General Services Administration: Overestimation of Federal Buildings Fund
Rental Revenue Projections (Testimony, 03/05/98, GAO/T-GGD-98-69).

GAO discussed the General Services Administration's (GSA) overestimation
of its rental revenue projects for the Federal Buildings Fund (FBF) in
fiscal years 1996, 1997, and 1998, and the actions it is taking to
improve its future revenue projections.

GAO noted that: (1) GSA had documentation supporting the dollar amounts
it attributed to six of the seven reasons it reported for the
overestimation of rental revenue; (2) in addition, GAO and others
identified several weaknesses in GSA's rental revenue estimation
process; (3) GSA was aware of these problems and has taken corrective
actions which GAO believes, if effectively implemented, should help
improve future rental revenue estimates; and (4) further, GSA recently
reported that the actual budget impact of its rental revenue
overestimation to be $634.4 million for fiscal years 1996 and 1997, and
reduced its fiscal year anticipated overestimation substantially.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  T-GGD-98-69
     TITLE:  General Services Administration: Overestimation of Federal 
             Buildings Fund Rental Revenue Projections
      DATE:  03/05/98
   SUBJECT:  Projections
             Federal office buildings
             Rental rates
             Facility maintenance
             Cost analysis
             Repair costs
             Maintenance costs
IDENTIFIER:  Federal Buildings Fund
             
******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed product.                         **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************


Cover
================================================================ COVER


Before the Subcommittee on Public Buildings and Economic Development,
Committee on Transportation and Infrastructure House of
Representatives

For Release
on Delivery
Expected at
9:00 a.m.  EST
Thursday, March 5, 1998

GENERAL SERVICES ADMINISTRATION -
OVERESTIMATION OF FEDERAL
BUILDINGS FUND RENTAL REVENUE
PROJECTIONS

Statement of Bernard L.  Ungar, Director
Government Business Operations Issues
General Government Division

GAO/T-GGD-98-69

GAO/GGD-98-69T


(240291)


Abbreviations
=============================================================== ABBREV


GENERAL SERVICES ADMINISTRATION: 
OVERESTIMATION OF FEDERAL
BUILDINGS FUND RENTAL REVENUE
PROJECTIONS
==================================================== Chapter STATEMENT

Mr.  Chairman and Members of the Subcommittee: 

I am pleased to be here today to address issues you raised about the
General Services Administration's (GSA) overestimation of its rental
revenue projections for the Federal Buildings Fund (FBF) in fiscal
years 1996, 1997, and 1998, and the actions it is taking to improve
its future revenue projections.  You asked us to verify, to the
extent practical, the amounts GSA has attributed to the individual
reasons for the $847 million overestimation that it identified in
January 1997 and determine whether GSA's corrective actions appear to
us to address these reasons.  (See app.  I.) Also, in response to
your request, we identify in this statement weaknesses that we and
others found in the rental revenue estimation process used by GSA, as
well as the actions it has planned or which are already under way to
address them. 

To verify the amount of the revenue overestimation that GSA
attributed to each of the reasons it identified, we examined
documents GSA had developed that explained the amount of
overestimation as well as supplied supporting details, such as a
listing of buildings associated with a particular problem, and we
discussed each reason with GSA program officials and staff.  To
develop an understanding of the rental revenue estimation process
used by GSA, we (1) discussed with GSA program officials and staff
the basic steps involved in the process used for fiscal years 1996
through 1999, and (2) reviewed studies of the process done by an
internal GSA review team, two consulting firms, and GSA's Inspector
General.  To determine the actions taken or planned by GSA, we
interviewed relevant GSA officials and staff, reviewed documentation
associated with the actions, and observed the operation of a new
management information system GSA is developing to, among other
things, help it estimate rental revenues. 

Briefly, we found that GSA had documentation supporting the dollar
amounts it attributed to six of the seven reasons it reported for the
overestimation of rental revenue.  In addition, we and others
identified several weaknesses in GSA's rental revenue estimation
process.  GSA was aware of these problems and has taken corrective
actions which we believe, if effectively implemented, should help
improve future rental revenue estimates.  Further, GSA recently
reported that the actual budget impact of its rental revenue
overestimation to be $634.4 million for fiscal years 1996 and 1997,
and reduced its fiscal year 1998 anticipated overestimation
substantially. 


   BACKGROUND
-------------------------------------------------- Chapter STATEMENT:1

Before proceeding further, Mr.  Chairman, I would like to briefly
explain the operation of FBF, which is administered by GSA's Public
Buildings Service (PBS).  In 1975 , FBF replaced appropriations to
GSA as the primary means of financing the operating and capital costs
associated with federal space owned or managed by GSA.  PBS charges
federal agencies rent, the receipts of which are deposited in FBF. 
Congress exercises control over FBF through the annual appropriations
process that sets annual limits on how much of the fund can be
expended for various activities.  In addition, Congress may
appropriate additional amounts for FBF. 

The specific activities the fund is used for include space
acquisition and the operation, maintenance, repair of, and
improvements to, government-owned and -leased buildings managed by
GSA.  FBF rent revenues have grown from about $2.5 billion in fiscal
year 1987 to about $4.8 billion in fiscal year 1997. 

Each year, as part of the budget process, PBS calculates a revenue
estimate that includes an estimate of rental revenue.  Under the
federal budget process, PBS' initial rental income estimate for a
given fiscal year is made 18 months in advance.  For example, PBS'
initial rental income estimate for fiscal year 1997 was made in the
spring of 1995, using rental revenue estimates for fiscal years 1995
and 1996 as a starting point.  At that time, however, PBS did not yet
have actual rental income data for all of fiscal year 1995, which had
not yet ended, and it also did not have actual rental income data for
fiscal year 1996, which had not yet begun.  Accordingly, as a
starting point, PBS updated and adjusted previous estimates it had
made for fiscal years 1995 and 1996 based on the most recent
information it had at the time, and it made assumptions about events
it believed would affect rental revenues in fiscal year 1997. 

Updated information PBS compiled for fiscal years 1995 and 1996
included expected (1) annualized changes in space inventories for
fiscal year 1995, (2) changes in the inflation rate, and (3) changes
in building delegations under which agencies pay their own building
operating costs and are refunded from FBF for the operating cost
portion of their rent payments.  PBS obtained this information from
GSA's regions and various other sources.  PBS' assumptions included
such factors as using a uniform national average for the number of
months new space would be occupied as well as the rental rate. 


   ACCURACY OF PBS' RENTAL REVENUE
   ESTIMATES
-------------------------------------------------- Chapter STATEMENT:2

As you can see, Mr.  Chairman, forecasting rental revenues is a
complex process carried out 18 months in advance that must be made
with incomplete information for the 2 years preceding the year being
estimated and is based on assumptions about future events.  Given
these circumstances, it would be reasonable to expect some
differences between actual and estimated rental revenues. 

In this regard, PBS' historical trends of estimated revenue versus
actual revenue show actual rent revenue (1) exceeded estimated rental
revenues for each year from fiscal year 1987 through fiscal year
1993--except for fiscal year 1992, when actual rental revenues were
less than estimated rental revenues by 1.6 percent--and (2) differed
from the estimate by less than 2 percent for each year during this
period, except for fiscal year 1990, when actual revenue exceeded the
estimate by 2.8 percent.  However, for each of fiscal years 1994
through 1997, PBS' data showed that annual actual rental revenues
were less than estimated rental revenues, ranging from an
overestimate of about $110.7 million, or 2.4 percent, in fiscal year
1995, to an overestimate of about $422.1 million, or 8.2 percent, in
fiscal year 1996.  (See app.  II.) For fiscal years 1994 and 1995,
PBS overestimated rental revenues by a combined total of $308.1
million, that according to the PBS' Chief Financial Officer, was
absorbed by reducing planned expenditures and using carryover
balances without the need for congressional action.  In addition, PBS
reported a combined overestimate of $773.5 million for fiscal years
1996 and 1997, which PBS dealt with as I will explain in a moment. 
It is important to note that the fiscal year 1996 and 1997
overestimation from the historical analysis does not match the amount
reported to Congress in January 1997 because the revenue estimates
were made at different times. 

In January 1997, PBS expected the total overestimation for fiscal
years 1996 and 1997 to be $847 million.  Subsequently, in July of
1997, PBS increased the anticipated overestimation for fiscal year
1997 by $86.8 million and reported an anticipated overestimation for
fiscal year 1998 of about $109.2 million.  This brought the total
anticipated overestimation for fiscal years 1996 through 1998 to
about $1.04 billion. 

PBS took two actions to deal with the $1.04 billion anticipated
overestimation.  First, PBS refrained from using about $680.5 million
in the spending authority provided by Congress in fiscal year 1997. 
Most of this spending authority was for new construction projects
that PBS expected to be delayed for other reasons.  Congress restored
this spending authority in the fiscal year 1998 FBF appropriation. 
Second, PBS reduced operating expenses in fiscal years 1997 and 1998
by deferring planned expenditures until later years to offset the
remaining $359.5 million.  However, after it had closed its fiscal
year 1997 books, PBS reported the actual budget impact of its
overestimation to be $634.4 million for fiscal years 1996 and 1997,
and reduced its fiscal year 1998 overestimation to $28.3 million. 


   WEAKNESSES IN PBS' REVENUE
   FORECASTING PROCESS
-------------------------------------------------- Chapter STATEMENT:3

As indicated, in January 1997, GSA informed Congress that it expected
its total overestimate of rental revenue for fiscal years 1996 and
1997 to be $847 million.  As shown in appendix I, PBS identified
seven reasons for the overestimation and linked specific dollar
amounts of the overestimate to each reason.  For example, PBS
attributed $209 million of the $847 million overestimate to rental
reductions in fiscal year 1995 in 18 metropolitan areas that had not
been factored into its original estimates.  PBS provided
documentation supporting the amount of the overestimation for six of
the seven reasons.  PBS could not provide data showing how the
amount--$86 million--attributed to the remaining reason (i.e., that
the original fiscal year 1995 rent revenue estimate was generally
higher than actual fiscal year 1995 revenues) was developed. 
Although we examined the documentation PBS provided, we did not trace
all the data compiled by PBS to explain its overestimation back to
the original source documents. 

In July 1997, PBS reported increased overestimates of rental revenue
for fiscal years 1997 and 1998 totalling $196 million, which, if
accurate, would have brought the total overestimation for fiscal
years 1996 through 1998 to over $1 billion.  However, PBS did not
identify the causes of the increased overestimation, and in January
1998, PBS identified the actual fiscal year 1997 budget impact of the
overestimate for fiscal year 1997 to be only about $14.1 million and
the estimated budget impact of the revised fiscal year 1998
overestimation to be about $28.3 million, for a combined total of
$42.4 million. 

During the course of our work, we determined that weaknesses in PBS'
estimation process contributed to the rental income overestimation. 
For example, while PBS reported that rent reductions in 18
metropolitan areas were not factored into the original estimates, PBS
did not report why the 18 areas were omitted.  Through discussions
with PBS staff and review of studies done by (1) the firms of Ernst
and Young and Arthur Andersen--consultants hired by GSA, (2) the GSA
Inspector General, and (3) the Rent Revenue Forecasting GO Team--an
internal GSA review team established to look at PBS' rental revenue
estimation process--we identified several weaknesses in the process
for estimating rental revenues.  These weaknesses included the
following: 

  -- Lack of documented policy and procedures for the rental revenue
     estimation process:  In attempting to determine how the rental
     estimates were made for fiscal years 1996 and 1997 and the
     initial estimate for 1998, we found limited documentation on how
     the calculations were made but no documented policy, procedural
     steps, or guidance on the process.  We had to determine the
     procedures and steps followed in calculating the rental revenue
     estimates through our discussions with PBS officials and staff. 
     The staff who made the basic calculations for the rental
     estimates for fiscal years 1996 through 1998 told us they were
     asked to perform this work on a part-time basis because staff
     who had previously done the calculations had left GSA or moved
     on to other positions due to downsizing, reorganization, and
     attrition.  The PBS staff members said they had no documented
     policies, procedures, or guidelines to follow and had learned
     from other PBS staff how to make the calculations.  Without a
     documented process, however, there is less assurance that, for
     example, all necessary steps and information, such as
     consideration of rental reductions from all metropolitan areas,
     are factored into the estimation process or that the process is
     carried out consistently from year to year.  Both GSA
     consultants, the GSA Inspector General, and the agency's GO team
     all found similar problems when they reviewed the process. 

  -- Unclear lines of responsibility and accountability for revenue
     estimates below the level of the PBS Commissioner:  Although we
     were able to identify PBS staff members who made the basic
     revenue estimate calculations, their responsibilities were not
     documented, nor were the responsibilities of other PBS offices
     and staff involved in the process.  This created a lack of
     control over the process for estimating rental revenues.  For
     example, one of the seven reasons for the overestimation PBS
     identified was an error (described by PBS as a technical error)
     in which $66 million was added in two different places instead
     of one.  In our opinion, an effective review process would have
     caught this error, but it was not clear who in PBS was
     responsible for reviewing and approving calculations.  A May
     1997 report by the internal GSA Rent Revenue Forecasting GO team
     also pointed out that, while many different persons and offices
     provided critical input for preparing estimates, there was no
     clear identification of responsibility or accountability for the
     revenue estimates within PBS except for the PBS Commissioner. 

  -- Lack of supporting documentation necessary to verify forecast
     information and assumptions:  In its May 1997 report, the
     internal GSA GO team noted that it had to rely on the memories
     of, and anecdotal information provided by, GSA employees to
     reconstruct the reasons for deficiencies in the rental revenue
     estimates.  To illustrate, one of the seven reasons for the
     overestimation identified by PBS was a change in assumptions
     about costs that was made in 1995 relative to the fiscal year
     1997 rent revenue estimate.  For example, one assumption changed
     was the estimated time that increased government-owned space
     would be occupied in a fiscal year.  The time was changed from 6
     months to 9 months, which resulted in an increase in the
     overestimation of rental revenue.  However, PBS staff said that
     they could not recall who had authorized the change in the
     assumptions.  Subsequently, PBS officials advised us that
     responsibility for the change in the assumptions was borne by
     the then PBS Commissioner.  We could not locate any
     documentation explaining why the change was made.  GSA's
     Inspector General also noted that PBS lacked documentation for
     the assumptions made and methodology used to increase the
     revenue gap expected for fiscal year 1997. 

  -- Use of national averages, rather than project-specific data, to
     forecast occupancy schedules and rental rates:  For fiscal years
     1996 and 1997, PBS reported that its use of national averages
     (which caused estimates of government-owned space increases to
     be too high) accounted for $142 million of the $847 million
     overestimation.  For example, PBS assumed that it would receive
     rent for all space coming on line, for 9 months of the year, at
     the national average per square foot rental rate.  In using
     national averages, PBS relied on less accurate data for
     estimating than if it had used project-specific data.  In
     addition, in its calculations, the national average rental rate
     for government-owned space was changed from $40 to $44 per
     square foot without supporting documentation explaining the
     reason for the change. 

Additional problems with PBS' rental revenue estimation process have
also been identified.  For example, in July 1997, Arthur Andersen
reported that PBS lacked documentation for its budget methodology,
including FBF, and had problems with its information and analysis
systems as well as its pricing policies and practices. 


   GSA'S EFFORTS TO IMPROVE ITS
   REVENUE PROJECTION PROCESS
-------------------------------------------------- Chapter STATEMENT:4

GSA is aware of the identified weaknesses in its revenue estimation
process and has corrective actions to improve this process either
already under way or planned.  These corrective actions include the
following: 

  -- Documentation is to be required for all decisions, assumptions,
     and steps involved in the rental revenue estimation process. 
     According to a PBS official, PBS will issue a directive on
     documenting the rental estimating process by April 1998.  In our
     review of PBS' fiscal year 1999 rental revenue estimate, we
     found that documentation had been prepared on the decisions,
     assumptions, and steps involved in the process. 

  -- Office of Financial and Information Systems (FIS), with overall
     responsibility for the rental revenue forecasting process, was
     established:  In July 1997, GSA issued an order establishing FIS
     with one of its responsibilities being to forecast rental
     revenue and monitor revenue status monthly to determine whether
     predictions of inventory changes, other technical assumptions,
     and income collected are occurring as anticipated.  Within FIS,
     there is a Rent Team, with six people (two additional positions
     are authorized, but not yet filled) responsible for executing
     these duties.  In the past, forecasting revenue was treated as a
     part-time task, and monitoring was done quarterly.  In addition,
     in April 1997, PBS hired a Chief Financial Officer to oversee
     FIS.  Also, each of GSA's regions was directed to appoint a
     revenue manager to be responsible for this issue.  According to
     a PBS official, as of February 20, 1998, 10 of the 11 regions
     have done so. 

  -- Project-specific data is to be used in occupancy schedules and
     rental rates instead of national averages:  In July 1996, PBS
     instructed each of its regions to submit monthly data on the
     changes expected in occupancy and rental rates for each property
     in its inventory.  This would include known changes caused by
     government downsizing.  However, PBS cautions that general
     estimates on downsizing that are not project-specific are too
     speculative to be used in making rental revenue forecasts.  PBS
     intends to use project-specific data, whenever possible, to
     provide a more realistic fact-based estimate on which to base
     future rental revenues. 

  -- New information system is being implemented to manage, track,
     and access data, with plans for a revenue forecasting module to
     be added to the system:  By January 1998, PBS had installed its
     new information system, called the System for Tracking and
     Administering Real Property (STAR), in all its regions.  PBS
     expects STAR to generate a more accurate inventory and greater
     integration of financial and operational data.  According to PBS
     officials, they plan to develop a rental revenue forecast module
     for STAR.  They expect this module to be completed by the spring
     of 1999 and to be used to develop the revenue projections for
     the fiscal year 2001 budget. 

It appears that the actions PBS has under way and planned to improve
its rental revenue estimation process address the weaknesses that we
and others have identified.  If effectively implemented, these
actions should help improve future revenue estimates.  However, as
PBS points out, and we agree, since its rental revenue estimate is a
forecast, it is unlikely to produce an estimate that is identical to
actual rental revenue.  While some variance is to be expected in any
estimating process, variances that go beyond a certain level can be
indicative of problems that need to be addressed. 

In this regard, we noted that PBS has not established an acceptable
margin of error against which it can measure the success of its
estimation process.  Having such a benchmark, we believe, would put
PBS in a better position to identify variances that need to be
investigated so that it can explore and fix the causes of excessive
variances, improve its estimation process, and determine its
effectiveness over time. 


   RECOMMENDATION
-------------------------------------------------- Chapter STATEMENT:5

We recommend that the Commissioner, PBS, establish an acceptable
margin of error for its rental revenue estimates, as well as a
process for exploring and resolving causes of variances outside the
margin adopted. 


------------------------------------------------ Chapter STATEMENT:5.1

Mr.  Chairman, that concludes my prepared statement.  I will be happy
to answer any questions the Subcommittee may have. 


PBS REASONS FOR THE OVERESTIMATION
OF REVENUE FOR FISCAL YEARS
1996-1998
=========================================================== Appendix I

                         (Dollars in millions)

Reason for overestimation                                       Amount
------------------------------------------------------------  --------
Less leased expansion space was delivered than was expected,      $232
 and at later dates than expected.
Fiscal year 1995 rental reductions in 18 metropolitan areas        209
 were not factored into the original estimates.
Estimates of the effect of government-owned-space increases        142
 were too high.
The fiscal year 1995 rental revenue estimate was generally          86
 higher than actual fiscal year 1995 revenues. Because of
 the timing of the budget, these high estimates were used as
 the basis for fiscal years 1996 and 1997 projections.
Assumptions concerning the costs of leased and government           82
 space were changed to make them less conservative.
A technical error was made in calculating the effect of             66
 indefinite authority in the rental of space.
Rental revenue decreases from buildings, or portions of             30
 buildings, becoming unoccupied were not factored into the
 original estimate.
======================================================================
Subtotal                                                          $847
In July 1997, GSA increased its estimate of the fiscal year         87
 1997 overestimation but did not identify the causes.
In July 1997, GSA identified an overestimation for fiscal          109
 year 1998 but did not identify the causes.
======================================================================
Total                                                           $1,043
----------------------------------------------------------------------
Note:  PBS' $847 million overestimate was reported to Congress in
January 1997 and subsequently revised downward.  PBS' July 1997
overestimates were also subsequently revised downward. 

Source:  GSA. 


FBF ESTIMATED AND ACTUAL RENTAL
INCOME FOR FISCAL YEARS 1987-1998
========================================================== Appendix II

                         (Dollars in millions)

                                     Revenues
                           ----------------------------
                                                               Percent
                                                                actual
                                                                 is of
Fiscal year                    Estimated         Actual      estimated
-------------------------  -------------  -------------  -------------
1987                            $2,442.7       $2,471.7         101.2%
1988                             2,921.9        2,969.1          101.6
1989                             3,039.8        3,074.2          101.1
1990                             3,188.2        3,278.6          102.8
1991                             3,490.5        3,493.1          100.1
1992                             4,001.1        3,939.0           98.4
1993                             4,429.2        4,432.0          100.1
1994                             4,831.2        4,633.8           95.9
1995                             4,600.1        4,489.4           97.6
1996                             5,117.0        4,694.9           91.8
1997                             5,155.1        4,803.7           93.2
1998                             4,846.1            N/A            N/A
----------------------------------------------------------------------
Note 1:  Rental income does not include reimbursables, outleasing,
and miscellaneous income; therefore, these numbers are less than
rental revenue in GSA's financial statements.  This historical
analysis does not match revenue overestimation reported to Congress
in January 1997 because the revenue estimates were made at different
times. 

Note 2:  N/A = Not available. 

Source:  GSA. 

*** End of document. ***