The Federal Employees' Retirement System: Potential Changes in Agency
Retirement Costs Following an Open Season (Testimony, 11/05/97,
GAO/T-GGD-98-27).
GAO discussed the potential impact on agency retirement costs of a
retirement system open season in which anyone who is a general civilian
employee of the federal government and is participating in the Civil
Service Retirement System (CSRS) or the CSRS Offset Plan would be
allowed to transfer to the Federal Employees Retirement System (FERS),
focusing on the potential changes in retirement costs charged to federal
civilian agencies that might follow such a transfer program.
GAO noted that: (1) it is difficult to predict who among eligible
employees would switch; (2) review of the first FERS transfer program in
1987 showed that although eligible employees were provided the
information and counseling that they would need to make a decision,
about 4 percent of the eligible employees transferred to FERS; (3) the
review suggests that employee decisions can be based on situational
factors that are economic as well as noneconomic; (4) assuming some
employees opt to switch, agency retirement costs would increase
following an open season because of differences in the way CSRS and FERS
are funded; (5) the amount of any increase would critically depend on
the number of employees who switch and their salary levels; (6) given
the uncertainty regarding how many employees might transfer, it is
correspondingly difficult to estimate whether agencies would have a
difficult time absorbing the cost increases; (7) although the largest
increase in retirement costs GAO calculated--$332 million--would appear
small in proportion to the costs of personnel benefits governmentwide,
some agencies might find the costs difficult to absorb, depending on
their different circumstances; and (8) regardless of the size of the
increases in costs, under the budget process discretionary spending is
capped, and Congress may choose not to provide agencies extra funding to
cover their increased retirement costs.
--------------------------- Indexing Terms -----------------------------
REPORTNUM: T-GGD-98-27
TITLE: The Federal Employees' Retirement System: Potential Changes
in Agency Retirement Costs Following an Open
Season
DATE: 11/05/97
SUBJECT: Fringe benefit costs
Federal employee retirement programs
Employee retirement plans
Civilian employees
Government retirement benefits
Economic analysis
Social security benefits
Cost analysis
IDENTIFIER: Federal Employees Retirement System
Civil Service Retirement System
Federal Thrift Savings Plan
Civil Service Retirement Offset System
******************************************************************
** This file contains an ASCII representation of the text of a **
** GAO report. Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved. Major **
** divisions and subdivisions of the text, such as Chapters, **
** Sections, and Appendixes, are identified by double and **
** single lines. The numbers on the right end of these lines **
** indicate the position of each of the subsections in the **
** document outline. These numbers do NOT correspond with the **
** page numbers of the printed product. **
** **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced. Tables are included, but **
** may not resemble those in the printed version. **
** **
** Please see the PDF (Portable Document Format) file, when **
** available, for a complete electronic file of the printed **
** document's contents. **
** **
** A printed copy of this report may be obtained from the GAO **
** Document Distribution Center. For further details, please **
** send an e-mail message to: **
** **
** **
** **
** with the message 'info' in the body. **
******************************************************************
Cover
================================================================ COVER
Before the Subcommittee on Civil Service Committee on Government
Reform and Oversight
House of Representatives
For Release on Delivery
Expected at
10:00 a.m., EST
Wednesday
November 5, 1997
THE FEDERAL EMPLOYEES' RETIREMENT
SYSTEM - POTENTIAL CHANGES IN
AGENCY RETIREMENT COSTS FOLLOWING
AN OPEN SEASON
Statement by Michael Brostek
Associate Director, Federal Management and
Workforce Issues
General Government Division
GAO/T-GGD-98-27
GAO/GGD-98-27T
(410225)
Abbreviations
=============================================================== ABBREV
CSRDF - Civil Service Retirement and Disability Fund
CSRS - Civil Service Retirement Systems
FERS - Federal Employees' Retirement System
GAO - General Accounting Office
OPM - Office of Personnel Management
TSP - Thrift Savings Plan
THE FEDERAL EMPLOYEES' RETIREMENT
SYSTEM: POTENTIAL CHANGES IN
AGENCY RETIREMENT COSTS FOLLOWING
AN OPEN SEASON
====================================================== Chapter SUMMARY
The Federal Employees' Retirement System (FERS) was implemented in
1987 and generally covers those employees who first entered federal
service after 1983. When FERS began, employees covered by the Civil
Service Retirement System (CSRS) were provided an opportunity to
transfer to FERS during a 6-month open season that ended in December
1987. The Federal Employees' Retirement Open Season Act of 1997,
passed by Congress but vetoed by the President, would have allowed
employees currently enrolled in CSRS or the CSRS Offset program a
second chance to transfer. Transferees under the act would have
received substantially the same coverage and benefits that were
provided to those who transferred during the first open season.
GAO was asked to examine the potential impact of a new FERS open
season on agency retirement costs. GAO developed illustrations that
estimated the effect if differing proportions of about 500,000
permanent general schedule CSRS and CSRS Offset Plan employees
(excluding employees of the Postal Service and certain other groups
such as Wage Grade employees) decided to switch. GAO used
information on current retirement practices and work it had performed
on the early implementation of FERS to provide perspective on the
difficulty of predicting how many employees might switch. GAO
applied different assumptions about the portion of eligible employees
who might switch and their salary levels and calculated the resulting
differences in agency retirement costs.
GAO found that:
-- It is difficult to predict who among the eligible employees
would switch. GAO's review of the first FERS transfer program
in 1987 showed that although eligible employees were provided
the information and counseling that they would need to make a
decision, about 4 percent of the eligible employees transferred
to FERS. The review suggested that employee decisions can be
based on situational factors that are economic as well as
noneconomic.
-- Assuming some employees opt to switch, agency retirement costs
would increase following an open season because of differences
in the way CSRS and FERS are funded. The amount of any cost
increase would critically depend on the number of employees who
switch and their salary levels. For example, assuming (1) a 10
percent switch rate--which would be twice the largest percentage
of employees who transferred in 1987 in the agencies GAO
reviewed--and (2) that employees earning higher salaries would
be most likely to switch, agency costs for the period January
1997 through December 1997 would have been an additional $332
million.
-- Given the uncertainty regarding how many employees might
transfer, it is correspondingly difficult to estimate whether
agencies would have a difficult time absorbing the cost
increases. Although the largest increase in retirement costs
GAO calculated--$332 million--would appear small in proportion
to the cost of personnel benefits governmentwide, some agencies
might find the costs difficult to absorb, depending on their
different circumstances. Regardless of the size of the
increases in costs, under the budget process discretionary
spending is capped, and Congress may choose not to provide
agencies extra funding to cover their increased retirement
costs.
THE FEDERAL EMPLOYEES' RETIREMENT
SYSTEM: POTENTIAL CHANGES IN
AGENCY RETIREMENT COSTS FOLLOWING
AN OPEN SEASON
==================================================== Chapter STATEMENT
Dear Mr. Chairman and Members of the Subcommittee:
I am pleased to be here today to discuss the potential impact on
agency retirement costs of a retirement system open season in which
anyone who is a general civilian employee of the federal government
and is participating in the Civil Service Retirement System (CSRS) or
the CSRS Offset plan would be allowed to transfer to the Federal
Employees Retirement System (FERS).\1 As you requested, the focus of
my statement is on the potential changes in retirement costs charged
to federal civilian agencies that might follow such a transfer
program. Because agencies would be responsible for paying any
additional costs, information about changes in retirement costs could
be important for workforce planning and management.
My statement has three main points. First, it is difficult to
predict the number and salary levels of those eligible employees who
would switch. Our review of the original FERS open season suggested
that employee decisions can be based on situational factors that are
economic as well as noneconomic. Second, assuming that some
employees opt to switch, agency retirement costs would increase
following an open season because of differences in the way CSRS and
FERS are funded. The amount of any such increase in agency
retirement costs would depend on the numbers of employees who switch
and their salary levels. Finally, given the uncertainty regarding
how many employees might transfer, it is correspondingly difficult to
estimate whether agencies would have a difficult time absorbing such
an increase. Although the largest increases in retirement costs that
we calculated could appear small in proportion to the costs of
personnel benefits governmentwide for the employees included in our
analysis, some agencies might find the costs difficult to absorb,
depending on their individual circumstances. Regardless of the size
of the cost increase, under the budget process discretionary spending
is capped, and Congress may choose not to provide agencies extra
funding to cover their increased retirement costs.
The Federal Employees' Retirement Open Season Act of 1997, passed by
Congress but vetoed by the President, would have allowed employees
currently enrolled in CSRS or the CSRS Offset program to transfer to
FERS. Under the act, an open season for transfers would have begun
on July 1, 1998, and run through December 31, 1998. Transferees
under the act would have received substantially the same coverage and
benefits that were provided to those who transferred during the first
FERS open season. Those who transferred would have received a single
annuity upon retirement. That annuity would have included a CSRS
component, calculated on the basis of the employee's final high-3
average salary and his or her years of CSRS service, and a FERS
component using the same high-3 salary and based on the employee's
years of FERS service. In addition, transferees would have received
Thrift Savings Plan (TSP) and Social Security benefits from their
FERS service.
Because of the Committee's interests, our analysis was only of the
potential changes in agency retirement costs that would affect
agencies' workforce and management decisions. We did not examine the
change in the total or the government's share of normal costs of the
CSRS and/or FERS programs, which is OPM's responsibility, or the
budgetary effects on discretionary and direct spending, which are the
responsibility of the Congressional Budget Office. It is worth
noting that any comparison of the total budgeted costs for CSRS and
FERS is driven by the fact, as discussed later in this statement,
that the costs are recorded in the budget on a different basis. Had
we examined the costs on a governmentwide basis the results may have
been different.
To assess the potential impact on agency retirement costs of a new
FERS open season, we developed several illustrations that estimated
the effect if differing proportions of the eligible employees
included in our analysis decide to switch. We used information from
work that we performed 10 years ago when we examined the early
implementation of FERS, including the first FERS open season and
employee decisions related thereto, and our knowledge of current
retirement practices to provide perspective on the difficulty of
predicting how many employees might switch to FERS.\2 In developing
our illustrations of the potential costs to agencies of a new open
season, we used benefit assumptions reflected in the recently vetoed
Federal Employees' Retirement System Open Enrollment Act of
1997--that transferees would receive substantially the same coverage
and benefits as those provided to employees who transferred during
the first FERS open enrollment. To illustrate the range of changes
in civilian agency retirement costs that might occur, we applied
different assumptions about the portion of eligible employees who
might switch and their salary levels, and calculated the resulting
differences in agency retirement costs. We developed two sets of
illustrative examples, each assuming that 1, 5, or 10 percent of the
eligible employees would transfer.\3 In making our calculations, we
used the most recent available data from the Federal Retirement
Thrift Investment Board on agency TSP payments, which were 1996 data.
We used 1997 Office of Personnel Management (OPM) data on the number
of nonpostal permanent civilian employees of the federal government
at the various grade levels\4 and 1996 OPM data on the share of CSRS
and FERS normal cost percentages charged to agencies for their CSRS
and FERS employees. As described in greater detail later in this
statement, normal cost is the term used to describe costs that are
calculated to reflect the cost of pension benefits as these benefits
are earned, rather than as they are paid. Our illustrations show how
agencies' retirement costs could have changed for the period from
January 1, 1997, through December 31, 1997, if all of the
transferring employees in each illustration were in FERS for the full
year.
--------------------
\1 Typically, the CSRS Offset plan covers employees who (1) had a
break in service that exceeded 1 year and ended after 1983 and (2)
had 5 years of creditable civilian service on January 1, 1987. CSRS
Offset plan coverage is described in greater detail in Federal
Retirement: Federal and Private Sector Retirement Program Benefits
Vary (GAO/GGD-97-40, April 7, 1997).
\2 Federal Personnel: Views From Two Agencies on Why More Employees
Did Not Join the New Retirement System (GAO/GGD-88-52FS, Mar. 11,
1988) and Federal Retirement: Implementation of the Federal
Employees Retirement System (GAO/GGD-88-107, Aug. 4, 1988).
\3 We selected 1 percent because it was the percentage used by the
Congressional Budget Office in making its preliminary cost estimate
of the vetoed Federal Employees' Retirement System Open Enrollment
Act of 1997. We selected 5 percent because it represented about the
largest percentage of employees who actually transferred in the
agencies we visited during our review of the first FERS open
enrollment. We added a 10-percent calculation to our analysis
because some observers have speculated that large numbers of eligible
employees would switch.
\4 OPM's Central Personnel Data File (CPDF) contains general schedule
employees, members of the Senior Executive Service (SES), blue collar
employees, and others. Of those employees in the CPDF, about 1.3
million are covered by the general schedule pay plan. Because the
methodology for our calculations used general schedule salaries for
calculating per employee increases in agency retirement costs, we
identified 1,042,112 permanent general schedule employees in Career,
Career-Conditional, Schedule A, and Schedule B appointments. Of the
general schedule employees who we used in our analysis, 548,377 were
covered by FERS, 457,084 were covered by CSRS, and 36,651 employees
were covered by the CSRS Offset plan. Had we analyzed changes in
agency retirement costs for non-general schedule employees (e.g.,
postal service employees, blue collar, and SES), our combined
estimate of these increases would have been higher.)
CSRS AND FERS ARE DESIGNED AND
FUNDED DIFFERENTLY
-------------------------------------------------- Chapter STATEMENT:1
CSRS and FERS are the two largest retirement programs for federal
civilian employees. At the beginning of fiscal year 1995, these
programs covered about 2.8 million federal employees, or 90 percent
of the civilian workforce, including postal employees. OPM
administers CSRS and FERS. CSRS and FERS pension benefits are
financed partly by federal agency and employee contributions and
partly by other government payments to the Civil Service Retirement
and Disability Fund (CSRDF).\5
--------------------
\5 The Department of the Treasury also makes annual payments that are
to cover interest on CSRS unfunded liability, payments for spouse
equity, and amortization payments to finance supplemental liabilities
for FERS, including those attributable to CSRS service for employees
who elected FERS. The Treasury Department's FERS payments are quite
small as a share of total normal costs. According to OPM actuarial
reports, in 1996 these payments were about $48 million, which
represents $1 in Treasury payments for each $138 in agency and
employee contributions. Thus, for all practical purposes, agency and
employee contributions cover the full normal cost of FERS benefits.
PROGRAM DESIGN DIFFERENCES
------------------------------------------------ Chapter STATEMENT:1.1
Although CSRS and FERS both provide pensions, the programs are
designed differently. CSRS was established in 1920 and predates the
Social Security system by 15 years. When the Social Security system
was established, Congress decided that employees in CSRS would not be
covered by Social Security through their federal employment. CSRS is
a stand-alone pension program that provides an annuity, determined by
a formula, as well as disability and survivor benefits. The program
was closed to new entrants after December 31, 1983. According to OPM
actuaries, the program is estimated to end in about 2070, when all
covered employees and survivor annuitants are expected to have died.
FERS was implemented in 1987, and generally covers those employees
who first entered federal service after 1983 as well as those who
transferred from CSRS to FERS. The primary impetus for the new
program was the Social Security Amendments of 1983, which required
that all federal employees hired after December 1983 be covered by
Social Security. FERS is a three-tiered retirement program that
includes Social Security and TSP benefits in addition to a pension.
Like CSRS, FERS provides disability and survivor benefits.
DIFFERENCES IN PROGRAM
FUNDING
------------------------------------------------ Chapter STATEMENT:1.2
CSRS and FERS also are funded differently. The costs of FERS
retirement benefits are paid by agencies and participating employees
as these benefits are earned, but some of the government's CSRS
pension costs are deferred.
To better appreciate how this occurs and the impact that the two
programs' different funding approaches have on agency retirement
costs, it is useful to understand differing methods for measuring
retirement program costs. The cost of a retirement plan is generally
not well measured by annual cash expenditures, because annual cash
expenditures simply cover the payments to existing beneficiaries in
any given year. An alternative measure of retirement costs assumes
that funds are set aside in roughly equal payments over the working
life of the employee to cover current and future liabilities for
benefit payments. "Normal cost" is the term used to describe costs
calculated in this manner; it measures the costs of pension benefits,
which are earned during an employee's working years but paid during
retirement, on an accrual rather than a cash basis. If resources
including interest amounts are set aside that are sufficient to fully
fund employees' retirement benefits as measured by normal cost
(including the pay increases for employees and cost of living
adjustments for annuitants), the retirement plan is said to be fully
funded on a dynamic normal cost basis.
It has long been our position that the appropriate way to calculate
and fund retirement costs is as the benefits are accruing.\6 When
done properly, recognizing retirement costs as they are being earned
reflects the full cost of providing these benefits to federal
personnel at the time their service is rendered. The annual normal
cost of CSRS, which is a much older program than FERS and predates
modern financing methods (which attempt to finance pension plans on
an actuarial basis),\7 is calculated for CSRS by OPM, but these costs
are not fully funded from agency and employee contributions.
According to OPM, when CSRS costs were estimated in this way, the
normal cost for fiscal year 1996 was 25.14 percent for CSRS. For
that year, the combined contributions of agencies and employees was
14 percent, or about 11 percent less than the full normal cost.
Beginning in fiscal year 1998, agencies will pay an additional 1.51
percent of pay towards their CSRS employees' retirement costs.
Employees covered by CSRS as well as those covered by FERS will also
pay an additional .25 percent in 1999, another 0.15 percent in 2000,
and a final extra 0.1 percent in 2001. Payment of a portion of the
difference between the full normal cost and agency and employee
contributions is funded through various means, and the remainder is
deferred.\8 When Congress established FERS in 1986, it adopted our
recommendation to charge agencies for all accruing retirement costs
not covered by employee contributions.
--------------------
\6 For example, see Federal Retirement Systems Unrecognized Costs,
Inadequate Funding, Inconsistent Benefits (GAO/FPCD-77-48, Aug. 3,
1977); The Design of a New Retirement Program for Federal Employees
Covered by Social Security (Apr. 12, 1985); Overview of Federal
Retirement Programs (GAO/T-GGD-95-172, May 22, 1995) ; and Federal
Retirement System Financing (GAO/T-GGD-95-197, Jun. 28, 1995).
\7 In establishing a defined benefit pension plan, an employer is
promising to pay benefits that will come due in the future.
Generally the money used to pay these benefits is obtained in one of
two ways, either through pay-as-you-go financing or through reserve
funding. The pay-as-you-go method would pay the pension benefits to
retired employees as these benefits come due out of appropriations.
Under the reserve funding method, contributions are made as benefits
are earned based on actuarial estimates of the value of the benefits.
CSRS was originally funded on a pay-as-you-go basis, and as a
consequence, built up an actuarial unfunded liability. Regarding
federal government pension plan liabilities, Public Law 95-595, 31
U.S.C. 9501-9504, enacted in 1978, established financial and
actuarial reporting requirements to such plans and required the plans
to report financial and actuarial information regarding plan
liabilities in annual reports. According to our summarization of
these reports in 1996, most of the 34 plans that we examined were
underfunded on a dynamic cost basis. However, FERS is fully funded,
and statutory provisions for the future elimination of the unfunded
benefit obligations of CSRS and the Military Retirement System have
already been enacted. See Public Pensions: Summary of Federal
Pension Plan Data (GAO/AIMD-96-6, Feb. 16, 1996).
\8 The funded portion is covered by other government contributions to
the retirement fund. OPM makes annual contributions to the fund from
its appropriation to amortize the liabilities created by employee pay
raises, once enacted, and other benefit improvements when they are
made; the Postal Service makes contributions to the fund to cover
retirement system liabilities resulting from collective bargaining
agreements with its employee unions and COLAS postal retirees
receive; and the Treasury pays the cost of benefits attributable to
military service and interest on the system's unfunded liability as
if it were funded. No provision exists to fund COLAs received by
nonpostal retirees. The remainder is addressed in the FERS statute
(Public Law 93-335) approved June 6, 1986, which requires that when
the budget authority in the retirement fund for CSRS benefits is
exhausted, automatic annual appropriations will be made to amortize
the shortfall over 30 years.
AGENCY COSTS FOR CSRS AND
FERS DIFFER
------------------------------------------------ Chapter STATEMENT:1.3
Because of these differences in the way in which CSRS and FERS are
designed and funded, agencies are charged more as a percentage of pay
for employees who are covered under FERS compared to their employees
who are covered under CSRS. As shown in table 1, when the three
components of the FERS benefit package are combined, beginning in
1999, agencies could contribute a total of 20.9 percent of pay for
their FERS employees, compared with 8.5 percent for their CSRS
employees.\9 Under these assumptions, agencies would pay a higher
percentage of pay for each employee who transfers from CSRS to FERS
than is currently paid. This is because agencies would need to
contribute 2.2 percent more for the FERS pension benefit, plus an
additional 6.2 percent for the employer's share of the Social
Security benefit up to a maximum salary of $65,400, and about 4
percent for the TSP benefit.\10 As the table also shows, the
percentage increase in agency costs per transferring employee would
be smaller in the case of employees under the CSRS Offset plan.
Table 1
Comparison of the Percentages of Pay
Agencies Could Contribute for Employees
Covered By the FERS, CSRS, and CSRS
Offset Retirement Plans
Federal retirement plans
--------------------------
CSRS
FERS CSRS Offset
------------------------------------------ ------ ------ ----------
Defined benefit 10.7% 8.5% 8.5%
Social Security\a 6.2% 6.2%
TSP (average match) 4.0%
Agencies' total contribution percentages 20.9% 8.5% 14.7%
Percentage increase in agencies' costs if 12.4% 6.2%
employees transfer to FERS
----------------------------------------------------------------------
\a As is true for private employers, this shows the employer
contribution to Social Security, not the full actuarial cost of the
system. See Retirement Income: Implications of Demographic Trends
for Social Security and Pension Reform (GAO/HEHS-97-81, July 11,
1997).
Source: 1997 OPM actuarial data.
Although agency costs generally could rise by 12.4 percent of pay for
each employee who transferred to FERS during the open season, the
actual dollar impact of each such transfer would vary depending on
employees' salaries, with the most highly compensated transferring
employees causing the greatest increase in agency retirement costs.
Table 2 provides the range of these dollar increases for general
schedule (GS) employees and shows that the smallest increase would be
$1,571 per employee, and the greatest would be about $9,769. Our
results for the CSRS Offset plan participants ranged from a low of
$785 to a high of $5,714 and are presented in appendix I.
Table 2
Comparison of Agency Cost Increases for
Each CSRS General Schedule Employee Who
Transfers to FERS, by Grade and Step
Steps
----------------------------------------------------------
Grades 1 2 3 4 5 6 7 8 9 10
-------------------- ---- ---- ---- ---- ---- ---- ---- ---- ---- ----
GS-1 $1,5 $1,6 $1,6 $1,7 $1,7 $1,8 $1,8 $1,9 $1,9 $1,9
71 23 75 28 80 11 62 14 16 65
GS-2 1,76 1,80 1,86 1,91 1,93 1,99 2,05 2,10 2,16 2,22
6 8 7 6 8 5 2 9 6 3
GS-3 1,92 1,99 2,05 2,12 2,18 2,24 2,31 2,37 2,44 2,50
7 1 6 0 4 8 3 7 1 5
GS-4 2,16 2,23 2,30 2,38 2,45 2,52 2,59 2,66 2,74 2,81
3 6 8 0 2 4 6 9 1 3
GS-5 2,42 2,50 2,58 2,66 2,74 2,82 2,90 2,98 3,06 3,14
0 1 2 3 3 4 5 6 6 7
GS-6 2,69 2,78 2,87 2,96 3,05 3,14 3,23 3,32 3,41 3,50
8 8 8 8 8 7 7 7 7 7
GS-7 2,99 3,09 3,19 3,29 3,39 3,49 3,59 3,69 3,79 3,89
8 8 8 8 8 8 8 8 8 8
GS-8 3,32 3,43 3,54 3,65 3,76 3,87 3,98 4,09 4,20 4,31
0 1 2 3 3 4 5 5 6 7
GS-9 3,66 3,79 3,91 4,03 4,15 4,27 4,40 4,52 4,64 4,76
8 0 2 4 7 9 1 3 6 8
GS-10 4,03 4,17 4,30 4,44 4,57 4,71 4,84 4,98 5,11 5,25
9 3 8 3 7 2 7 1 6 1
GS-11 4,43 4,58 4,73 4,88 5,02 5,17 5,32 5,47 5,62 5,76
7 5 3 1 9 7 5 3 1 9
GS-12 5,31 5,49 5,67 5,85 6,02 6,20 6,38 6,56 6,73 6,91
8 6 3 0 8 5 2 0 7 4
GS-13 6,32 6,53 6,74 6,95 7,16 7,37 7,58 7,80 8,01 8,16
4 5 6 7 8 8 9 0 1 6
GS-14 7,47 7,72 7,97 8,16 8,29 8,41 8,53 8,66 8,78 8,91
3 3 2 5 0 4 9 3 8 3
GS-15 8,45 8,59 8,74 8,89 9,03 9,18 9,32 9,47 9,62 9,76
0 7 3 0 6 3 9 6 2 9
--------------------------------------------------------------------------------
Source: GAO analysis of OPM data.
--------------------
\9 The CSRS Offset percentage is 14.7. The figures presented in
figure 1 and used for our analysis are only designed to illustrate
the potential changes in agency retirement costs. As noted earlier
in this statement, in 1996, agencies actually contributed 7 percent
for the CSRS defined benefit and will continue to contribute this
percentage through fiscal year 1997. Also in this regard, in 1996,
agencies contributed 11.4 percent for the FERS defined benefit,
although as noted elsewhere in this statement and as used by CBO in
its cost estimate of the budgetary impacts of an open season, the
most recent OPM estimate of this cost is 10.7 percent.
\10 With respect to the pension benefit, agencies are to pay the
difference between employee contributions of 0.8 percent and the full
dynamic normal cost. According to the most recent OPM actuarial
estimates, this difference is 10.7 percent. With respect to TSP, the
maximum employer potential cost, or liability, for contributions for
FERS-covered employees is 5 percent of pay--consisting of up to 4
percent in matching contributions, plus 1 percent in nonmatching
contributions. In 1996, the most recent year for which data were
available, the Thrift Board reported that agencies contributed about
$2 billion to FERS employees' TSP accounts. According to OPM
actuaries, the dynamic normal cost of TSP benefits has been rising
and is now estimated to be about 4 percent of pay.
IT IS DIFFICULT TO PREDICT THE
NUMBER OR SALARY LEVEL OF
EMPLOYEES WHO WOULD TRANSFER TO
FERS
-------------------------------------------------- Chapter STATEMENT:2
Because each transferring employee would add to an agency's
retirement costs, assessing the effect of an open season on agencies'
retirement costs is critically dependent on the number of employees
who would switch and their salary levels. Exclusive of the Postal
Service, about 457,000 permanent civilian GS employees covered by
CSRS and about 37,000 covered by the CSRS Offset plan would be
eligible and could transfer to FERS during a new open season. Given
the potential range of impacts on agency retirement costs and the
difficulty of making a precise estimate of these costs, it is useful
to consider what factors employees might consider in making their
decisions and the salary levels of those who might be most likely to
switch.
When FERS was created, Congress asked us to evaluate the act's
initial implementation. During the course of our work, we developed
information on the reasons why eligible employees chose not to
transfer to FERS. One lesson from this work was that it is difficult
to predict whether an individual employee will decide to change his
or her retirement coverage. For example, at the time of the initial
open season in 1987, the Office of Management and Budget estimated
that as many as 40 percent of eligible employees would transfer to
FERS, while after the open season ended in January 1988, OPM
confirmed that about 86,000 CSRS employees (about 4 percent) actually
transferred..
Also, transfer rates during the first FERS open season varied across
the agencies that we reviewed. At sites we visited, we found that
transfer rates ranged from less than 1 percent to more than 4
percent, which translates into a fourfold difference in the increases
in the agencies' retirement costs.\11
During our 1987 work, we also examined the manner in which agencies
fulfilled their implementation responsibilities during the open
season, in part to understand the role that information provided to
employees may have played in their decisions. We found that although
fewer eligible employees than expected actually transferred, there
were no underlying deficiencies in the implementation of the transfer
program that might have accounted for the low percentages. FERS
information was widely available and distributed. Advisors who were
to provide individual counseling were trained by OPM, and these
advisors were available to assist employees. Computer models were
also widely available, and analyses and estimates were generally
provided to employees who requested them. Notwithstanding such
agency efforts, briefings at the sites we visited were not well
attended. Also, only a small percentage of employees requested
computer estimates of their potential retirement benefits.
We interviewed personnel officials and advisors who were responsible
for counseling employees during 1987 on the advantages and
disadvantages of transferring to FERS. On the basis of the views of
advisors and personnel officials at the sites we visited, we
identified four primary reasons why employees decided not to transfer
to FERS. The reasons--both economic and noneconomic--were as
follows:
-- Employees regarded FERS as too complex to understand.
-- Employees believed they could not afford to contribute to TSP.
-- Employees planned to make the federal government their career
and believed that CSRS provided greater benefits for career
employees than FERS.
-- Employees did not trust various aspects of the design or
stability of FERS, including the viability of the Social
Security system and potential for future changes in FERS benefit
levels.
In contemplating the relevance of employees' 1987 reasoning for a new
open season, it is worth noting that much has changed in the past
decade. For instance, although the number of FERS retirees is still
small relative to CSRS retirees, FERS is no longer an unknown or
fledgling retirement program. The benefits available from
participation in FERS likely are better understood today than in 1987
when the program was created. Of particular importance may be a
general improvement in public understanding of 401(k) plans and the
role that they now play in retirement savings. Although the growth
of these plans may be due in part to their popularity with employers,
the plans also enjoy increased employee popularity because they are
seen as an important element of retirement income and as having the
advantage of portability. Also, notwithstanding the recent
volatility of stock markets worldwide, sustained economic growth has
boosted plan earnings and increased total assets substantially over
the past decade, including those of the TSP C fund. On the other
hand, should the U.S. economy falter or market gains abate or turn
to losses, TSP might seem less attractive to those employees who are
considering whether to change their retirement coverage.
As to the TSP affordability issue cited by employees, CSRS
participants--who are allowed to contribute up to 5 percent of pay on
a tax-deferred basis to their own TSP accounts--are doing so in large
numbers. In 1996, more than half of the CSRS-covered employees in
civilian agencies contributed to TSP. Participation rates ranged
from a low of about 42 percent at the Postal Service to a high of
about 75 percent at the Education Department. Also, more than 70
percent of CSRS employees who were contributing to TSP in 1996
contributed the maximum allowed.
In part, because 10 years have elapsed since the last open season,
individual employees' personal circumstances could be considerably
different. In general, more CSRS employees would be closer to
retirement eligibility, many may have moved to higher salary levels,
and still others' family responsibilities may have changed--either by
increasing or decreasing.
It is reasonable to expect that such economic considerations would
play an important role in employees' decisions about transferring to
FERS. Given such changes, some employees might find a financial
advantage in changing their retirement coverage now who did not in
1987. For example, some economic incentives to switch include the
following:
-- Changes in personal circumstances such as advancing to higher
salary levels could be contributing to the increased CSRS
employee participation in TSP. For example, as shown in figure
1, about 80 percent of CSRS employees earning between $60,000
and $65,000 contributed the maximum 5 percent of their salaries
into TSP. Some of these employees might be attracted to FERS to
take advantage of the opportunity to put up to 10 percent of
their salaries, which would result in a 5-percent contribution
by the government.
-- After a 5-year period of FERS participation, any CSRS employee
who switched would become eligible upon retirement for the full
spousal benefit provided under Social Security. CSRS retirees
have these benefits partially or entirely offset by their CSRS
pensions.
-- The approximately 4,000 CSRS employees who have worked long
enough to qualify for the maximum retirement benefit of 80
percent of their high-3 average salary could earn greater
retirement benefits by transferring to FERS, because they would
qualify for agency matching contributions to TSP, Social
Security, and other benefits. The pensions that they would
receive as transferees would include the maximum allowable CSRS
benefit, plus the additional FERS annuity.
-- Lower-salaried employees may fare well under FERS, because
Social Security benefits are weighted toward lower income
workers. Although employees who participate in CSRS do not
receive Social Security benefits from their federal service,
those participants who transfer to FERS would receive pensions
combining CSRS and FERS benefits, and they would be eligible to
receive Social Security benefits from their federal service.
However, this advantage to low wage earners could be
considerably reduced by the Windfall Elimination provision,
which would significantly reduce Social Security benefits for
transferees who do not have at least 21 years of substantial
Social Security coverage.
Figure 1: Percent of CSRS
Participants Who Contributed 5
Percent of Income in 1996, by
Selected Salary Brackets
(See figure in printed
edition.)
Source: Federal Retirement Thrift Investment Board data.
On the other hand, concerns about economic risks or uncertainty could
dissuade some employees from transferring. As noted above, concern
about the reliability of Social Security benefits was an important
factor in employee decisions during the first open enrollment season.
Current opinion polls suggest that many Americans remain concerned
about whether these benefits will be a reliable source of retirement
income. Our review of the last open season also suggests that
factors such as the perceived complexity of the FERS program and its
attendant risks also can play important roles in employees'
decisions. For example, some employees have less than 40 years of
service and thus have not reached the maximum CSRS benefit. These
employees might not want to exchange the larger annuities that
additional CSRS service would provide for smaller FERS annuities that
are coupled with the risk of losses from TSP investments that could
result an overall lower retirement income.\12
Finally, although economic factors logically would play an important
role in employee decisions, noneconomic considerations could also
play a significant role. Choices about when is the right time to
retire are very personal and reflect individual values and
circumstances as well as economic considerations. Thus, these
decisions could be based on factors that are hard, if not impossible,
to quantify. Upon transferring to FERS, for example, employees would
need to work an additional 5 years to avoid the public pension offset
rules that apply to CSRS employees. However, working the additional
5 years might take some individuals beyond the date at which they had
planned to retire, and the trade-off between additional retirement
income and a shorter retirement would be a difficult personal
decision. In the final analysis, much like employee decisions to
take or forego recent early-out opportunities, it may be possible to
estimate who would benefit from a FERS open season transfer on
economic grounds, but the deciding factors might in fact be known
only to the employees.
--------------------
\11 We visited 23 Department of the Army and Veterans Administration
field activities. We selected two large agencies, one military and
one civilian. Both employed a large number of civilian employees and
had numerous field activities widely dispersed throughout the
country. However, because we did not randomly select the agencies or
their field locations, the information we obtained could not be
projected to portray the implementation of FERS throughout the Army,
Veterans Administration, or the government.
\12 The amount of the reduction in annuity benefits would depend on
the number of years of FERS service that would be applied to the
employee's eventual annuity calculation.
AGENCY RETIREMENT COST
INCREASES WOULD DEPEND ON THE
NUMBER AND SALARY LEVELS OF
EMPLOYEES WHO TRANSFER
-------------------------------------------------- Chapter STATEMENT:3
Given the difficulty of predicting the number and characteristics of
the employees who ultimately might transfer to FERS, we developed two
sets of examples to illustrate the range of increases in retirement
costs that agencies could face following a new FERS open season.
Using two different sets of assumptions, we calculated the increase
in agency retirement costs if 1, 5, and 10 percent of the eligible
employees transferred to FERS. For the first set of calculations, we
assumed that the distribution of transferring employees for each GS
pay grade matched their actual distribution, governmentwide. That
is, if 2.2 percent of the eligible CSRS employees were in grade GS-4,
then 2.2 percent of the employees included in the 1-, 5-, and
10-percent calculations would also be in grade GS-4. For the second
set of calculations, we assumed that a larger proportion of the
transferring employees would be earning salaries at higher GS grades.
This assumption helps to illustrate the sensitivity of the agency
retirement cost increases to the salary level of transferring
employees and may be especially pertinent if employees nearing the
end of their careers, or those who wish to take advantage of the FERS
TSP higher maximum contribution rate are most likely to transfer.
Table 3 illustrates the potential increases in agency retirement
costs if 1, 5, and 10 percent of the eligible employees covered by
CSRS transferred to FERS, assuming that the transfer rates across the
GS grades would match the current distribution of employees by grade
and step. As the table shows, under these assumptions the increase
in agency costs for the full calendar year 1997 would have ranged
from about $24 million more if 1 percent of the employees transferred
to about $244 million more if 10 percent transferred. Our results
for CSRS Offset plan participants ranged from less than $1 million to
about $8 million and are presented in appendix I.
Table 3
Comparison of the Estimated Increase in
Agency Costs for General Schedule
Employees Covered by CSRS
Total cost of selected rates of
transfer
----------------------------------
Grades 1 percent 5 percent 10 percent
---------------------------------- ---------- ---------- ----------
GS-1 $133\a $663 $1,326
GS-2 2,694 13,472 26,944
GS-3 39,971 199,856 399,711
GS-4 268,896 1,344,478 2,688,957
GS-5 850,251 4,251,255 8,502,510
GS-6 806,650 4,033,250 8,066,499
GS-7 1,389,418 6,947,090 13,894,179
GS-8 687,775 3,438,877 6,877,753
GS-9 2,017,484 10,087,421 20,174,842
GS-10 319,857 1,599,285 3,198,569
GS-11 4,321,219 21,606,095 43,212,189
GS-12 6,483,941 32,419,705 64,839,410
GS-13 4,273,407 21,367,034 42,734,067
GS-14 1,985,738 9,928,692 19,857,384
GS-15 945,820 4,729,099 9,458,198
======================================================================
Total $24,395,34 $121,965,7 $243,931,4
0 21 42
----------------------------------------------------------------------
Note 1: For this table, we estimated the total costs associated with
each selected rate of transfer by assuming that the proportion of
CSRS employees who transfer would match the actual distribution of
these employees across GS grades governmentwide.
Note 2: Due to rounding, totals may not equal the transfer rate
percentages shown above.
\a Because only seven of the CSRS employees whom we included in our
analyses were in grade GS-1, 1 percent of these seven represents the
cost difference for less than one person.
Source: GAO analysis of OPM data.
Table 4 shows the increase in agency costs if 1, 5, or 10 percent of
the eligible employees covered under CSRS transferred to FERS under
the different assumption that transfer rates would be higher for
employees in grades GS-13 through GS-15. Following this logic, the
calculations assume that 75 percent of the total number of employees
who would transfer would be in grades GS-13 through GS-15, and the
remaining 25 percent of the transferring employees would be in grades
GS-1 through GS-12. Under these assumptions, the 1 year agency costs
would range from a low of about $32 million if 1 percent of the
employees transferred to a high of $320 million if 10 percent
transferred. Our results for CSRS Offset plan participants ranged
from about $1 million to about $13 million and are presented in
appendix I.
Table 4. Comparison of the Estimated
Increase in Agencies' Costs for General
Schedule Employees Covered by CSRS if
the Concentration of Those Transferring
to FERS Were at the Higher Grade Levels.
Total cost of selected rates of
transfer
----------------------------------
Grades 1 percent 5 percent 10 percent
---------------------------------- ---------- ---------- ----------
GS-1 to GS-12 $ 26,964,573 53,934,002
5,393,886 $
$
GS-13 to GS-15 26,585,497 132,919,51 265,831,04
0 4
======================================================================
Total $31,979,38 $159,884,0 $
3 83 319,765,04
6
----------------------------------------------------------------------
Note 1: For this table, we estimated the total costs associated with
each selected rate of transfer by assuming that 75 percent of CSRS
employees who transfer would be in grades GS-13 through GS-15, and 25
percent of the employees would be in grades GS-1 through GS-12.
Note 2: Due to rounding, totals may not equal the transfer rate
shown.
Source: GAO analysis of OPM data.
AGENCIES MIGHT HAVE TO PAY ANY
ADDITIONAL RETIREMENT WITHOUT
ADDITIONAL RESOURCES
-------------------------------------------------- Chapter STATEMENT:4
Regardless of the size of the increase in agencies' retirement costs
may be, the way in which agencies' budget resources are controlled
under the current budget process suggests that agencies might have to
absorb any increase. Agency spending to pay retirement costs is
discretionary spending. Under the current process, discretionary
spending is subject to fixed-dollar caps that are implemented through
the budget and appropriations processes.\13 As a consequence, unless
Congress chose to provide additional resources to fund the added
retirement costs that agencies could be charged, agencies would need
to absorb any increases, within the limits of their annual
appropriations.\14
Given the uncertainty regarding how many employees might transfer to
FERS during an open season, it is correspondingly difficult to
estimate whether agencies would have a difficult time absorbing such
an increase. By one measure--the percentage that the increase in
retirement costs might be of agencies' total expenditures for
salaries and benefits--the increased retirement costs do not look
imposing. For example, our highest estimate of increased
governmentwide costs, about $332 million, would represent
approximately 1 percent increase in expenditures for salaries and
benefits for the employees used in our analysis, based on fiscal year
1997 figures. Of course, depending on how "personnel-intensive"
specific agencies' operations may be or what other cost increases may
arise from other sources, even such a small overall percentage
increase could be difficult to absorb in certain situations, and
might, for example, result in reductions in staff levels or capital
spending. Finally, agencies would be affected differently, depending
on their particular grade structure. For example, an agency with a
relatively larger share of highly salaried employees who opt to
transfer would experience larger increases in its retirement costs
compared to an agency with a smaller share, assuming that both
agencies experienced the same transfer rates.
--------------------
\13 In particular, as noted in Budget Policy: Issues in Capping
Mandatory Spending (GAO/AIMD-94-155, July 18, 1994), congressional
budget resolutions set totals by budget function and accompanying
statements to the conference reports allocate funds to the
appropriations committees for discretionary programs. House and
Senate appropriations committees subsequently allocate these totals
among their subcommittees. OMB keeps score by tracking congressional
actions, and Congress has established spending levels in each
congressional budget resolution. Should appropriations exceed the
discretionary cap, the Budget Enforcement Act provides for
eliminating the overage through the sequestration of discretionary
spending. Policymakers vote annually on these discretionary program
appropriations.
\14 To provide additional appropriations for this purpose would
likely require an increase in the discretionary caps.
------------------------------------------------ Chapter STATEMENT:4.1
That concludes my prepared statement. I would be pleased to answer
any questions you or other Members of the Committee may have.
=========================================================== Appendix I
Table I.1. Comparison of Agency Cost
Increases for Each CSRS Offset Employee
Who Transfers to FERS, by Grade and Step
Steps
----------------------------------------------------------
Grades 1 2 3 4 5 6 7 8 9 10
-------------------- ---- ---- ---- ---- ---- ---- ---- ---- ---- ----
GS-1 785 812 838 864 890 905 931 957 958 982
GS-2 883 904 933 958 969 997 1,02 1,05 1,08 1,11
6 4 3 2
GS-3 964 996 1,02 1,06 1,09 1,12 1,15 1,18 1,22 1,25
8 0 2 4 6 8 1 3
GS-4 1,08 1,11 1,15 1,19 1,22 1,26 1,29 1,33 1,37 1,40
2 8 4 0 6 2 8 4 0 6
GS-5 1,21 1,25 1,29 1,33 1,37 1,41 1,45 1,49 1,53 1,57
0 1 1 1 2 2 2 3 3 3
GS-6 1,34 1,39 1,43 1,48 1,52 1,57 1,61 1,66 1,70 1,75
9 4 9 4 9 4 9 4 9 4
GS-7 1,49 1,54 1,59 1,64 1,69 1,74 1,79 1,84 1,89 1,94
9 9 9 9 9 9 9 9 9 9
GS-8 1,66 1,71 1,77 1,82 1,88 1,93 1,99 2,04 2,10 2,15
0 6 1 6 2 7 2 8 3 8
GS-9 1,83 1,89 1,95 2,01 2,07 2,13 2,20 2,26 2,32 2,38
4 5 6 7 8 9 1 2 3 4
GS-10 2,01 2,08 2,15 2,22 2,28 2,35 2,42 2,49 2,55 2,62
9 7 4 1 9 6 3 1 8 5
GS-11 2,21 2,29 2,36 2,44 2,51 2,58 2,66 2,73 2,81 2,88
9 3 7 1 5 9 3 6 0 4
GS-12 2,65 2,74 2,83 2,92 3,01 3,10 3,19 3,28 3,36 3,45
9 8 7 5 4 2 1 0 8 7
GS-13 3,16 3,26 3,37 3,47 3,58 3,68 3,79 3,90 4,00 4,11
2 8 3 8 4 9 5 0 5 1
GS-14 3,73 3,86 3,98 4,11 4,23 4,36 4,48 4,60 4,73 4,85
7 1 6 0 5 0 4 9 3 8
GS-15 4,39 4,54 4,68 4,83 4,98 5,12 5,27 5,42 5,56 5,71
5 2 8 5 1 8 4 1 7 4
--------------------------------------------------------------------------------
Source: GAO analysis of OPM data.
Table I.2. Comparison of the Estimated
Increase in Agency Costs for General
Schedule Employees Covered by the CSRS
Offset Plan
Total cost of selected rates of
transfer
----------------------------------
Grades 1 percent 5 percent 10 percent
---------------------------------- ---------- ---------- ----------
GS-1\a $0 $0 $0
GS-2 156 779 1,558
GS-3 4,310 21,550 43,101
GS-4 29,292 146,459 292,919
GS-5 76,991 384,957 769,914
GS-6 68,269 341,346 682,692
GS-7 91,249 456,245 912,491
GS-8 32,855 164,273 328,547
GS-9 77,874 389,370 778,741
GS-10 8,232 41,160 82,321
GS-11 118,355 591,774 1,183,547
GS-12 144,041 720,203 1,440,406
GS-13 92,328 461,639 923,277
GS-14 39,548 197,742 395,484
GS-15 22,982 114,910 229,820
======================================================================
Total $806,661 $4,033,305 $8,064,370
----------------------------------------------------------------------
Note 1: For this table, we estimated the total costs associated with
each selected rate of transfer by assuming that the proportion of
CSRS Offset employees who transfer would match the actual
distribution of these employees across GS grades governmentwide.
That is, because 6.1 percent of the eligible CSRS employees were in
grade GS-4, 6.1 percent of the CSRS Offset employees included in the
1-, 5-, and 10-percent cost estimates were also in grade GS-4.
Note 2: Due to rounding, totals may not equal the transfer rate
percentages shown above.
\a Because no CSRS Offset employees were in grade 1, the cost
difference is $0.
Source: GAO analysis of OPM data.
Table I.3. Comparison of the Increase in
Agency Costs for General Schedule
Employees Covered by the CSRS Offset
Plan if the Concentration of Those
Transferring to FERS Were at the Higher
Grade Levels
Total cost of selected rates of
transfer
----------------------------------
Grades 1 percent 5 percent 10 percent
---------------------------------- ---------- ---------- ----------
GS-1 to GS-12 $182,590 $912,951 $1,825,901
GS-13 to GS-15 1,079,569 5,397,844 10,791,690
======================================================================
Total $1,262,159 $6,310,795 12,617,591
----------------------------------------------------------------------
Note 1: For this table, we estimated the total costs associated with
each selected rate of transfer by assuming that 75 percent of CSRS
Offset employees who transfer would be in grades GS-13 through GS-15,
and 25 percent of the employees would be in grades GS-1 through
GS-12.
Note 2: Due to rounding, totals may not equal the transfer rate
percentages shown.
Source: GAO analysis of OPM data.
*** End of document. ***