i is 1996 assessment base for institution i and
j = 1 through N represents all institutions. Institutions that did
not exist on December 31, 1996 or are not successors to institutions
in existence then would have 1996 ratios set to zero.
An institution's dividend share for each succeeding quarter
(DSi,t) would be:
[GRAPHIC] [TIFF OMITTED] TP18SE07.012
where DSi,t is institution i's dividend share at time t,
t is the end of the most recent quarter for which the fund balance
is available, ai,t is institution i's fund allocation at
time t and Ft is the adjusted fund balance at time t.
Institution i's fund allocation at time t, ai,t, in
the equation (4) is derived as:
[GRAPHIC] [TIFF OMITTED] TP18SE07.013
where ht is an adjustment factor accounting for the
growth or shrinkage of the adjusted fund balance (as defined above)
from t-1 to t after excluding eligible premiums for the quarter
ending at time t-1 that were collected at time t, rt is a
redistribution factor that redistributes the shares of institutions
that failed after time t-1 but before time t and
pi,t is eligible premiums paid by institution
i at time t for the quarter ending at time t-1.
The adjustment factor for the growth or shrinkage of the
adjusted fund balance, ht, is calculated as:
[GRAPHIC] [TIFF OMITTED] TP18SE07.014
where mt is all institutions in existence at time t. The
redistribution factor, rt, is calculated as:\22\
---------------------------------------------------------------------------
\22\ However, an institution might fail after the end of the
quarter on which dividend shares are calculated (which will always
be the fourth quarter), but before distribution of a dividend.
Consequently, a final adjustment of dividend shares may be
necessary. This share would be calculated as follows:
See equation 8 above.
where DSi,B is institution i's dividend
share at the time a dividend is distributed, B is the time at which
a dividend is distributed, and mB is all institutions at
time t that had not failed as of time B.
[GRAPHIC] [TIFF OMITTED] TP18SE07.015
[GRAPHIC] [TIFF OMITTED] TP18SE07.016
Definition and Description of the Payments Method
An institution's dividend share, DSi,t, would be
defined as:
[[Page 53196]]
[GRAPHIC] [TIFF OMITTED] TP18SE07.017
where DSi,T is institution i's current dividend share, T
is the end of the most recent quarter for which assessment base data
is available, wT is the weight assigned to the 1996 ratio
for period T, ab96,i is the 1996 assessment base for
institution i, T-k is the earliest period to be covered, which could
be all periods after 2006 or some recent period, such as the most
recent 3, 5, 10 or 15 years, pi,t is eligible premiums
paid by institution i at time t for the quarter ending at time t-1,
and mT is total institutions as of time T.\23\, \24\
---------------------------------------------------------------------------
\23\ Under Variation 2 described in the text, T-k would not
include any year before 2007. When a dividend share in any year
depended upon premiums paid before 1997, the premiums would be
factored into wT rather than being included in
pi,t.
\24\ If an institution failed after the end of the quarter on
which dividend shares were calculated (which will always be the
fourth quarter), but before distribution of a dividend, a final
adjustment of dividend shares may be necessary. This share would be
calculated simply by deleting the failed institution's payments and
1996 ratio from the preceding formulas.
---------------------------------------------------------------------------
Appendix B--Model Assumptions
Among other things, the model assumes the following:
1. Investment income in 2007 equals 4.7 percent of the start-of-
year fund balance. For each year thereafter, it equals 4.57 percent
of that year's starting fund balance. These estimates are based on
projections from an investment model that relies on Blue Chip
forecasts of the yield curve through 3rd quarter 2008.
2. The initial assessment rate schedule is 3 basis points above
the base rate schedule; thus, the initial minimum rate is 5 basis
points. Rates fall to base rates the year after the fund reserve
ratio reaches or exceeds 1.25 percent. Risk Category I institutions
that pay rates between the minimum and maximum rate for the category
are assumed to pay 0.6 basis points above the minimum rate, which
reflects the current weighted average rate for the group.
3. Any restoration plan is assumed to be a 5 year plan.
Surcharges in a restoration plan are estimated using an iterative
procedure to account for the effect of credit use. During a
restoration plan, an institution may use no more than 3 basis points
in credit use.
4. Operating expenses for 2007 are $988 million and grow at an
annual rate of 5 percent thereafter.
5. Insured and domestic deposits are assumed to grow at 5
percent per year.
6. The beginning fund balance at 2007 equals $50,165 million.
7. Credit use is limited by the 90 percent rule during 2008,
2009, and 2010. (No institution may apply credits to offset more
than 90 percent of an assessment for these years.)
8. Institutions are assigned to 1 of 10 credit groups and 1 of 6
assessment rate groups based on December 31, 2006 Call Report and
TFR data, CAMELS information, and one-time credits. An institution's
credits are determined by its share of the December 31, 1996
assessment base. An institution's credit group is determined by the
ratio of its credits to its December 31, 2006 deposits. Because an
institution's initial relative dividend share is determined
analogously, based upon the ratio of its share of the December 31,
1996 assessment base to its share of the December 31, 2006 deposits,
institutions in the same credit group will have similar relative
dividend shares. In the tables and charts in the text comparing the
relative dividend shares under alternative allocation methods, the
``oldest'' group refers to the credit group with the most credits
relative to their December 31, 2006 deposits, those whose credits
are more than 12 basis points of their December 31, 2006 deposits.
The initial weighted average of credits-to-deposits for the credit
group is 15.6 basis points.
9. High fund losses correspond to the losses incurred by the
Bank Insurance Fund from 1987 to 1994, with losses measured relative
to total domestic deposits. Low fund losses assume losses are equal
to 0.1 basis points of domestic deposits each year.
Dated at Washington, DC, this 11th day of September, 2007.
By order of the Board of Directors.
Federal Deposit Insurance Corporation.
Robert E. Feldman,
Executive Secretary.
[FR Doc. 07-4596 Filed 9-17-07; 8:45 am]
BILLING CODE 6714-01-P