Private Pensions: Information on Cash Balance Pension Plans	 
(03-NOV-05, GAO-06-42). 					 
                                                                 
The nation's private defined benefit (DB) pension system, a key  
contributor to the financial security of millions of Americans,  
is in long-term decline. Since 1980, the number of active	 
participants in Pension Benefit Guaranty Corporation (PBGC)	 
insured single employer DB plans has dropped from 27.3 percent of
all national private wage and salary workers in 1980, to about 15
percent in 2002, and more recently the PBGC has assumed billions 
of dollars in unfunded benefit obligations from bankrupt plan	 
sponsors. Some analysts have identified hybrid DB plans like cash
balance (CB) plans as a possible means to revitalize this	 
declining system. However, conversions from traditional DB plans 
to CB plans have sometimes been controversial because of the	 
effect conversions may have on the benefits of workers of	 
different ages. As House and Senate committees consider 	 
comprehensive pension reform legislation that includes efforts to
resolve uncertainties about CB plans, GAO was asked to (1) review
current research about the implications of CB conversions for	 
employee benefits, (2) describe the prevalence and type of	 
transition provisions used to protect workers' benefits in past  
CB conversions, and (3) estimate the effects of CB conversions on
the benefits of individual participants under a hypothetical	 
conversion to a typical CB plan from a typical traditional DB	 
plan.								 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-06-42						        
    ACCNO:   A40459						        
  TITLE:     Private Pensions: Information on Cash Balance Pension    
Plans								 
     DATE:   11/03/2005 
  SUBJECT:   Employee retirement plans				 
	     Fringe benefits					 
	     Older workers					 
	     Pensions						 
	     Retirement benefits				 
	     Retirement income					 
	     Comparative analysis				 
	     Cash balance plans 				 
	     Defined benefit plans				 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO 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.                                         **
**                                                              **
******************************************************************
GAO-06-42

     

     * Appendix I: Information on Cash Balance Pension Plans
     * Appendix II: Review of Literature on Cash Balance Plans
     * Appendix III: Analysis of Form 5500 Data on Cash Balance Plans
     * Appendix IV: Analysis of Simulated Cash Balance Plans and Traditional
       Final Average Pay Plans
     * Appendix V: GAO Contacts and Staff Acknowledgments

                 United States Government Accountability Office

GAO

                                  October 2005

PRIVATE PENSIONS

                   Information on Cash Balance Pension Plans

GAO-06-42

PRIVATE PENSIONS

Information on Cash Balance Pension Plans

  What GAO Found

Current pension and economic literature provides little conclusive
evidence about the effects of CB plan conversions on benefits. In many
cases, data and other methodological issues (e.g., sampling methods) limit
the generalization of results. Nonetheless, cash balance research
indicates that the effects of a conversion depend on many factors,
including the generosity of the CB plan, transition provisions that might
limit any adverse effects on current employees, and firm-specific employee
demographics. CB plan conversions are posited to have distributional
effects on expected pension wealth: younger, more mobile workers usually
benefit while older workers with long job tenure are likelier to
experience a loss, particularly if they are nearly eligible for early
retirement.

GAO's analysis of a representative sample of plan conversions determined
that most conversions occurred between 1990 and 1999 and primarily in the
manufacturing, health care, finance and insurance industries. Most
conversions set participants' opening account balances equal to the
present value of benefits accrued under the previous plan, although the
interest rate used to calculate the balance varied around the 30-year
Treasury bond rate. Most plans provided some form of transition provisions
to mitigate the potential adverse effects of a conversion on workers'
expected benefits for at least some employees. About 47 percent of all
conversions grandfathered at least some of the employees into the former
traditional DB plan. In most cases, grandfathering eligibility was limited
to employees meeting a specified minimum age and/or years of service.

GAO's simulations of the effects of conversions on pension benefits show
the following:

     o In conversions from a traditional DB plan to a typical CB plan, most
       workers, regardless of age, would have received greater benefits under
       the DB plan. Unless grandfathered into the former plan, older workers
       experience a greater loss of expected benefits than younger workers.
     o In comparing a typical CB plan to a terminated FAP plan, all vested
       workers would do better under the CB plan.
     o In conversions from a traditional DB plan to a CB plan of equal cost
       to the sponsor and more generous than the typical CB plan, while more
       workers at age 30 have benefit increases under the CB plan, this was
       not true for those at age 40 and 50.
     o In comparing a equal cost CB plan to a terminated FAP plan, again all
       vested workers would do better under the CB plan.
     o GAO's comparisons focusing on the lifetime present value of benefits
       did not change the basic findings of GAO's analysis of monthly
       benefits.

                 United States Government Accountability Office

Contents

  Letter 1

Concluding Observations 8 Agency Comments 9

Appendix I Information on Cash Balance Pension Plans

Appendix II Review of Literature on Cash Balance Plans

  Appendix III Analysis of Form 5500 Data on Cash Balance Plans

Appendix IV Analysis of Simulated Cash Balance Plans and Traditional Final
Average Pay Plans

Appendix V GAO Contacts and Staff Acknowledgments

  Tables

T able 1: Cash Balance Plan Sample Disposition 57 Table 2: Conversions
Costs of Typical Cash Balance Plan by

Conversion Age 66

Abbreviations

APS            age plus service                                           
BLS            Bureau of Labor Statistics                                 
CB             cash balance                                               
CBOLT             Congressional Budget Office's long-term social security 
                  model                                                      
DB             defined benefit                                            
DC             defined contribution                                       
LABOR          Department of Labor                                        
GAM            Group Annuitant Mortality                                  
FAP            final average pay                                          
PBGC           Pension Benefit Guaranty Corporation                       
PENSIM         pension policy microsimulation model                       
PSG            Policy Simulation Group                                    
PSID           Panel Study of Income Dynamics                             
SIPP           Survey of Income and Program Participation                 
SPD            summary plan description                                   

This is a work of the U.S. government and is not subject to copyright
protection in the United States. It may be reproduced and distributed in
its entirety without further permission from GAO. However, because this
work may contain copyrighted images or other material, permission from the
copyright holder may be necessary if you wish to reproduce this material
separately.

United States Government Accountability Office Washington, DC 20548

November 3, 2005

The Honorable George Miller Ranking Minority Member Committee on Education
and the Workforce United States House of Representatives

The Honorable Bernard Sanders Ranking Minority Member Subcommittee on
Financial Institutions

and Consumer Credit Committee on Financial Services United States House of
Representatives

The Honorable Tom Harkin United States Senate

The nation's private defined benefit (DB) pension system, 1 a key
contributor to the financial security of millions of American workers and
their families, is in long-term decline. The number of single employer DB
plans has declined dramatically over the past several decades, 2 from over
95,000 in 1980 to less than 35,000 in 2002, with the number of active
participants in such plans dropping from 27.3 percent of all national
private wage and salary workers in 1980, to about 15 percent in 2002. 3
Structural problems in industries like airlines, steel, and auto parts
have led to large bankrupt firms terminating their DB plans, with
thousands of workers losing some of their benefits and saddling the
Pension Benefit

1

In DB plans, formulas set by the employer determine employee benefits. DB
plan formulas vary widely, but benefits are frequently based on
participant pay and years of service.

2

Single employer plans provide benefits to employees of one employer or, if
under common control, employees of several related employers.
Multiemployer plans are DB plans created by collective bargaining
agreements covering more than one employer and generally operate under the
joint trusteeship of labor and management. These plans cover over 9.7
million participants or about 22 percent of all workers and retirees
insured by the Pension Benefit Guaranty Corporation (PBGC). See GAO,
Private Pensions: Multiemployer Plans Face Short and Long-Term Challenges,
GAO 04-423 ( Washington, D.C.: Mar. 26, 2004).

3

Pension Benefit Guaranty Corporation, Pension Insurance Data Book,
(Washington, D.C., 2004).

Guaranty Corporation (PBGC) with billions of dollars in unfunded benefit
guarantees. 4 In response, several congressional committees have proposed
comprehensive pension reform legislation that, among other issues, would
address the underfunding of single employer defined benefit plans. 5

Some analysts have identified hybrid DB plans like cash balance (CB) plans
as a possible means to revitalize this declining system. CB plans are
referred to as hybrid plans because legally they are DB plans but contain
certain features that resemble defined contribution plans. Similar to
traditional DB plans, CB plans use a formula to determine pension
benefits. However, unlike traditional final average pay (FAP) plans that
pay retirement benefits on the basis of an annuity amount calculated using
years of service and earnings, CB plans express benefits as a hypothetical
individual account balance that is based on pay credits (percentage of
salary or compensation) and interest credits, rather than an annuity.

In the late 1990s, many pension plan sponsors converted their traditional
final average pay plans to CB plans. Conversions to CB plans have been
controversial because of the effect they may have on pension benefits of
workers of different ages and years of service. 6 In particular, CB plan
conversions can sometimes result in so-called "wearaway" situations where
some workers do not earn additional pension benefits while other workers
continue to do so. 7 The legality of CB plans has recently been questioned
in a court ruling regarding whether a CB plan is age

4

The PBGC is the federal corporation that insures certain benefits of
vested participants in DB plans. PBGC insures both single employer and
multiemployer defined benefit plans. As of the end of fiscal year 2004,
PBGC reported an accumulated deficit in its single employer program of
$23.3 billion. See GAO, Private Pensions: Recent Experiences of Large
Defined Benefit Plans Illustrate Weaknesses in Funding Rules , GAO-05-294
(Washington, D.C.: May 31, 2005).

5

For example, see H.R. 2830 and S. 1783.

6

As we noted in past GAO work, plan participants could benefit from
receiving clearer information regarding the conversions they face. See
GAO, Private Pensions: Implications of Conversions to Cash Balance Plans,
GAO/HEHS-00-185 (Washington, D.C.: Sept. 29, 2000) and Cash Balance Plans:
Implications for Retirement Income, GAO/HEHS-00-207 (Washington, D.C.:
Sept. 29, 2000).

7

See GAO, Private Pensions: Implications of Conversions to Cash Balance
Plans, GAO/HEHS-00-185 (Washington, D.C.: Sept. 29, 2000).

discriminatory. 8 Employers report this legal uncertainty has made CB plan
conversions less popular than in the past. In 2000, we reported on the
implications of conversions to CB plans and recommended legislative and
executive agency actions to address the regulatory uncertainty concerning
CB plans and to improve disclosure to affected participants. 9

In response to the problems facing the DB system, committees in both the
House and the Senate have recently proposed legislation that would address
many issues facing defined benefit plans, including the legal uncertainty
regarding the formation of new CB plans or the conversion of traditional
DB plans to CB plans. 10 To help in your deliberations, you asked us to
provide information on the incidence, features, and effects of CB plan
conversions. More specifically, you asked: (1) What does the current
research say about the implications of CB plan conversions for workers'
benefits? (2) What is the prevalence and types of transition provisions
provided to protect workers' benefits in past conversions to CB plans? (3)
How do individual participants fare under a hypothetical conversion to a
typical CB plan compared to the typical FAP plan? On September 1, 2005,
and again on October 12, 2005, we briefed your staff on the results of our
analysis. This report formally conveys the information provided during
those briefings. (See app. I).

To determine the results of current research, we conducted a review of
academic and business literature regarding CB plans and the conversion of
traditional DB plans to CB plans. To identify the prevalence and types of
transition provisions in CB plans, we worked with the 2001 Form 5500 to

8 See Cooper v. IBM Pers. Pension Plan, 274 F.Supp.2d 1010 (S.D. Ill.
2003). Compounding this uncertainty, in September 1999, the Internal
Revenue Service announced that it would begin requiring that applications
for the approval of cash balance formula designs be forwarded to its
headquarters for technical review, resulting in an effective moratorium on
approving conversions to cash balance plans.

9

For more information, see GAO/HEHS-00-185 and GAO, Cash Balance Plans:
Implications for Retirement Income , GAO/HEHS-00-207 (Washington, D.C.:
Sept. 29, 2000).

10

See for example, H.R. 2830 and S.1783.

identify and examine CB plan conversions for their design features. 11 We
first identified all 843 plans with 100 or more participants that
indicated a CB or hybrid plan component on Form 5500. We then selected a
random sample of 205 of these plans. Our sample was comprised of the 45
largest plans (the smallest of which has about 17,500 participants) and a
random sample of 160 other plans. 12 Of these 205 plans, we identified 31
large plans and 102 smaller plans as conversions from traditional DB plans
to CB plans. (For our methodology, see apps. II, III and IV.)

To analyze the effects of a CB plan conversion on individual workers, we
used a pension policy microsimulation model (PENSIM). PENSIM simulates
lifetime retirement benefits for over 100,000 participants in the 1955
birth cohort. We calculated and compared monthly retirement income for
workers from the 1955 birth cohort who are projected to be alive at age
68, and vested in a job covered by a typical FAP plan that is converted to
a CB plan. The model allows comparison of benefits received from CB plans
and ongoing traditional FAP plans, as well as terminated FAP plans.

We conducted four simulations:

     o typical CB plan versus typical FAP plan,
     o typical CB plan versus terminated FAP plan,
     o equal cost CB plan versus typical FAP plan, and
     o equal cost CB plan versus terminated FAP plan.

Plan characteristics for the traditional FAP plan and typical CB plan were
based on Bureau of Labor Statistics' (BLS) employee compensation and
benefit data, our analysis of CB conversions as designated in the 2001

11

The Form 5500 contains considerable information on plan assets,
liabilities, contributions, design features, including whether a plan is a
cash balance plan. Although the Form 5500 provides the most comprehensive
data, its problems are well documented. Our analysis focused on the
features of the CB plan at the time of conversion and thus would not
include information on how these plans might have been amended since that
date. It is possible that some sponsors have amended their plans since the
initial conversion, in light of employee reactions and recent court
decisions. Also, it is possible that some sponsors have changed other
employee benefit plans to help mitigate the potential reduction in some
workers' future benefits resulting from a CB plan conversion, but
determining the nature and extent of such changes was outside the scope of
our work.

12

Estimates based on our random sample of plans are subject to sampling
error. We are 95 percent confident that the true population values are
within +/- 9 percentage points of the estimated percentages.

Form 5500 data base and discussions with industry actuaries and
consultants knowledgeable about CB plans and DB plans generally. We
developed the features of our equal cost CB plan by starting with the
design features of the typical CB plan and then increasing both the base
pay credit and the weighted pay credits (a percentage of pay that
increases as an employee's age and/or years of service increase) until the
cost was equivalent with a traditional FAP plan with a workforce of
identical actuarial, demographic, and labor market characteristics. (See
slides 19 to 25 in app. I.) We conducted our work between September 2004
and September 2005 in accordance with generally accepted government
auditing standards.

In summary, we found the following:

     o The pension and economic literature provides little conclusive
       evidence about the effects on benefits and other aspects of CB plan
       conversions, particularly with regard to why sponsors convert to CB
       plans in the first place. (See slides 9 and 10.) In many cases, data
       and other methodological issues (e.g., sampling methods) limit the
       generalization of results. The effects of a conversion depend on a
       variety of factors including the generosity of the CB plan itself,
       transition provisions that might limit any adverse effects on current
       employees, 13 and firm-specific employee demographics. CB plan
       conversions are posited to have distributional effects on expected
       pension wealth: younger, more mobile workers usually benefit while
       older workers with long job tenure are more likely to experience a
       loss, particularly if they are near the age and service requirements
       for early retirement. Less research is available on the actual benefit
       distributional effects of such conversions, e.g., how participants are
       likely to fare under a CB plan compared to the traditional DB plan
       that is being replaced.
     o Our analysis of plan conversions determined that most conversions
       occurred between 1990 and 1999, and primarily in the manufacturing,
       health care, finance and insurance industries. Most conversions set
       participants' opening account balances equal to the present value of
       their accrued benefits under the previous plan, although the interest

Some firms protected workers against a potential reduction in future
benefits by grandfathering, at the time of conversion, all or some plan
participants. Grandfathering allows eligible participants to continue to
accrue benefits under the prior formula or entails operating both formulas
and providing eligible participants with the greater benefit.
Grandfathering can be implemented in various ways, affecting different
groups of workers.

Page 5 GAO-06-42 Pension Plans

rate used to calculate the balance varied plus or minus 1 percent of the
30-year Treasury bond rate. (See slides 11 to 18.) 14 The use of interest
rates above the 30-year Treasury rate is more likely to result in a
wearaway situation, unless otherwise mitigated. Most plans provided some
form of transition provisions to mitigate the potential adverse effects of
a conversion on workers' expected benefits for at least some employees.
About 47 percent of all conversions used some form of grandfathering that
was applied to at least some of the employees in the former traditional DB
plan. 15 In most cases, grandfathering eligibility was limited to
employees meeting a specified minimum age or years of service or both.
Most conversions also used some form of ongoing weighted pay credit.

o  Our comparison of a typical FAP plan that is converted to a typical CB
plan finds that, regardless of a worker's age, more workers would have
received greater benefits under the FAP than under the typical CB plan. 16
(See slides 26 to 28.) For workers who receive less under the CB
conversion, median benefit decreases range from $59 per month at age 30 to
$238 per month at age 50. For the workers who receive more under the
conversion, median benefit increases range from $15 per month at age 30 to
$27 per month at age 50. 17 Those who experience either benefit increases
or decreases are more likely to be men, except for those at the age 50
conversion, where they are more likely to be women.

14

IRC section 417(e)(3) stipulates that DB sponsors that permit lump sum
distributions must, among other conditions, calculate distributions to
departing participants using an interest rate no greater than 30-year
Treasury rate. Using a higher interest rate would result in a lower lump
sum distribution.

15

There is a range of types of grandfathering that can be used by plan
sponsors. They can include provisions such as giving employees a choice of
whether to stay in the old FAP plan or join the new CB plan, providing a
minimum benefit where the employee is guaranteed to at least earn the
benefit of their former plan until a future specified date, or making
grandfathering available to only some or all employees in the former plan.

16

These comparisons are based on amounts of annuity benefits and do not take
into account death benefit coverage before an annuity begins. For the
purpose of this report, it is recognized that participants do not have an
entitlement to future or expected benefits.

17

We also conducted a comparison of lifetime benefits for workers under a
traditional FAP and those converted to a typical CB plan as well as to an
equal cost CB plan. In these cases, while the number of workers faring
better under the CB plans is greater at each age compared to the numbers
in the monthly benefits calculation, the basic results found under the
monthly benefit comparison are not changed in either case. (See slides
31-32, 38-39.)

     o In comparing a conversion to a typical CB plan with a terminated FAP,
       all vested workers would do better under the CB plan. Median monthly
       benefits increase at conversion ages 30, 40, and 50, with increases
       ranging from $150 per month for conversions at age 30 to $305 per
       month for conversions at age 50. (See slides 29 to 30.) The increase
       in benefits for older workers is because grandfathered benefits are
       included in these results. Although the analysis focuses on vested
       workers at the time of conversion, under a terminated FAP plan, by law
       all previously unvested workers (those with less than 5 years service)
       are immediately vested. 18
     o Under a traditional FAP plan conversion to an equal cost CB plan,
       larger numbers of workers at all ages have benefit increases than
       under the typical CB plan/FAP plan scenario. 19 (See slides 33 to 35.)
       Grandfathering again protects the benefits of those older workers who
       were covered. However, while more workers who are converted at age 30
       fare better under the CB plan, this was not true at other ages. A key
       factor is the greater generosity of the equal cost CB plan compared to
       the typical CB plan we also simulated. 20 Under the equal cost
       scenario, median reductions range from $75 per month for conversions
       at age 30 to $128 per month for conversions at age 50, while median
       increases range from $90 per month for conversions at age 30 to $29
       per month for conversions at age 50. For all conversion ages, those
       with longer job tenure and who are not covered by grandfathering
       protections are more likely to lose than those workers with shorter
       tenure. At each conversion age, a greater percentage of those who are
       more likely to experience benefit increases are men rather than women.
     o In comparing a conversion to an equal cost CB plan with a terminated
       FAP plan, again all vested workers do better under the CB plan. Median
       increases range from $283 per month for conversions at age 30 to $396
       per month for conversions at age 50. (See slides 36 to 37.) The
       increase in benefits for older workers comes about because
       grandfathered benefits are included in these results.

18

In our simulations, about 36 percent of our sampled individuals (57,049)
who participated in at least one private sector FAP or CB plan never
vested in such plans.

19

Again, these comparisons are based on amounts of annuity benefits and do
not take into account death benefit coverage before an annuity begins.

20

This plan's pay credits were more generous than virtually all of the 136
plan conversions we analyzed.

Page 7 GAO-06-42 Pension Plans

Concluding Observations

Our analysis illustrates one of the difficult choices facing the Congress
in crafting comprehensive DB pension reform legislation, including the
controversial issues surrounding the legal status of CB plans, and
particularly CB conversions. The current confusion concerning CB plans is
largely a consequence of the present mismatch between the ongoing
developments in pension plan design and a regulatory framework that has
failed to adapt to these designs. Although CB plans legally are DB plans,
they do not fit neatly within the existing regulatory structure governing
DB plans. This mismatch has resulted in considerable regulatory
uncertainty for employers as well as litigation with potentially
significant financial liabilities. For many workers, this mismatch has
raised questions about the confidence they may have in the level of income
they expect at retirement, confidence that has already been shaken by the
termination of large pension plans by some bankrupt employers. 21

CB plans may provide more understandable benefits and larger accruals to
workers earlier in their careers, advantages that may be appealing to a
mobile workforce. However, conversions of traditional FAP plans to CB
plans redistribute benefits among groups of workers and can result in
benefits for workers, particularly those who are longer tenured, that fall
short of those anticipated under the prior FAP plan. Our simulations
suggest that grandfathering plan participants who are being converted can
protect those workers' expected benefits, and, in fact, such protections,
in some form, are fairly common in conversions. Our simulations also show
that without such mitigation, many workers can receive less than their
expected benefits when converted from a traditional FAP plan, even in
cases where the CB plan is of equal cost to the FAP plan it is replacing.
As a result, as we noted in our 2000 report, 22 additional protections are
needed to address the potential adverse outcomes stemming from the
conversion to CB plans. For example, requirements for setting opening
account balances could protect plan participants, especially older
workers, from experiencing periods of no new pension accruals after
conversion while other workers continue to earn benefits.

Our simulated comparison of CB plans with the termination of a FAP plan
leads to several important observations. First, the immediate vesting of
all unvested workers requirement in a plan termination actually leads to a
greater number of workers getting some retirement benefits and highlights

21 See GAO-05-294. 22 See GAO/HEHS-00-185.

Page 8 GAO-06-42 Pension Plans

                                Agency Comments

the portability limitation of DB plans. Workers in an ongoing DB plan only
receive benefits if they are vested. Appealing to a mobile workforce would
seem to place an even greater significance on pension portability. Yet
even CB plans, which often feature lump sum provisions in their design, do
not address this issue because they typically have similar vesting
requirements as traditional FAPs.

In our simulations, vested workers under either a typical or equal cost CB
plan still fare better than if the FAP plan is terminated. We note further
that some sponsors of CB plans have already exited the DB system, a system
that has been declining in sponsorship and participation for several
decades now. There is a crucial balance between protecting workers'
benefit expectations with unduly burdensome requirements that could
exacerbate the exodus of plan sponsors from the DB system. Congress, as it
grapples with the broader components of pension reform, has the
opportunity not only to protect the benefits promised to millions of
workers and eliminate the legal uncertainty surrounding CB plans that
employers face, but also to craft balanced reforms that could stabilize
and possibly permit the long-term revival of the DB system.

We provided a draft of this report to the departments of Labor, Treasury,
and the PBGC. No written comments were provided by these agencies. They
did, however, provide technical comments, which we incorporated as
appropriate.

We plan to provide copies of this report to the Secretaries of the
Department of Labor and the Department of Treasury and to the Pension
Benefit Guaranty Corporation and interested congressional offices. We will
make copies available to others upon request. In addition, the report will
be available at no charge on the GAO Web site at http://www.gao.gov.

If you have any questions concerning this request please contact me at
(202) 512-5932. Other major contributors to the report are listed in
appendix VI.

Barbara D. Bovbjerg

Director, Education, Workforce and Income Security Issues

Appendix I: Information on Cash Balance Pension Plans

                                   Objectives

I. Literature Review: Evaluate current research on the implications of
cash balance (CB) plan conversions

II. Plan Analysis: Review CB plans for the prevalence and types of
transition provisions provided to protect workers' benefits when
converting to CB plans

III. Simulations: Analyze how participants may fare under hypothetical CB
plan conversions compared to the typical final average pay (FAP) plan and
to a terminated FAP plan.

                                   Background

        * CB pension plans:
             o Are a type of hybrid defined benefit (DB) plan that expresses
               benefits as a hypothetical account balance based on pay,
               service, and interest credits.
             o Are classified as DB plans because participants' benefits are
               determined by a benefit formula.
     o FAP plans are a type of DB plan where participants' benefits are
       derived from a formula that is based, in part, on the employee's final
       average pay.

                               Background, cont.

     o Some conversions to CB plans have been controversial because of the
       effect they may have on pension benefits of workers with different
       ages and years of service. At the same time, CB plans have been noted
       for providing lump sum benefits that can be rolled over upon
       separation and providing benefit accruals based on pay and length of
       service.
     o Wearaway periods: CB plan conversions can sometimes result in
       situations where some workers do not earn additional pension benefits
       while other workers continue to do so.
          * Wearaway can occur for a variety of reasons. Examples of when
            wearaway can occur are:
               o At conversion when a participant's hypothetical opening
                 account balance is set at less than the present value of the
                 prior accrued benefits (the level of benefits received if
                 paid out as a lump sum).
               o After conversion because of a fall in the federally mandated
                 discount rate used to determine a lump sum amount.
               o In relation to annuity benefits earned as of conversion. It
                 is dependent on the the form of annuity selected by the
                 participant and the design of early retirement benefits in
                 the prior plan's formula.

                               Background, cont.

     o During wearaway, pay and interest credits do not represent new benefit
       accruals until the CB account exceeds the value of benefits that could
       be paid under the old plan.
     o Wear-away periods tend to be longer for older workers.
     o Status of CB plans has been questioned after a court's ruling that at
       least one CB plan is age discriminatory. In late 1999, the Treasury
       Department stopped issuing IRS determination letters approving CB plan
       conversions.
     o Proposed pension reform legislation includes provisions that could
       clarify some legal issues concerning CB plans.
     o Some analysts believe that CB plans represent a potential opportunity
       to stem the decline or even revitalize the declining DB system.

                                  Methodology

I. Conducted review of academic and business literature.

II. Analyzed Form 55001 information and attachments from 2003 and earlier
years capturing design features of CB plan conversions at the point of the
initial conversion. Initial conversions from a traditional DB plan to a CB
plan with most covering the period from early 1990s to 2003 with a few
plan conversions in the mid 1980s. Subsequent changes to CB plans' design
were not part of the analysis nor were changes made to other plan
benefits.

     o Identified 843 plans with 100+ participants that indicated CB on Form
       5500
     o Selected the 45 largest plans (1.8 million participants) and a random
       sample of 160 other plans
     o Of these 205 plans, 31 large plans and 102 smaller plans met criteria
       as conversions

1The Form 5500 Report, which is completed and filed by the Plan sponsor,
is the primary source of information for both the federal government and
the private sector regarding the operation, funding, assets, and
investments of private pension plans and other employee benefit plans. The
Form 5500 does not provide enough detail to determine the number of
participants affected by a conversion.

                               Methodology, cont.

o  Estimates are based on a random sample of plans, so slightly different
estimates could result from a different random sample. We are 95%
confident that the true population values are within +/-9 percentage
points of the estimate percentages based on our sample.

III. Simulate effects of a conversion to a CB plan and other scenarios -
Used a pension policy micro-simulation model (PENSIM).

        * Model simulates lifetime retirement benefits for over 100,000
          participants in the 1955 birth cohort. Lifetime and monthly
          retirement income is analyzed for those who are:
             o projected to be alive at age 68, and
             o vested in a job covered by a typical FAP plan that is
               converted to a CB plan (typical or equal cost).
     o Model allows comparison of benefits received from CB plans, ongoing
       traditional final average pay plans, and terminated FAPs.
     o See appendixes II, III, and IV for a discussion of our methodology.

                              Summary of Findings

I. Literature provides few generalizable conclusions, particularly with
regard to:

     o why sponsors convert to CB plans
     o the benefit distributional effects of such conversions.

II. Analyzed plan conversions show most, but not all:

     o converted accrued benefits into an opening account balance and offered
       some form of transition provisions.
     o had age and service eligibility restrictions on transition provisions.

                           Summary of Findings, cont.

III. Regardless of age, workers who were converted from an FAP plan to a
typical CB plan generally had reductions from expected FAP benefits. A
majority of younger workers received larger benefits under a conversion to
an equal cost CB plan.

     o Analysis of lifetime benefits under a conversion to an equal cost CB
       plan does not change basic findings.
     o Vested workers receive larger benefits under a CB conversion of either
       type compared to benefits received under termination of an FAP.

I. Literature Review

      Research Provides Limited Evidence to Generalize About CB
      Conversions

ta and othe thodolog issues

o   o DDaata and othethodolog issues (e.g., sampling methods) limit
generaizaon of results.

limit genera iza on of results.

     o Conversion impact depends on a variety of factors including plan
       generosity, transition provisions, and firm specific employee
       demographics.
     o Also, because of the different accrual patterns in a CB plan compared
       to a FAP plan, for a variety of workers, the impact of a conversion
       varies.

I. Literature Review

    Research Provides Limited Evidence to Generalize About CB Conversions,
    cont.

ent search prov ed evidence as t

o   o CCuurrrrentsearch proved evidence as t

y sponso to plans.

        * WWhhy sponso to plans.
        * par pant kel fa under
     o HHooww parpantkel fa under

re tive to the tra ona DB that is ing repl

retive to the traona DB that ising repl

o  CB plan conversions have distributional effects on pension wealth:

     o Younger, more mobile workers who vest usually benefit.
     o Older workers with long job tenure likely to experience a loss,
       particularly if they are near age and service requirements for early
       retirement.

Note: About 36 percent of our sampled individuals (57,049) who
participated in at least one private sector FAP or CB plan never vested in
such plans.

         Methods for Determining Opening Account Balances at Conversion

        * There were 2 primary methods for setting the opening account
          balance:
            formula that may include factors
       such as interest rates, employer-added incentives, early retirement
       benefits, & other assumptions.

Conversion Interest Rates and Transition Provisions Are Key Factors in Wearaway

     o 23 of 39 plans with data available used conversion interest rates
       within 1% of the prior month's 30-year Treasury rate
     o Wearaway may occur when a participant's hypothetical opening account
       balance is set at less than the present value of its accrued benefits
       using 30-year Treasury rate, as specified under the Internal Revenue
       Code.
     o Transition provisions (e.g., grandfathering, transition pay credits)
       are important factors in mitigating wearaway.

o  Grandfathering prevents wearaway for participants who continue to
accrue benefits under the prior plan formula.

            About Half of Plans Offered Some Form of Grandfathering

o  Grandfathering was offered in 47% of all conversions and in 55% of the
largest converted plans, although most of these provisions had some form
of age or service restrictions.

        * Eligibility requirements in plans offering grandfathering included:
             o age plus service
             o all employees
             o age or service
     o Age plus service was the method most often used.

II. Plan Analysis

               Most Conversions Used Weighted Ongoing Pay Credits

     o 62% of all conversions used some form of weighted pay credits (those
       that increase based on the participant's age and/or service).
     o 36% of all conversions used level pay credits (those that are a level
       function of salary).
     o About 42% of large conversions used an age plus service method for
       providing ongoing pay credits.
     o Weighted pay credits tend to benefit older and longertenured workers
       relatively more than level pay credits.

                                III. Simulations

                        Simulations of Plan Conversions

        * Compare monthly and lifetime retirement income for workers from the
          1955 birth cohort who were converted at different ages to a CB plan
          and were either:
             o Not vested in a typical, traditional FAP plan at the time of
               conversion but stay on the job and later vest; or
             o vested at the time of conversion in a typical, traditional FAP
               plan;
        * 4 Simulations:
             o Typical CB plan vs. typical FAP plan
             o Typical CB plan vs. terminated FAP plan
             o Equal cost CB plan vs. typical FAP plan
             o Equal cost CB plan vs. terminated FAP plan

                                       19

                                III. Simulations

              Plan Characteristics: Typical Final Average Pay Plan

     o Immediate eligibility and 5-year cliff vesting and normal retirement
       age 65, early retirement age 55 with 10 years of service with early
       retirement benefit reduction of 5 percent per year.
     o Immediate disability retirement benefits for those vested, no
       survivors benefits or joint-and-survivor annuities.
     o Benefits paid as a nominal annuity (i.e., no benefit COLA).
     o Terminal earnings (final pay) is final five-year average.
     o Benefits formula is excess integrated with base rate of 1.5 percent of
       final pay per year of service and has a rate of 0.45 percent of final
       pay per year of service for those amounts in excess of the social
       security maximum.
     o Typical FAP plan design based on prior GAO reports, literature
       reviews, and discussions with pension actuaries, consultants
       knowledgeable about DB plans.

                                III. Simulations

                Plan Characteristics: Typical Cash- Balance Plan

     o Immediate eligibility and five-year cliff vesting; base pay credit of
       3.0 percent of salary for employee with age-plus-service (APS) =< 35.
     o Pay credit rises gradually until it is 6.0 percentage points above the
       base pay credit for employee with APS >= 70.
     o Cash-balance account crediting rate is the Treasury rate.
     o Employee rolls over account balance at separation and earns Treasury
       rate. Balances converted to nominal single-life annuity at retirement
       using the Treasury rate and the GAM 83 mortality table adjusted to the
       pertinent year.
     o Typical CB plan design is based on plans analyzed in GAO's Form 5500
       data, and confirmed by pension actuaries, consultants knowledgeable
       about CB plans.
     o Some typical CB plan design features may have changed in light of
       recent court decisions and congressional interest

III. Simulations

               Plan Characteristics: Equal Cost Cash Balance Plan

        * Same assumptions as the typical CB plan except:
             o Base pay credit of 7.35 percent of salary for employee with
               age-plus-service (APS) =< 35.
             o Pay credit rises gradually until it is 6.0 percentage points
               above the base pay credit for employee with APS >= 70.
        * Equal cost CB plan used for our simulations is:
             o More generous pay credits than virtually all plans in our Form
               5500 analysis
             o More generous than those specified in pension research
             o Though not explicitly modeled, to some extent, our equal cost
               cash balance plan could be considered to implicitly include
               other enhancements made by employers to other benefits, such
               as those provided by a DC plan, for example.

                                III. Simulations

              Plan Characteristics: Cash-Balance Plan Conversions

     o Opening cash balance equal to the present value of accrued finalpay
       benefit at plan conversion date. Discount rate is the 30-year Treasury
       rate. Mortality table is GAM 83 projected for mortality improvements
       to the pertinent year.
     o Employees who meet an age-plus-service (APS) >= 60 eligibility
       requirement at plan conversion date are grandfathered under the FAP
       plan and receive benefits according to that plan's provisions.
          * Treatment of early retirement benefits:
               o The FAP plan considered in this report has a modest
                 earlyretirement subsidy: benefits are reduced by 5 percent
                 for each year benefits are claimed before age 65.
               o Federal anti-cutback rules are simulated correctly in that
                 when a FAP plan is converted or terminated, employees who
                 remain with the firm until early retirement age are eligible
                 for early retirement benefits under the old plan.

                                III. Simulations

                   Plan Characteristics: Terminated FAP Plan

        * Same assumptions as the typical FAP plan and:
             o Terminated FAP plan has immediate cessation of additional
               benefit accrual.
             o Current law on plan terminations requires "immediate vesting"
               for "non-vested" workers regardless of years of service. This
               results in previously ineligible workers now receiving a small
               benefit.
     o Analysis focuses on "vested" workers only --those with at least five
       years service.
     o FAP plan and termination scenarios provide benchmark range of possible
       comparisons, including plan freezes

                                III. Simulations

        Monthly Retirement Income Results vs. Lifetime Benefits Results

     o Results are shown in terms of present value of lifetime benefits for
       those alive at age 68 and monthly retirement income for those alive at
       age 68. Age 68 is the age when the largest number of individuals are
       retired and alive in our sample.
          * Monthly benefit and lifetime benefit comparisons for those alive
            at age 68 will have slightly different results:
               o For example, vested workers under CB plans who typically
                 separate earlier in their careers may start benefits at a
                 different age compared to similar workers who separate from
                 an FAP plan.
               o Thus, the present value of lifetime benefits paid to these
                 workers under CB plans may be distributed over a different
                 time period than for similar workers under FAP plans. So
                 monthlybenefits may be slightly different.

                                III. Simulations

Comparisons of Median Monthly Retirement Income: Typical CB Plan vs. Typical FAP
                                      Plan

o  Regardless of age at conversion, more workers who are converted from a
FAP plan to the typical CB plan experience benefit reductions. (See figure
5).

     o Key factor is lack of generosity of the typical CB plan.
     o Grandfathering protects those workers who meet eligibility

requirements.

o  For those not grandfathered, at conversion ages 30, 40, and 50: (See

figure 6)

        * Reductions in median monthly income range from $59 for conversions
        * at age 30 to $238 for conversions at age 50.
     o Increases range from $15 per month for conversions at age 30 to $27

per month for conversions at age 50.

     o Those who benefit as well as those who lose from conversion at ages 30
       and 40 are more likely to be men and at age 50 are more likely to be
       women.
     o At all conversion ages, those experiencing greater benefits from
       conversion are generally more highly educated and have higher incomes.

                                III. Simulations

Comparisons of Median Monthly Retirement Income: Typical CB Plan Conversion vs.
                          Terminated Typical FAP Plan

     o Regardless of age, all vested workers who converted to a typical CB
       plan experienced monthly benefit increases compared to a terminated
       FAP plan.
     o At conversion ages 30, 40, and 50, increases range from $150 per month
       for conversions at age 30 to $305 per month at age 50. Grandfathered
       benefits for those eligible under the CB plan greatly impact results
       shown for older workers.(See figure 7.)

o  Terminated plan benefits are shown for only those participants who were
vested in the typical CB plan.

                                III. Simulations

Workers Converted to Typical CB Plan from Typical FAP at Earlier Ages Generally
                       Receive Reduced Lifetime Benefits

o  Comparison of lifetime benefits for typical CB plan and typical FAP
plan does not change basic findings from monthly benefit comparisons.

     o Regardless of age at conversion, more workers who are converted from a
       FAP plan to the typical CB plan have lower present value of lifetime
       benefits. (See figure 8.)
     o Nearly half of workers experiencing a conversion at age 50 are
       grandfathered in their FAP benefit.

                                III. Simulations

Grandfathering Protects Eligible Older Workers' Monthly Benefits When Converted
                to an Equal Cost CB Plan from a Typical FAP Plan

     o Grandfathering protects eligible older workers' benefits converted to
       an equal cost CB Plan from a FAP Plan (See figure 9.)
          * More workers who converted from a FAP plan to an equal cost CB at
            age 30 generally experience monthly benefit increases
                  * Increases range from $90 per month for conversions at age
                    30
                  * to $29 per month for conversions at age 50. (See figure
                    10.)
               o Reductions range from $75 per month for conversions at age

30 to $128 per month for conversions at age 50.

o  For all conversion ages, those with a longer job tenure and who are

not covered by grandfathering protections are more likely to

experience lower benefits than those with shorter tenure

                                III. Simulations

    Comparisons of Median Monthly Retirement Income: Equal Cost CB Plan vs.
                          Terminated Typical FAP Plan

     o Regardless of age, all vested workers who converted to an equal cost
       CB plan experience benefit gains compared to a terminated FAP.
     o Median increases range from $283 per month for conversions at age 30
       to $396 per month for conversions at age 50. Grandfathered benefits
       for older workers under the CB greatly impact results.(See figure 11.)

o  Terminated plan benefits are shown for only those participants who were
vested in the equal cost CB plan.

                                III. Simulations

Workers Converted to Equal Cost CB Plan from Typical FAP at Age 30 Receive
                           Greater Lifetime Benefits

o  Comparison of lifetime benefits for equal cost CB plan and typical FAP
plan consistent with basic findings from monthly benefit comparisons (See
figure 12).

     o More workers converted to an equal cost CB plan from a typical FAP at
       age 30 receive greater present value of lifetime benefits through
       conversion than would at later conversion ages.
     o Nearly half of workers experiencing a conversion at age 50 are
       grandfathered in their FAP benefit, while a significant number (41%)
       of unprotected workers converted at age 50 experience a lower present
       value of lifetime benefits.
     o Outside of grandfather protections, results show a redistribution of
       benefits from older workers to younger workers.

Appendix II: Review of Literature on Cash Balance Plans

GAO compiled a comprehensive list of the academic literature on CB pension
plans since our last reports on the subject issued in 2000, 1 focusing on
those studies that contained original and material empirical work on the
issue. After constructing a list of the relevant literature, we eliminated
partial or incomplete studies, those that did not contain material
empirical work and those that exhibited serious methodological concerns.
We then conducted a more detailed review of the remaining studies,
including several surveys of CB plans. The review concentrated on the
studies' findings and on the methodological issues that may limit
conclusions that can be reached. There is a list of the studies and
surveys reviewed for this report at the end of this appendix. 2

Although there are academic studies that attempt to go beyond anecdotal
information, the literature remains in its infancy. Data and other
methodological issues often limit the conclusions that the empirical
studies examining the impact of plan conversions can reach and, the
ability to generalize their results. In general, the results of all
studies are sensitive to assumptions regarding earnings growth, interest
rates, investment returns, and turnover rates. Because some specifics of
the simulations presented in some studies do not include sufficient
detail, it is difficult to evaluate the quality of the estimates in some
cases. 3

1

See GAO, Cash Balance Plans: Implications for Retirement Income,
GAO/HEHS-00-207 (Washington, D.C.: Sept. 29, 2000) and GAO, Private
Pensions: Implications of Conversions to Cash Balance Plans,
GAO-HEHS-00-185 (Washington, D.C.: Sept. 29, 2000).

2

Since we focused on empirical literature produced since 2000, we did not
include one older study that is cited in the literature in our detailed
review (Kopp and Sher, "A Benefit Value Comparison of a Cash Balance Plan
with a Traditional Average Pay Defined Benefit Plan," The Pension Forum
[Society of Actuaries, October 1998]). The study contains data and other
methodological limitations, as well as making similar conclusions. For
example, because the study examines hypothetical rather than actual plan
conversions, it is not clear that the results extend to the broader
workforce. Additional limitations include that fact that the authors had
limited wage information and therefore relied on simple wage assumptions
rather than actual wage histories and did not test the sensitivity of the
results to the assumptions made regarding key variables.

3

See for example, Watson Wyatt Worldwide, The Unfolding of a Predictable
Surprise: A Comprehensive Analysis of the Shift from Traditional to Hybrid
Plans (2000) and Robert Clark, and Sylvester Schieber, "The Transition to
Hybrid Pension Plans in the United States: An Empirical Analysis," Private
Pensions and Public Policies, eds. W. Gale et al. (Washington, D.C.:
Brookings Institution, 2004).

Lack of Available Data Limits Empirical Studies

Because of the limited availability of data on actual conversions and on
the workforce associated with a particular conversion, few empirical
studies have the ability to examine actual conversions. 4 Because a range
of factors that are unique to each conversion influence the final impact
on workers-including demographic characteristics, the transition benefits
offered during the conversion and the generosity of the new CB plan
relative to the old plan it is replacing-it is difficult to extend the
results of the literature to the actual experience of workers. For
example, in the conversion to a new plan, a sponsor may eliminate early
retirement subsidies-a significant reason why older workers may receive
lower benefits. Similarly, some employers may offer transition benefits
that can help to ameliorate the adverse effects of plan changes on the
more senior segment of the workforce, while others do not. Other studies
focus on "hypothetical" or "prototypical" workers instead of actual
employees and therefore cannot make definitive statements about many
segments of the population or actual workers in the plans analyzed. 5

In addition, the majority of the research simulates the effects of plan
conversions on the workforce assuming that the conversion is cost neutral
(the cost of the new CB plan is equal to the cost of the old DB plan so
that overall pension benefits remain constant). However, some research
suggests that the retirement benefit implications due to a shift to a less
generous CB plan differ materially from the effects of a cost-neutral
conversion. 6 Moreover, several studies were limited to plans that include
transition benefits that often ensure that existing workers do not suffer
significant losses in pension wealth during plan conversions and exclude

4

Virtually all researchers studying this issue, including GAO, have
suffered these data limitations.

5

The use of hypothetical workers is also a limitation of prior GAO reports
on cash balance plans. See GAO-02-207 and GAO-00-185.

6

See, for example, Watson Wyatt Worldwide (2000). In simulating the effects
of one conversion to a cost-reducing cash balance plan, the authors find
that the majority of workers receive lower benefits. However, another
simulation of a shift to a cost neutral plan finds that the majority of
the workers receive higher benefits and, although the losses are
disproportionately borne by the older workers, they are lower than the
losses experienced in the cost-reducing case.

Small Sample Size Limits Survey Reports

pension wealth on previous jobs. 7 Thus, their inclusion/omission may lead
to a bias in the empirical findings either in favor or against CB plan
designs.

Some studies examine only a few plan conversions or rely on assumptions
based upon information extracted from the limited surveys discussed below.
Since the plans analyzed may not be representative, the outcomes may not
generalize to the typical CB conversion or related to the broader
workforce. 8

A few widely cited studies which use survey data in an attempt to
determine the reasons why employers initiate CB plan conversions contain
methodological limitations and base their conclusions on employers' self
perceptions along with additional biases, and cannot be extended beyond
the small samples of firms studied. 9 For example, one study is limited by
a low response rate (20 percent) and insufficient information about the
population and sampling method, survey instrument and its development,
while the others raise concerns over the potential for sample bias and/or
the additional bias due to the fact that over half of the plans evaluated
were those for which the researchers were the primary design

7

This pertains to the majority of the literature we reviewed. Although
Johnson and Uccello

(R.W. Johnson, and C. Uccello, "Cash Balance Plans and the Distribution of
Pension Wealth," Industrial Relations, 42(4) [2003], 745-773) include
pension wealth on previous jobs, analyze actual workers and capture a
greater diversity of outcomes, the results do not generalize to
cost-reducing plan conversions or conversions where the defined benefit
plan incorporated early retirement incentives (see below for more on the
cost neutral assumption). Moreover, the pension wealth in cash balance
plans may be exaggerated because of issues with the data and the
assumptions regarding turnover rates.

8

For example, see several studies conducted by Schieber (including Clark
and Schieber [2004]) which are derived from data on 77 plans collected and
analyzed initially by Watson Wyatt Worldwide in their 2000 study. Given
estimates of the number of actual cash balance conversions and their
growth since 2000, it is not clear that this work can be used as a
reliable guide to gauging the impact of a typical cash balance conversion
on workers. Moreover, GAO found other research studies that were based on
significantly fewer cash balance conversions, e.g., Clark and F.W.
Munzenmaier (R. Clark, and F.W. Munzenmaier. "Impact of Replacing a
Defined Benefit Pension with a Defined Contribution Plan or a Cash Balance
Plan." North American Actuarial Journal, 5 (1). (2003-4): 32-56. (2001).

9

See, for example, Watson Wyatt Worldwide (2000); PriceWaterhouseCoopers,
Survey of Conversions from Traditional Pension Plans to Cash Balance Plans
(2000); and Mellon Financial Corporation, 2004 Survey of Cash Balance
Plans (Secaucus, N.J.: 2004).

Studies and Surveys Reviewed

consultants. 10 In general, we determined that the results from these
surveys may not be representative of the population of CB plan conversions
and methodological limitations suggest that the results should be
interpreted with caution.

Clark, Robert. "Pension Plan Options: Preferences, Choices and the
Distribution of Benefits." Pension Research Council Working Paper, PRC WP.
2003-24.

Clark, Robert, and Fred W. Munzenmaier. "Impact of Replacing a Defined
Benefit Pension with a Defined Contribution Plan or a Cash Balance Plan."
North American Actuarial Journal, 5 (1). (2003-4): 32-56.

Clark, Robert, and Sylvester Schieber. "The Transition to Hybrid Pension
Plans in the United States: An Empirical Analysis." Private Pensions and
Public Policies, eds. W. Gale et al. Washington, D.C.: Brookings
Institution. 2004.

Coronado, Julia, and Philip Copeland. "Cash Balance Pension Plan
Conversions and the New Economy." Federal Reserve Board Working Paper.
November 2003.

D'Souza, Julia, John Jacob, and Barbara Lougee. Why Do Firms Convert to
Cash Balance Pension Plans? An Empirical Investigation. Cornell
University, December 2004.

Johnson, R.W., and C. Uccello. "Cash Balance Plans and the Distribution of
Pension Wealth." Industrial Relations, 42(4) (2003): 745-773.

Mellon Financial Corporation. 2004 Survey of Cash Balance Plans. Secaucus,
N.J.: 2004.

Niehaus, Greg, and Tong Yu. "Cash-Balance Plan Conversions: Evidence on
Excise Taxes and Implicit Contracts." The Journal of Risk and Insurance,
72(2). 2005.

PriceWaterhouseCoopers. Survey of Conversions from Traditional Pension
Plans to Cash Balance Plans. July 2000.

10

PriceWaterhouseCoopers (2000).

Purcell, Patrick. Pension Issues: Cash Balance Plans. Washington, D.C.:
Congressional Research Service, August 2003.

Rao, A., L. Higgins, and S. Taylor. "Cash Balance Pension Plans: Good News
or Bad News." Journal of Applied Business Research, 18 (3). 2002.

Samwick, Andrew, and Jonathan Skinner. "How will 401(k) Plans Affect
Retirement Income?." American Economic Review, Vol. 94, No.1. March 2004.
11

Schieber, Sylvester. "The Shift to Hybrid Pensions by U.S. Employers: An
Empirical Analysis of Actual Plan Conversions." Pension Research Council
Working Paper, PRC WP. 2003-23.

Watson Wyatt Worldwide. The Unfolding of a Predictable Surprise: A
Comprehensive Analysis of the Shift from Traditional to Hybrid Plans.
2000. 12

11

Findings of this paper are not directly discussed in this appendix as it
involves an assessment of plan conversions from traditional defined
benefit plans to defined contribution plans.

12

Looking at 78 plan conversions, Watson Wyatt Worldwide (2000) found that
56.4 percent of the plans were cost-reducing, 20.5 percent were
cost-neutral and 23.1 were costincreasing. However, when the authors
assumed workers took full advantage of the enhancements to the defined
contribution plan that occurred contemporaneously with the transition,
44.9 percent of the plans were cost reducing, 17.9 percent were cost
neutral and

37.2 percent were cost increasing. This work has led some researchers to
deduce that the average cash balance conversion is cost neutral, since the
majority of the plans (55.1 percent) were cost-neutral or increasing.
However, as we indicated earlier, it is not clear that this small sample
of conversions is representative. Also, some recent statistics do not
support the assumption of full participation used by Watson Wyatt to
incorporate these enhancements. For example, some estimates suggest that a
significant percentage of employees do not participate in their 401(k)
program at all, and the majority of those that do participate do not
maximize the value of the plan. See Congressional Research Service,
Automatic Enrollment in Section 401(k) Plans (Washington, D.C., Oct. 14,
2004). The CRS found that because enrollment in most S:401(k) plans is
voluntary, not all workers whose employers offer a plan choose to
participate. `The Bureau of Labor Statistics reports that in 2003, 51
percent of workers in the private sector were employed at establishments
that offered a defined contribution plan, but just 40 percent of employees
at private establishments participated in a plan. Consequently, the
participation rate among employees whose employer offered a DC plan was 78
percent." Also see Alicia H. Munnell and Annika Sunden, Coming Up Short:
The Challenge of 401(k) Plans (Washington, D.C.: Brookings Institution
Press, 2004).

The authors conclude that one in four employees do not participate in a
401(k) plan, and less than 10 percent contribute the maximum.

Appendix III: Analysis of Form 5500 Data on Cash Balance Plans

                          Sample of Cash Balance Plans

To obtain information about CB plan conversions, we reviewed 2001 Form
5500 data for a random sample of CB plans. We drew this sample from the
population of plan sponsors that indicated on their Form 5500 that they
sponsored a CB plan. 1 The study population consisted of all CB plans as
of 2001 having at least 100 active participants, supplemented with an
additional 96 CB plans that were identified by PBGC based on 2002 and 2003
data not yet available to the GAO. For the purpose of this report, we
excluded plans having fewer than 100 participants in order to focus on the
plans with the greatest number of participants. 2 This resulted in a total
of 843 plans in our study population.

We used the Form 5500 as our primary source of information for analyzing
the prevalence of transition provisions used by plan sponsors when they
converted to a CB plan because it was a cost effective way of obtaining
conversion information for a large number of plans. It would have been
optimal to obtain summary plan descriptions (SPD) from plan sponsors.
However, since plan sponsors are no longer required to file SPDs, direct
contact with such a large number of plan sponsors would have been cost
prohibitive. 3

Although it is the most comprehensive pension data available, using Form
5500 data also presented limitations and weaknesses. We had limited
ability to determine the full scope of conversions beyond tax year 2001
since this was the most current and complete 5500 data publicly available
from the Department of Labor (Labor) when we began our analysis. In
addition, we also had difficulty obtaining Form 5500 filings for some
years, particularly from the early 1990s and before. As previously
reported by GAO, 4 statutory reporting requirements, processing issues,
and current

1

The Form 5500 is a disclosure form that private sector employers with
qualified pension plans are required to file with the Internal Revenue
Service (IRS), Labor's Employee Benefit Security Administration (EBSA),
and Pension Benefit Guaranty Corporation (PBGC). This dataset contains all
private sector single employer DB plans that are insured by the PBGC.

2

There were 1590 plans of any participant size that indicated they were
cash balance plans in the Form 5500.

3

Effective August 5, 1997, with the passage of the Taxpayer Relief Act of
1997, plan sponsors were no longer required to file summary plan
descriptions or related documents with the Department of Labor. Instead,
plans are required to furnish this information only upon request.

4

See GAO, Private Pensions: Government Actions Could Improve the Timeliness
and Content of Form 5500 Pension Information , GAO-05-491 (Washington,
D.C.: June 2005).

                                 Sample Design

Labor practices affect the timeliness of the release of available Form
5500 information, in some cases, resulting in a 3-year lag between data
reporting and its release. In addition, information provided on the form
and attachments proved, in some instances, to be inconsistent from one
plan sponsor to another. This inconsistency hampered our data collection
efforts, and subsequently, we were unable to provide meaningful results on
all of the information our data collection instrument was designed to
capture. For example, we found that not all plans reported having a lump
sum feature for those who separate before retirement although we believed
some of those plans did so. In addition, some plans provided extensive
details on discount rates and formulas used in their opening account
balance calculations while others provided no information. In situations
where we could not find information on the form or its attachments, we
recorded this as "information not found." Finally, although the Form 5500
provides information on the number of active participants in the entire
plan, it was often impossible to determine how many of those participants
were converted to the CB plan in instances where only certain employee
groups were converted. Nevertheless, our estimates are based on plan-level
data.

The sample design for this study was a stratified random sample of CB
plans, with the 45 largest plans comprising the first stratum, and an
additional 160 plans selected from the remaining plans, producing a total
sample of 205 plans. Of these sampled plans, we obtained sufficient plan
information for 165, we found 21 plans to be out-of-scope for our study
(not CB plans), and for 19 plans we could not obtain sufficient
information on the plans. Also, of these 205 sampled plans, 7 plans
started a new CB plan only for the new employees, while keeping their
existing employees in the traditional DB plan. We did not include these
plans in our analysis since they were start-up CB plans and not converted
CB plans.

This sample disposition information is summarized in table 1.

 Table 1: Cash Balance Plan Sample Disposition
                          Not CB  Sufficient           Insufficient Completion 
 Stratum Population Sample plana information Converted information        rate 
                                             plan                   
 1.                                                                            
 Largest         45     45   2            40        31            3        93%
 45                                                                 
 plans                                                              
 2. Rest of         798 160  19          125       102           16        90% 
 plans                                                              
 Total              843 205  21          165       133           19 

Description of the Review

Source: GAO analysis of sampled Form 5500 data.

a

Sampled plans that were determined to be hybrid plans other than CB plans
were outside the scope of this study.

After obtaining Form 5500s, attachments, and summary plan descriptions
where available 5 for sampled plans, we recorded plan features on a
standardized instrument containing 51 questions designed to capture
information about

     o characteristics of the traditional DB plan, such as the conversion
       date and the type of DB plan in place before the conversion;
     o the conversion such as when it took place, which employees were
       affected, and the type of transition provisions used; and
     o the ongoing features of the CB plan, such as pay credits and interest
       credits provided at the time of conversion.

Estimates Estimates of converted CB plans were based on our sample of CB
plans. Estimates for this target population were formed by weighting the
survey data to account for both the sample design and the completion rate.

Sampling Error Because we surveyed a sample of CB plans, our estimates are
subject to sampling errors that are associated with samples of this size
and type. A

5

We had some summary plan descriptions available to us as a result of past
GAO work on cash balance issues. See GAO, Private Pensions: Implications
of Conversions to Cash Balance Plans , GAO/HEHS-00-185 (Washington D.C.:
September 2000) and GAO, Cash Balance Plans: Implications for Retirement
Income, GAO/HEHS-00-207 (Washington D.C.: September 2000). We determined
that some plans that had supplied summary plan descriptions reviewed in
those studies were also included in the sample of this study. In addition,
for this study a few plan sponsors provided plan documents upon our
request for additional information and information on a few other plans
was available via the Internet.

                               Nonsampling Error

different random sample could produce slightly different estimates. Our
confidence in the precision of the results from this sample is expressed
in 95 percent confidence intervals. The 95 percent confidence intervals
are expected to include the actual results for 95 percent of the samples
of this type. We calculated confidence intervals for our study results
using methods that are appropriate for a stratified, probability sample.
For the percentages presented in this report, we are 95 percent confident
that the results we would have obtained if we had studied the entire study
population are within +- 9 or fewer percentage points of our results. For
example, we estimate that 47 percent of the CB plan conversions offered
some form of grandfathering. The 95 percent confidence interval for this
estimate would be no wider than +- 9 percent, or from 38 percent to 56
percent.

In addition to sampling error, the practical difficulties in conducting
sample file reviews of this type may introduce other types of errors,
commonly referred to as nonsampling errors. For example, questions may be
misinterpreted, or errors could be made in keying questionnaire data. We
took several steps to reduce these errors.

To minimize some of these errors, each completed data collection
instrument was verified for accuracy, and a process of content analysis
was undertaken to resolve interpretation differences. We performed 100
percent verification of all keypunched questionnaire data. We also traced
and verified the data collection instrument to descriptive statistics and
output generated by GAO data analyst staff. In the event of changes, the
entire verification process was again performed which included 100 percent
verification of the new keypunched data, additional content analysis to
verify the change being made, and reverifying the output generated by the
data analyst staff.

In addition, we were only to record a plan as having a characteristic if
evidence of that characteristic was found in the file review. For example,
it is possible that some CB plans had transition provisions at conversion
that were not clearly indicated in the 5500 files and attachments. We can
only conclude that evidence of transition provisions being offered was not
found in the 5500 data for this plan.

Appendix IV: Analysis of Simulated Cash Balance Plans and Traditional
Final Average Pay Plans

To analyze the effects of a CB plan conversion on individual workers, we
used a pension policy simulation model PENSIM. 1 PENSIM is a dynamic
microsimulation model for analysis of the retirement income implication of
government policies affecting employer-sponsored pensions. The model has
been developed by the Policy Simulation Group (PSG) since 1997 with
funding by the Office of Policy and Research at the EBSA of the U.S.
Department of Labor. To meet GAO's needs for this project the model
includes several enhancements that permit the analysis of CB plan
conversions.

PENSIM uses discrete event simulation methods to generate a sample of life
histories that reflect the effects of individual risks (mortality,
disability, earnings, etc.). The likelihood and timing of simulated life
events are represented by a variety of probability models, including
hazard functions and multinomial logit models that have been estimated
using various survey data sets. The timing of job history events and
employer pension sponsorship are estimated using longitudinal SIPP data
and longer-term longitudinal PSID data. Simulated life histories contain
information on educational attainment, disability, mortality, and a
complete job history that includes details on earnings and pension
accumulation for each job. Details of pension plan(s) covering a worker on
a job are assigned using a pension characteristics imputation model, which
has been estimated with late-1990s BLS Employee Benefit Survey data. 2
Life histories simulated by PENSIM generate social security benefit and
payroll tax results similar to those generated by the Congressional Budget
Office's long-term social security model (CBOLT).

PENSIM simulates the pension accruals of employees as they move from job
to job over their lifetime and estimates their retirement income from a
lifetime of pension coverage. With its CB plan analysis capability, PENSIM
can also simulate changes in retirement income caused by conversions from
traditional defined benefit pension plans to CB pension plans. PENSIM
produces a large random sample of simulated life histories for people born
in a given year and for their spouses who may have been born in a
different year. For our report, we do not include spousal benefits in the
analysis. The members of the birth cohort sample experience demographic
and economic events, the incidence and timing of which vary

1

For more information on PENSIM, go to http://www.polsim.com/PENSIM.html.

2

For more information on the pension characteristics imputation model, go
to http://www.polsim.com/penchar.pdf.

Page 59 GAO-06-42 Pension Plans Appendix IV: Analysis of Simulated Cash
Balance Plans and Traditional Final Average Pay Plans 06-42templated.doc

Simulated Pension Plans

by age, gender, education, disability, and employment status. The types of
life events that are simulated in PENSIM include

     o demographic events (birth, death);
     o schooling events (leaving school at a certain age, receiving a certain
       educational credential);
     o family events (marriage, divorce, childbirth);
     o disability events;
     o initial job placement;
     o job mobility events (earnings increases while on a job, duration of a
       job, movement to a new job, or out of the labor force);
     o pension events (becoming eligible for plan participation, choosing to
       participate, becoming vested, etc.); and
     o retirement events.

This broad scope of simulated life events is necessary in order to
simulate lifetime pension accruals with any realism.

Three pension plans are used in this study to simulate several kinds of
private-sector plan conversions and terminations. The baseline from which
the conversion/termination analysis starts is a typical final-pay defined
benefit pension plan ("typical FAP"). This typical FAP plan has common
private-sector characteristics and a benefit formula that produce an
employer cost of providing the pension equal to the average cost of the
full variety of final-pay plans observed in BLS Employee Benefit Survey
data. 3 The second plan considered in the analysis is a typical CB pension
plan ("typical CB") that has been specified to have characteristics found
to be typical of the plans we analyzed in the GAO Form 5500 data
collection

3

We chose to evaluate the effects of converting or terminating a typical
FAP to determine the changes in benefits that would be experienced by
those currently participating in a FAP plan. An alternate approach would
be to base the typical plan on characteristics of FAP plans that elected
to convert or terminate. However, this would have required additional
information and analysis related to the individual circumstances of such
FAP plans that were outside of the scope of our study. While such an
alternative could be used to evaluate the effect of past conversions and
terminations on affected participants, the results would be limited in
predicting the effect of future conversions or terminations on those
currently covered by a FAP pension plan.

Appendix IV: Analysis of Simulated Cash Balance Plans and Traditional
Final Average Pay Plans 06-42templated.doc

conducted as part of this study. 4 The third plan is a more generous
version of the typical CB pension plan ("equal-cost CB") that has been
constructed to have the same employer cost as the typical FP plan.

The typical FAP plan has the following characteristics

     o immediate eligibility;
     o 5-year cliff vesting;
     o normal retirement age of 65;
     o early retirement age of 55 with 10 years of service with benefits
       reduced by five percent for each year of early retirement (i.e., fifty
       percent reduction at age 55);
     o immediate unreduced disability retirement benefit for those who are
       vested;
     o no survivors' benefit for those who die on the job;
     o selection of single-life annuity at retirement (no selection of joint
       and survivor annuity because study ignores survivors' benefits);
     o benefit paid as a nominal annuity (i.e., no benefit COLA);
     o FAP is the highest consecutive five-year average; and
          * benefit formula is excess integrated with a base rate of 1.5
            percent of final pay per year of service and has a rate of 0.45
            percent of final pay per year of service for those amounts over
            the social security maximum.
          * The typical CB plan has the following characteristics
     o immediate eligibility;
     o 5-year cliff vesting;
     o base pay credit of 3.0 percent of salary for employee with age plus
       service of less than or equal to 35;
     o pay credit rises gradually until it is 6.0 percentage points above the
       base pay credit for employee with age plus service greater than or
       equal to 70 (this results in a maximum pay credit of 9.0 percent of
       salary);
     o interest credit is calculated using current 30-year Treasury rate;

The typical CB plan features derived from GAO's Form 5500 data were, in
part, established by employee and sample selection weighting. As stated
previously in this report, it is not known how many participants of the
plan were actually affected by the conversion to a CB plan. However, for
the purposes of construction, we applied the employee weights assuming 100
percent of participants were affected. Results for how participants will
fare under our typical CB plan, when taken in conjunction with our equal
cost CB plan, provide two polar views of how a distribution of individuals
may be affected when converted.

  Appendix IV: Analysis of Simulated Cash Balance Plans and Traditional Final
                      Average Pay Plans 06-42templated.doc

     o employee always rolls over full account balance into an IRA at job
       termination; 5
     o rollover account earns current 30 year Treasury rate each year;
     o account balances are converted to a nominal single-life annuity at
       retirement using the treasury rate, current projected mortality rates,
       and projections of future reductions in mortality. An annuity loading
       fee was used such that it ensures the provider is solvent (i.e., 1.5
       percent for women and 3.0 percent for men);
     o at conversion, opening account balance is equal to the statutory
       present value of accrued benefit under old plan;
          * at conversion, employee with age plus service greater than or
            equal to 60 is grandfathered in old plan so that benefit at job
            end can never be lower than it would have been if the old plan
            had continued operating
          * The equal-cost CB plan has the following characteristics
     o same characteristics as the typical CB plan except the base pay credit
       is 7.35 percent of salary for employee with age-plus-service (APS) =<
       35, rather than the 3.0 percent of salary in the typical CB plan, and
          * pay credit rises gradually until it reaches a maximum of 6
            percentage points above the base pay credit for employee with age
            plus service greater than or equal to 70.
          * These three plans are used to simulate the following conversion
            and termination situations
     o typical CB plan versus typical FAP plan;
     o typical CB plan versus FAP plan that is terminated with no replacement
       of any kind;
     o equal cost CB plan versus typical FAP plan; and
     o equal cost CB plan versus FAP plan that is terminated with no
       replacement of any kind.

All PENSIM runs conducted for this study simulate a 3 percent sample of

Simulation

the 1955 birth cohort using historical information through the present and
Assumptions 2004 OASDI Trustees Report intermediate-cost assumptions for
the future

5

One claimed benefit for CB plans is the ability to rollover account
balances upon separation. Our simulation model fully captures this
feature. This contrasts with a traditional FAP plan where a participant
who leaves before early retirement loses both future final pay increases
and the early retirement subsidy.

Page 62 GAO-06-42 Pension Plans Appendix IV: Analysis of Simulated Cash
Balance Plans and Traditional Final Average Pay Plans 06-42templated.doc

projection. The resulting cohort sample consists of 151,263 individuals
born in 1955 either in the U.S. or elsewhere (and immigrated to the U.S.
in a subsequent year).

The PENSIM runs differ only in their assumptions concerning privatesector
sponsorship of the typical FAP plan (which is assumed to be offered by all
private-sector employers who are simulated to offer a FAP DB plan) and the
typical or equal-cost CB plan (which is assumed to be offered by all
private-sector employers who are simulated to offer a CB DB plan). The
employment history of each individual and coverage/participation in
employer-sponsored DB and DC plans are a key component to determining the
lifetime benefits for each individual. Pension benefits accumulated as a
result of movement to different employers during a person's entire work
history is included in reported results. Pension coverage across a
lifetime may include participation in a variety of DB and DC plans or no
coverage at all. Workers who are not covered under either a private sector
FAP or a CB pension plan are excluded from the study analysis. Most of the
study analysis focuses on those who have vested in at least one
private-sector FAP or CB plan. Public-sector FAP plans are assumed to be
unchanged across all runs, and all other types of DB plans (i.e., other
than FAP or CB) and all types of DC pension plans in all sectors are
assumed to be unchanged across all the runs. Additionally, all the PENSIM
runs used in this study contain the exact same life histories and job
careers for the cohort sample. That is, the only change that takes place
in all PENSIM runs is whether the private sector DB plan is a FAP or a CB
plan. 6 The simulation analysis provides the following general results
about the cohort sample

     o sample individuals who had at least one private-sector FAP or CB
       pension plan: 57,049 (100.0 percent);
     o sample individuals who never vest in such a plan: 20,274 (35.5
       percent);
          * sample individuals who vest in such a plan but die before age 68:
            6882
          * (12.1 percent);
     o sample individuals who vest in such a plan and live to age 68: 29,893

(52.4 percent);

o  of the 29,893, 87.0 percent vest in just one FAP or CB pension plan
over their lifetime, while 12.3 percent vest in two plans, and all but
three of the rest vest in three such plans; and

We did not attempt to model any changes in employee behavior that may
affect job tenure as a result of a conversion to a CB plan.

Page 63 GAO-06-42 Pension Plans Appendix IV: Analysis of Simulated Cash
Balance Plans and Traditional Final Average Pay Plans 06-42templated.doc

o  of the 26,018 who vest in just one FAP or CB pension plan, only 10.2
percent accumulate thirty years or more of service on that job.

The study makes four pair-wise comparisons between PENSIM runs: (1)
typical CB plan versus ongoing typical FAP plan, (2) equal-cost CB plan
versus ongoing typical FAP plan, (3) typical CB plan versus terminated
typical FAP plan, and (4) equal-cost CB plan versus terminated typical FAP
plan. In each comparison, the difference in lifetime pension income
between the two runs is calculated for each sample individual. Lifetime
pension income includes all pension benefits earned during a person's
career even if they are unaffected by the assumed change in employer
pension sponsorship between the two runs. Lifetime pension income is
expressed in one of two ways: the present value of all pension income
received over the individual's lifetime or the monthly pension income
received at age 68. In both cases, the monetary amounts are expressed in
2004 dollars.

The conversion/termination of the typical FAP plan is assumed to occur at
one of eight ages: 25, 30, 35, 40, 45, 50, 55, and 60. The entire cohort
sample was put through eight separate simulation runs --one simulation run
for each age. Results are shown for those who were vested in a job that
was caught in a conversion. The conversion provisions (opening balance and
grandfathering) described above for the typical and equal-cost CB plans
were found to be typical in our analysis of the Form 5500 sample drawn for
this study. Based on our Form 5500 sample plan analysis and meetings with
consultants who are experts on CB plans, there was concurrence that the
opening CB would be equal to the present value of accrued benefit under
the old plan at the conversion date. The expected present value of the
accrued benefit is calculated using a GAM83 mortality table adjusted to
the proper year and the current Treasury rate as the discount rate. 7 If
eligible for grandfathering, an individual receives the higher of two
amounts at job termination: the accumulated CB under the new plan and the
expected present value of the benefit the individual would have received
if the typical FAP plan had not been converted. The expected present value
is calculated using the same mortality and discount assumptions as used in
the opening balance calculation. All individuals affected by a conversion
or termination are covered under the federal

We have no wearaway-neither initial nor inadvertent wearaway, or any other
form-in our modeling.

Page 64 GAO-06-42 Pension Plans Appendix IV: Analysis of Simulated Cash
Balance Plans and Traditional Final Average Pay Plans 06-42templated.doc

anticutback rules. The PENSIM runs use these same mortality and discount
assumptions for the anticutback calculations.

The employer cost of sponsoring a pension plan is defined as the
percentage ratio of the present value of benefits paid to all individuals
who Estimates worked on a job where that pension plan was sponsored and
the present

value of earnings paid to all individuals who worked on a job where that

pension plan was sponsored. The present value calculations use Treasury

rates to discount both the benefit and earnings cash flows. For a FAP
plan,

the benefit cash flow is the annuity payment stream. For a CB plan, the

benefit cash flow is the CB amount paid at job termination.

All employer cost estimates are for the 1955 birth cohort. Using a younger
birth cohort would produce a higher employer cost rate for the typical FAP
plan because of rising life expectancy and about the same employer cost
rate for the typical CB plan because of its earnings-based benefit
formula. The estimated employer cost rates are as follows

     o full variety of private-sector final-pay plans in BLS data: 7.547
       percent;
     o typical FAP plan: 7.545 percent (by construction equal to cost of full
       variety) Note: the employer cost of providing disability retirement
       benefits in the typical FAP plan to the 1955 birth cohort is 0.487
       percent out of 7.545 percent. Note: in order to simplify the study
       presentation, the typical FAP plan is assumed to have no survivor
       benefits, which are actually a typical benefit under FAP plans, and
       thus the 7.545 percent is an underestimate of a typical FAP plan's
       cost;
     o typical CB plan with no conversion costs: 5.006 percent (i.e.,
       conversion age 15); 8
     o typical CB plan with conversion costs by conversion age (see table 2);
       and
     o equal-cost CB plan with averaged conversion costs: 7.547 percent. 9

8

This is an estimate of the ongoing cost of the typical CB plan after all
conversion costs have been paid.

9

The base pay credit rate of the equal-cost CB plan has been adjusted so
that the plan's employer cost equals that of the typical FAP plan.

Page 65 GAO-06-42 Pension Plans Appendix IV: Analysis of Simulated Cash
Balance Plans and Traditional Final Average Pay Plans 06-42templated.doc

Comparison of Estimated Employer Cost of the Typical CB Plan Immediately
After the Conversion with the Estimated Employer Cost of the Typical FAP
Plan

Table 2: Conversions Costs of Typical Cash Balance Plan by Conversion Age

                             Conversion age Percent

                                    25 4.843

                                    30 4.645

                                    35 4.464

                                    40 4.680

                                    45 5.602

                                    50 6.937

                                    55 7.925

                                    60 7.866

Cost averaged over 8 conversions agesa,b

Source: GAO analysis using the PENSIM Model.

aNote that calculating a simple average of the eight cost rates assumes a
uniform conversion-age distribution, which is analogous to assuming a
uniform employee age distribution at the plan conversion date. While this
assumption may not be exactly true for individual plans, there is no
publicly available data that provide information that would support an
assumption of a nonuniform employee age distribution for all plan
conversions.

b

This 5.870 percent is an estimate of employer cost immediately after the
conversion from the typical FAP plan to the typical CB plan when
conversion costs are being paid. Given the widespread belief that typical
cash balance conversions have not produced substantial immediate pension
cost savings for employers, the reasons for the difference between the
7.545 percent and 5.870 percent are discussed below.

There are several reasons why the estimated employer cost of the typical
CB plan immediately after the conversion of 5.870 percent is below the
estimated employer cost of the typical FAP plan of 7.545 percent by about
22 percent. First, the typical FAP plan has been constructed to reflect
the full variety of private-sector FAP plans contained in the BLS Employee
Benefits Survey data used to impute plan characteristics in PENSIM. The
characteristics of the typical CB plan are drawn from the Form 5500 sample
used for this study and from discussions with pension experts and
actuaries who confirmed that the characteristics were in the range of what
they believe was typical for CB plans. This sample of CB plans is the
largest available sample, and the only sample to be drawn using
statistical sampling methods.

The difference in the estimated employer cost rates for these two plans is
consistent with prior research. Specifically, the cost difference reported
here is somewhat smaller than the cost difference for typical plan

Appendix IV: Analysis of Simulated Cash Balance Plans and Traditional
Final Average Pay Plans 06-42templated.doc

conversions reported in a widely cited study by Watson Wyatt Worldwide. 10
In the Watson Wyatt study, the employers who convert typical (i.e., middle
of the cost distribution) FAP plans to CB plans-the 20 percent in deciles
5 and 6 in table 9-experience an immediate definedbenefit pension employer
cost reduction of about 19 percent (18.72 percent in fifth decile and
19.76 percent in sixth decile). However, the 22 percent cost reduction
estimated in this study and the 19 percent cost reduction estimated in the
Watson Wyatt study are not comparable because of differences in the Watson
Wyatt life history simulations, which ignore disability events, and
therefore, underestimate the cost of the FAP plans. To make our estimates
comparable to the Watson Wyatt estimates, we subtracted the 0.487 percent
disability costs from 7.545 percent yielding a without-disability employer
cost estimate for the typical FAP plan of 7.058 percent. Our estimate of
the immediate cost of the typical CB plan is 5.870 percent, which is about
17 percent below the 7.058 percent without-disability estimate. This 17
percent immediate employer definedbenefit cost reduction is about the same
as the 19 percent reduction found in the Watson Wyatt study.

See Watson Wyatt Worldwide, The Unfolding of a Predictable Surprise: A
Comprehensive Analysis of the Shift from Traditional Pensions to Hybrid
Plans (2000), 18-19.

Appendix V: GAO Contacts and Staff Acknowledgments

Barbara D. Bovbjerg, Director (202) 512-7215

GAO Contacts

Staff Acknowledgments

The following staff members made major contributions to this report:
Charles A. Jeszeck, Assistant Director, Kimberley M. Granger,
Analyst-in-Charge, Joseph Applebaum, Kevin Averyt, Richard Burkard,
Virginia Chanley, Tamara Cross, David Eisenstadt, Lawrence Evans Jr.,
Benjamin Federlein, Nila Garces-Osorio, Sharon Hermes, Jason Holsclaw,
Gene Kuehneman Jr., Michael Maslowski, Amanda Miller, Michael Morris,
Luann Moy, Macdonald Phillips, Mark Ramage, Tovah Rom, Nyree Ryder, George
Scott, and Roger Thomas.

(130410)

GAO's Mission

The Government Accountability Office, the audit, evaluation and
investigative arm of Congress, exists to support Congress in meeting its
constitutional responsibilities and to help improve the performance and
accountability of the federal government for the American people. GAO
examines the use of public funds; evaluates federal programs and policies;
and provides analyses, recommendations, and other assistance to help
Congress make informed oversight, policy, and funding decisions. GAO's
commitment to good government is reflected in its core values of
accountability, integrity, and reliability.

The fastest and easiest way to obtain copies of GAO documents at no cost
is through GAO's Web site ( www.gao.gov) . Each weekday, GAO posts GAO
Reports and newly released reports, testimony, and correspondence on its
Web site. To

have GAO e-mail you a list of newly posted products every afternoon, go to
www.gao.gov and select "Subscribe to Updates."

                             Order by Mail or Phone

The first copy of each printed report is free. Additional copies are $2
each. A check or money order should be made out to the Superintendent of
Documents. GAO also accepts VISA and Mastercard. Orders for 100 or more
copies mailed to a single address are discounted 25 percent. Orders should
be sent to:

U.S. Government Accountability Office 441 G Street NW, Room LM Washington,
D.C. 20548

To order by Phone: Voice: (202) 512-6000 TDD: (202) 512-2537 Fax: (202)
512-6061

To Report Fraud, Contact:

Web site: www.gao.gov/fraudnet/fraudnet.htm

E-mail: [email protected]

Federal Programs Automated answering system: (800) 424-5454 or (202)
512-7470

Gloria Jarmon, Managing Director, [email protected] (202) 512-4400 U.S.
Government Accountability Office, 441 G Street NW, Room 7125 Relations
Washington, D.C. 20548

Paul Anderson, Managing Director, [email protected] (202) 512-4800

Public Affairs

U.S. Government Accountability Office, 441 G Street NW, Room 7149
Washington, D.C. 20548

PRINTED ON

RECYCLED PAPER
*** End of document. ***