Financial Management: Review of VA's Actuarial Model for Veterans'
Compensation Benefits (Letter Report, 01/29/99, GAO/AIMD-99-46).
In a report on the government's consolidated financial statements for
fiscal year 1997, GAO noted that the Department of Veterans Affairs'
(VA) estimated liability for veterans' compensation benefits was
materially understated primarily because it did not include estimates
for anticipated changes in disability ratings and for incurred claims
not yet reported. (See GAO/AIMD-98-127, Mar. 1998.) Because of these
limitations, VA's methodology for computing the liability did not comply
with Statement of Federal Financial Accounting Standards (SFFAS) No. 5,
which prescribes accounting standards for the federal government's
liabilities. VA revised its model to comply with SFFAS No. 5 before
issuing its own audited financial statements for the Department in April
1998. Using the revised model, VA's estimated liability as of September
30, 1997, in its April 1998 report was $466 billion--an increase of $270
billion over that reported in the government's consolidated financial
statements for fiscal year 1997. This report discusses the improvements
that VA has made to the model and makes recommendations for additional
improvements that should enhance the reliability of estimates produced
by the model.
--------------------------- Indexing Terms -----------------------------
REPORTNUM: AIMD-99-46
TITLE: Financial Management: Review of VA's Actuarial Model for
Veterans' Compensation Benefits
DATE: 01/29/99
SUBJECT: Military benefits claims
Accounting standards
Financial records
Veterans benefits
Financial statement audits
Internal controls
Future budget projections
Government liability (legal)
Actuarial tables
******************************************************************
** This file contains an ASCII representation of the text of a **
** GAO report. This text was extracted from a PDF file. **
** Delineations within the text indicating chapter titles, **
** headings, and bullets have not been preserved, and in some **
** cases heading text has been incorrectly merged into **
** body text in the adjacent column. Graphic images have **
** not been reproduced, but figure captions are included. **
** Tables are included, but column deliniations have not been **
** preserved. **
** **
** Please see the PDF (Portable Document Format) file, when **
** available, for a complete electronic file of the printed **
** document's contents. **
** **
** A printed copy of this report may be obtained from the GAO **
** Document Distribution Center. For further details, please **
** send an e-mail message to: **
** **
** **
** **
** with the message 'info' in the body. **
******************************************************************
FrtCover.book GAO United States General Accounting Office
Report to the Secretary of Veterans Affairs
January 1999 FINANCIAL MANAGEMENT
Review of VA's Actuarial Model for Veterans' Compensation Benefits
GAO/AIMD-99-46
GAO/AIMD-99-46
United States General Accounting Office Washington, D. C. 20548
Lett er
Page 1 GAO/AIMD-99-46 VA Actuarial Model
GAO
Accounting and Information Management Division Lett er
B-281738 January 29, 1999 The Honorable Togo D. West, Jr. The
Secretary of Veterans Affairs
Dear Mr. Secretary: In our report on the fiscal year 1997
consolidated financial statements of the U. S. government issued
on March 31, 1998, 1 we noted that the Department of Veterans
Affairs' (VA) estimated liability for veterans' compensation
benefits was materially understated primarily because it did not
include estimates for anticipated changes in disability ratings
and for
incurred claims not yet reported. Given such limitations, VA's
methodology for computing the liability did not comply with
Statement of Federal Financial Accounting Standards (SFFAS) No. 5,
which prescribes accounting standards for liabilities of the
federal government. Prior to the
issuance of its own audited financial statements for the
department on April 30, 1998, VA revised its model to comply with
SFFAS No. 5. Using the revised model, VA's estimated liability as
of September 30, 1997, in its April 30, 1998 report was $466
billion-- an increase of $270 billion over that reported in the
consolidated financial statements of the U. S. government for
fiscal year 1997, issued in March 1998.
The objectives of our review were to determine whether VA's
revised actuarial model complied with SFFAS No. 5 and whether
future liability estimates could be improved. This report
discusses improvements that VA has made to the model, and our
recommendations for additional improvements that should enhance
the reliability of estimates produced by
the model. Results in Brief VA's revised model resulted in an
estimate for fiscal year 1997 that was
much more consistent with SFFAS No. 5 than that previously used
because it included estimates for anticipated changes in
disability ratings and for incurred but not reported claims.
During our review we informed VA officials that the model needed
further refinements to comply more fully with SFFAS No. 5.
Specifically, the 1 Financial Audit: 1997 Consolidated Financial
Statements of the United States Government (GAO/AIMD-98-127, March
31, 1998).
B-281738 Page 2 GAO/AIMD-99-46 VA Actuarial Model
revised model did not include 1.5 million current active military
personnel, some of whom have sustained injuries and may qualify
for future benefits. Instead, it only included those who had
separated from active military service as of September 30, 1997.
In response, VA modified its model for fiscal year 1998 by
including the current military population, which,
according to VA officials, will further increase the liability
estimate by about $12 billion. The VA Office of the Inspector
General (OIG) is currently testing the underlying data used in the
model as part of its audit of the fiscal year 1998 financial
statements. Also, we noted certain limitations in the data used in
the model to develop the estimated liability. VA's fiscal year
1997 liability estimate was based on 1 to 3 years of experience
for various data elements (e. g. beneficiary type, age, gender).
Studying the experience over such a limited time period may not
fully represent the universe of veterans and therefore could cause
distortions in predicting future benefits. In addition, VA did not
group claimants by conflict- related exposures, such as Agent
Orange; therefore the model did not reflect the impact of such
events on future benefits. Finally, VA's model did not consider
the time lag between date of discharge and date of the initial
award, although a veteran's likelihood of filing a claim decreases
the longer he/ she is out of the service. These data, if
considered along with age and beneficiary type, would provide a
better basis for predicting the likelihood that a veteran would
seek a compensation award. With better experience data over a more
extended
period of time, VA's model would be more predictive, resulting in
a more reasonable liability estimate.
Background VA provides veterans or their dependents with
compensation benefits if the veteran was disabled or died of
military service- connected causes. In fiscal year 1997, 2. 5
million veterans and survivors received compensation
benefits of approximately $16 billion. VA calculates an estimated
liability for compensation benefits expected to be paid in future
years to veterans and, if applicable, their survivors, who have
met or are expected to meet defined eligibility criteria.
The estimated liability for veterans' compensation benefits first
became an area of focus in June 1993, when the Department of
Veterans Affairs' Office of Inspector General (VA- OIG) gave an
adverse opinion on VA's financial
statements for the fiscal year ended September 30, 1992. The
adverse opinion was due in part to the fact that VA did not record
in the Statement of Financial Position the present effects of the
probable future payments
B-281738 Page 3 GAO/AIMD-99-46 VA Actuarial Model
for compensation and pension benefits as required by Office of
Management and Budget (OMB) Bulletin 93- 02.
VA subsequently entered into a contract with a consultant firm to
develop a model for calculating an estimate for the present value
of future veterans' benefits based on VA's definition of the
liability. This model contained
numerous spreadsheets with basic assumptions or specific
calculations and beneficiaries were segregated according to age
and type (veteran, spouse, or child). Based on mortality
assumptions and the age of the beneficiaries, the model applied
Cost Of Living Adjustment (COLA) assumptions to the expected
future benefits, assuming a disability rating would not worsen.
Using U. S. Treasury rates, the stream of future benefits was then
discounted back to September 30 of the fiscal year. VA continued
to use this model through fiscal year 1996. As part of its annual
financial statement audit, the VA- OIG separately contracted with
an independent public accountant (IPA) to review and report on
VA's estimated liability for veterans' compensation benefits.
Based on our review of the 1996 estimated liability, we determined
that the model used was limited in that it only considered those
beneficiaries who were on the payment rolls as of September 30 of
the fiscal year. In order to comply with SFFAS No. 5 the estimated
liability must recognize veterans who had experienced an event
(injury) but had not yet filed a claim and for increases in
benefits other than COLAS. That recognition depends on estimates
of veteran and survivor mortality, analysis of probabilities of
future disability, and other predictions requiring actuarial
expertise.
VA revised its model in an attempt to reflect each of these
factors in its liability estimate disclosed in its fiscal year
1997 consolidated financial statements. Approximately 64 percent
of the $270 billion increase over what was reported in the
consolidated financial statements of the U. S.
government can be attributed to the (1) inclusion of veterans not
currently receiving benefits but who had been injured while on
active duty and (2) inclusion of survivors of veterans, such as
dependent children and spouses, not currently receiving benefits
but who are expected to. The increase was also significantly
affected by factors not exclusively related to a specific cohort.
These included factors such as changes in assumptions relating to
mortality projections, estimates of nondeath terminations such
as eligible veterans electing not to claim benefits or veterans
being institutionalized, and projected changes in the rate of new
entrants in all cohorts.
B-281738 Page 4 GAO/AIMD-99-46 VA Actuarial Model
Both the old and revised model project beneficiary payments for 72
years and then discount them back to the present value. However,
another major difference is the original model incorporated a
perpetuity formula using assumptions from the 28th and 29th years
of projection for years 30 through 72, while the new model
specifically projects based on
assumptions for each of the 72 years. 2 Scope and Methodology
This review was done as part of our audit of the fiscal year 1997
consolidated financial statements of the U. S. government. To meet
our objectives, we contracted with the Ernst & Young LLP (E& Y)
Actuarial Services Group (contractor) to assess VA's methodology
and the
appropriateness of actuarial- based assumptions and other
actuarial judgments applied to VA's calculation of the fiscal year
1997 estimated liability for veterans' benefits. Neither we nor E&
Y audited the underlying data used in the model such as age,
gender, and type of beneficiary (veteran, spouse, children, and
mothers and fathers), and percentage of disability. These data
come from compensation and pension systems within VA and external
sources such as the Department of Defense and the Congressional
Budget Office. The accuracy and completeness of the underlying
data used in the model are equally important as the model. VA
management is responsible for establishing, maintaining, and
assessing the internal controls over the systems that produce such
data. The OIG is currently reviewing the internal controls over
the systems that produce the underlying data as well as
obtaining assurance as to whether the data are accurate and
complete as part of the audit of the fiscal year 1998 financial
statements. Therefore, we are not reporting on whether the
estimated liability reported in the financial statements is
reasonable or reliable.
In order to rely on the work of the contractor, we evaluated the
qualifications and independence of the contractor's staff,
reviewed and approved the contractor's approach plans and work
programs, attended key meetings between the contractor and VA
personnel, and reviewed the contractor's working papers to
determine (1) the nature, timing, and extent of work performed,
(2) the extent of quality control methods used, and (3) whether
evidence in the working papers supported the contractor's 2 The
actuary used 72 years, which is roughly the same as the period
used by the Office of the Actuary of
the Social Security Administration of 75 years.
B-281738 Page 5 GAO/AIMD-99-46 VA Actuarial Model
conclusion concerning VA's methodology for calculating the
reserve. We briefed VA officials and made the contractor's results
available to VA for calculating the fiscal year 1998 estimate. We
performed our review of VA's actuarial model from September 1997
through November 1998 in accordance with generally accepted
government auditing standards. We requested written comments on a
draft of this report from the Secretary or his designee. The
Secretary of Veterans Affairs provided us with written comments,
which are discussed in the "Agency Comments" section and are
reprinted in appendix I.
Model Expanded to Reflect Active Military
In our report on the fiscal year 1997 consolidated financial
statements of the U. S. government, we reported that VA's
estimated liability for veterans' compensation benefits was
materially understated because it did not include estimates for
anticipated changes in disability ratings and for incurred claims
not yet reported. Without including these estimates, VA's
methodology did not comply with SFFAS No. 5. 3 VA revised its
model to comply with SFFAS No. 5 prior to the issuance of VA's
audited financial statements on April 30, 1998. Using the revised
model, VA's estimate as of September 30, 1997, was $466 billion--$
270 billion more than the $197 billion estimate that was reported
in the consolidated financial
statements of the U. S. government. While the revised model
provides decisionmakers with significantly better information
about probable future obligations than the estimates produced
under the previous model, we found that the model needed further
refinements to conform with SFFAS No. 5. Specifically, VA's new
model
did not account for the effect of active military personnel who--
due to events that occurred while on duty-- will probably be
eligible for disability benefits upon leaving military service.
This population needed to be factored into the provision for
claims incurred but not reported, which is one of the elements
required for a loss reserve.
Some of the 1.5 million active military personnel as of September
30, 1997, had already experienced the event (injury) that would
result in
3 SFFAS No. 5, Accounting for Liabilities of the Federal
Government, became effective in fiscal year 1997. It requires
recognition of an expense and the related liability for
compensation benefits when a future outflow or sacrifice of
resources is probable and measurable on the basis of events
occurring on or before the reporting date.
B-281738 Page 6 GAO/AIMD-99-46 VA Actuarial Model
compensation benefits subsequent to leaving the service. Although
the total number of active duty personnel was only 6. 6 percent of
the population of veterans not then collecting compensation or
pension
benefits, historically the largest group of new beneficiaries are
those recently separated from military service. Thus, active
military may be the most significant population of potential new
beneficiaries. We informed VA that the estimated liability
associated with these incurred but not reported
claims should be recognized and VA revised its model accordingly.
VA officials told us that the projection of estimates for this
population will increase the reported liability in its fiscal year
1998 financial statements by about $12 billion.
Data Need to Reflect More Historical Experience
VA's model for estimating the future liability for veterans'
compensation benefits needs to be further refined to include
changes and developments in veterans' disabilities over an
expanded period of time. In general, the fiscal year 1997
liability estimate was based on the experience of the compensation
program only over the past 3 fiscal years when data were
available. For some components, only 1 or 2 years of data were
available. This was due in part to the fact that VA had just begun
to compile these data, some of which were not readily available,
in an attempt to comply with the requirements of SFFAS No. 5. For
example, only 1 year of data were available to predict future
increases in veterans benefits, such as changes in disability
ratings.
Relying on such a short time period can cause distortions in
predicting future benefits. For example, if future benefits are
predicted based on a limited period that represents only the time
frame immediately following a conflict, when there may be an
unusually large number of claims filed, the resulting estimate
will most likely be inflated. Conversely, if the time period
considered is during peacetime, when one would expect a low
number of claims filed, the liability estimate will most likely be
understated. Expanding the period over which experience is studied
increases the credibility of the underlying assumptions and
identifies trends and sensitivity to changes in economic
conditions, resulting in more reasonable estimates. For fiscal
year 1997, the model grouped beneficiaries into cohorts 4 by
4 Cohorts are identifiable groups within beneficiaries and
potential beneficiaries. For example, 40- year old veterans are
considered a cohort.
B-281738 Page 7 GAO/AIMD-99-46 VA Actuarial Model
beneficiary type (veteran, spouse, child) and age. Future benefit
payments were projected for each beneficiary type. VA's estimate
relied on the historical experience with older veterans who served
in World War I, World War II, and the Korean conflict and applied
the result to younger veterans who served in Vietnam or the Gulf
War in projecting new entrants, deaths, or other terminations.
This methodology assumes that there are no more recent conflict-
related exposures that would affect the projections. However,
younger veterans had certain latent exposures that older veterans
did not. For example, the experience of World War II and Korean
War veterans would not have predicted some of the disabilities
arising from the Vietnam War, such as Agent Orange. If VA
beneficiaries had also been
grouped by conflict, the model would have reflected the impact of
conflictrelated exposures.
Additionally, VA's model does not consider the time lag between
date of discharge and date of initial benefit. This is an
important consideration because the likelihood of a veteran filing
a claim decreases the longer he/ she is out of the service. For
example, a 50- year old veteran who retired from military service
a year ago would have a higher probability of receiving a
disability benefit than a 50- year old veteran who was discharged
from military service 25 years ago. If these data were considered
along with age and benefit type, the model could better predict
the likelihood of a veteran securing a benefit rather than just
grouping veterans by age and beneficiary type.
Conclusions By revising its model, VA has significantly improved
the reasonableness and reliability of its process for estimating
its liability for veterans
compensation benefits. It can improve the model further by
examining the experience of more distinct groupings over an
extended period of time. This would give VA a better chance of
finding trends in the data that would
lead to a more predictive and reasonable model. Recommendations We
recommend that the Secretary of Veterans Affairs direct the Under
Secretary for Benefits to further improve its model for estimating
VA's liability for veterans' compensation benefits by: Refining
the estimates by continuing to evaluate the actual experience of
the compensation programs and expanding the period over which
experience is studied.
B-281738 Page 8 GAO/AIMD-99-46 VA Actuarial Model
Considering the impact of conflict- related exposures and the
time lag between discharge and date of initial award, in addition
to age and type of veterans and beneficiaries, as indications of
the propensity for veterans to secure compensation benefits.
Agency Comments In written comments on a draft of this report, the
Secretary of Veterans Affairs concurred with our recommendations
for improving the actuarial
model. VA plans to obtain a contractor's assistance to help it
make the improvements.
This report contains recommendations to you. The head of a federal
agency is required by 31 U. S. C. 720 to submit a written
statement on actions taken on these recommendations to the Senate
Committee on Governmental Affairs and the House Committee on
Government Reform within 60 days of the date of this report. A
written statement also must be sent to the House and Senate
Committees on Appropriations with the agency's first request for
appropriations made more than 60 days after the
date of this report. We are sending copies of this report to the
Chairmen and Ranking Minority Members of the House and Senate
Committees on Veterans Affairs and the Director of the Office of
Management and Budget. We will provide copies to others upon
request. Should you or your staff have any questions, please
contact me at (202) 512- 4476 or Alana B. Stanfield, Assistant
Director, Health, Education and Human Services, Accounting and
Financial Management Issues, at (202) 512- 3197.
Sincerely yours, Gloria L. Jarmon Director, Health, Education, and
Human Services
Accounting and Financial Management Issues
Page 9 GAO/AIMD-99-46 VA Actuarial Model
Appendix I Comments From the Department of Veterans Affairs
Appendi x I
Page 10 GAO/AIMD-99-46 VA Actuarial Model
Appendix II Major Contributors to This Report Appendi x I I
Accounting and Information Management Division, Washington, D. C.
Alana B. Stanfield, Assistant Director W. David Grindstaff,
Assistant Director Martin J. Eble, Senior Auditor
(919274) Let t er
Ordering Information The first copy of each GAO report and
testimony is free. Additional copies are $2 each. Orders should be
sent to the following address, accompanied by a check or money
order made out to the Superintendent of Documents, when necessary,
VISA and MasterCard credit cards are accepted, also.
Orders for 100 or more copies to be mailed to a single address are
discounted 25 percent.
Orders by mail: U. S. General Accounting Office P. O. Box 37050
Washington, DC 20013
or visit: Room 1100 700 4 th St. NW (corner of 4 th and G Sts. NW)
U. S. General Accounting Office Washington, DC
Orders may also be placed by calling (202) 512- 6000 or by using
fax number (202) 512- 6061, or TDD (202) 512- 2537.
Each day, GAO issues a list of newly available reports and
testimony. To receive facsimile copies of the daily list or any
list from the past 30 days, please call (202) 512- 6000 using a
touchtone phone. A recorded menu will provide information on how
to obtain these lists.
For information on how to access GAO reports on the INTERNET, send
an e- mail message with info in the body to: info@ www. gao. gov
or visit GAO's World Wide Web Home Page at: http:// www. gao. gov
United States General Accounting Office Washington, D. C. 20548-
0001
Official Business Penalty for Private Use $300
Address Correction Requested Bulk Rate
Postage & Fees Paid GAO Permit No. GI00
*** End of document. ***