Disaster Relief Fund: FEMA's Estimates of Funding Requirements Can Be
Improved (Letter Report, 08/29/2000, GAO/RCED-00-182).

Pursuant to a legislative requirement, GAO reviewed how the Federal
Emergency Management Agency (FEMA) determines current and future funding
requirements for the Disaster Relief fund, focusing on the: (1) accuracy
and timeliness of FEMA estimates of costs for past disasters; (2)
reasonableness of FEMA's approach to estimating the timing and cost of
future disasters; and (3) impact of FEMA's initiatives on the rate of
obligating disaster relief funds.

GAO noted that: (1) problems with both the accuracy and timeliness of
the information FEMA provides monthly to Congress on estimated remaining
costs for past disasters; (2) FEMA's headquarters staff relies on data
provided by its regional offices to produce a monthly report to Congress
on funding requirements for past disasters; (3) however, as of the end
of August 1999, for a third of all past disasters, GAO found that staff
from headquarters and regional offices disagreed on the amount of funds
obligated to date; (4) collectively, differences, in reported
obligations between headquarters and regional staff totaled nearly $250
million--with headquarters reporting obligation amounts that were $18
million higher than regions reported obligations; (5) obligation amounts
reported by regional staff were both higher and lower than amounts
reported by headquarters staff, and FEMA officials could not tell GAO
which amounts were correct; (6) when regional offices questioned the
accuracy of the data, FEMA headquarters staff responsible for
reconciling discrepancies initially failed to determine the cause and
make the needed corrections; (7) instead, FEMA used the inaccurate data
to report on remaining costs for past disasters; (8) in addition,
because of the time needed to assemble and analyze the data, the
information on remaining costs in the monthly report is based on
obligation data that are 4 to 6 weeks old; (9) FEMA officials
acknowledge that data problems exist and have taken steps to correct
them; (10) FEMA can improve its approach for estimating the timing and
cost of disasters anticipated to occur during the remainder of the
current fiscal year and in the forthcoming fiscal year; (11) FEMA's
recent initiatives designed to expedite the closeout of its funding
activities for past disasters has had a significant impact on FEMA's
rate of obligating disaster relief funds; (12) in particular, the
formation of three teams of staff from FEMA's Office of Financial
Management, referred to as "closeout teams," to facilitate the closure
of funding activity for disasters occurring from fiscal years 1989
through 1997 increased FEMA's obligations and recoveries for past
disasters considerably; (13) at the end of September 1997, 419 old
disasters, with a projected remaining cost to FEMA of over $3 billion,
remained open; and (14) it is less clear what impact other FEMA
initiatives, such as setting goals to provide disaster funding more
quickly, are having on the Fund's requirements because sufficient time
has not elapsed to measure their effectiveness.

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

 REPORTNUM:  RCED-00-182
     TITLE:  Disaster Relief Fund: FEMA's Estimates of Funding
	     Requirements Can Be Improved
      DATE:  08/29/2000
   SUBJECT:  Disaster relief aid
	     Emergency preparedness
	     Data integrity
	     Reporting requirements
	     Future budget projections
	     Funds management
IDENTIFIER:  FEMA Public Assistance Program
	     FEMA Disaster Relief Fund
	     FEMA Hazard Mitigation Grant Program
	     FEMA Individual and Family Grant
	     FEMA Integrated Financial Management Information System

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Testimony.                                               **
**                                                              **
** 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/RCED-00-182

Resources, Community, and
Economic Development Division

B-284058

August 29, 2000

The Honorable Christopher S. Bond
Chairman
The Honorable Barbara A. Mikulski
Ranking Minority Member
Subcommittee on VA, HUD, and
Independent Agencies
Committee on Appropriations
United States Senate

The Federal Emergency Management Agency's (FEMA) Disaster Relief Fund is the
major source of federal disaster recovery assistance for state and local
governments when a disaster occurs. The Fund receives an annual
appropriation of up to $320 million. Amounts needed above the base level
generally have been provided as emergency supplemental appropriations,
usually in response to large disasters. The Fund received about $3.4 billion
in regular (nonemergency) appropriations for disaster relief over the
previous 10 years, whereas about $24 billion was provided through emergency
supplemental appropriations. Because appropriations made to the Fund are
available until spent, the Fund usually carries balances that are obligated
as needed for relief or recovery projects. The balance in the Fund
fluctuates continually as funds are obligated.

FEMA is required by the House and Senate Appropriations Committees to report
monthly on the Fund's status. The reports are intended to enable the
Congress to better monitor the availability of funds and allow for the
timely preparation of requests for supplemental funds. The reports provide a
snapshot of the funds available, obligations to date, remaining costs from
past disasters,1 and the anticipated costs of disasters that might occur in
the remaining months of the fiscal year.

In early 1999, FEMA mistakenly projected that the Fund would have a
carryover of $700 million at the end of fiscal year 1999. A month later,
FEMA revised its monthly report to indicate that the Fund faced a potential
funding shortfall and that it would run out of money before the end of the
fiscal year. Concerned about whether FEMA is providing the Congress with
accurate and timely information with which to make budget decisions, as well
as about the impact of recent FEMA initiatives to fund projects more
quickly, your Committee mandated that we, among other things, review how
FEMA determines current and future funding requirements for the Disaster
Relief Fund.

As agreed, this report addresses (1) the accuracy and timeliness of FEMA's
estimates of remaining costs for past disasters, (2) the reasonableness of
FEMA's approach to estimating the timing and cost of future disasters, and
(3) the impact of FEMA's initiatives on the rate of obligating disaster
relief funds.

We found problems with both the accuracy and timeliness of the information
FEMA provides monthly to the Congress on estimated remaining costs for past
disasters. FEMA's headquarters staff relies on data provided by its regional
offices to produce a monthly report to the Congress on funding requirements
for past disasters. However, as of the end of August 1999, for a third of
all past disasters, we found that staff from headquarters and regional
offices disagreed on the amount of funds obligated to date. Collectively,
differences in reported obligations between headquarters and regional staff
totaled nearly $250 million--with headquarters reporting obligation amounts
that were $18 million higher than the regions reported obligations.
Obligation amounts reported by regional staff were both higher and lower
than amounts reported by headquarters staff, and FEMA officials could not
tell us which amounts were correct. When regional offices questioned the
accuracy of the data, FEMA headquarters staff responsible for reconciling
discrepancies initially failed to determine the cause and make the needed
corrections. Instead, FEMA used the inaccurate data to report on remaining
costs for past disasters. In addition, because of the time needed to
assemble and analyze the data, the information on remaining costs in the
monthly report is based on obligation data that are 4 to 6 weeks old. FEMA
officials acknowledge that data problems exist and have taken steps to
correct them. For example, FEMA is developing a new automated system to
estimate remaining disaster costs on a real-time basis. The new system is
projected to be operational in August 2000.

FEMA can improve its approach for estimating the timing and cost of
disasters anticipated to occur during the remainder of the current fiscal
year and in the forthcoming fiscal year. We recognize that it is difficult
to estimate the number, severity, and timing of future disasters. However,
in estimating the timing and cost of future disasters, FEMA does not
adequately account for the timing of past disasters. FEMA uses the 5-year
annual average level of obligations for past disasters, adjusted for
inflation, as its estimate of the total cost of disasters anticipated to
occur during the current fiscal year. To estimate when during the year the
disasters will occur, FEMA simply allows the 5-year annual average to
decline at a constant rate (8 percent) each month during the fiscal year.
Using this approach, FEMA estimates that disasters costing about $500
million will occur in August and September 2000. However, these months
represent the height of the hurricane season, and over the last 5 years, the
average cost for disasters to FEMA has been twice this amount. We believe a
better way of estimating the cost and timing of future disasters would be
for FEMA to use actual monthly data on the estimated total cost of disasters
that have occurred over the last 5 years. This report recommends ways for
FEMA to improve its methods for projecting the timing and cost of future
disasters.

FEMA's recent initiatives designed to expedite the closeout of its funding
activities for past disasters has had a significant impact on FEMA's rate of
obligating disaster relief funds. In particular, the formation of three
teams of staff from FEMA's Office of Financial Management, referred to as
"closeout teams," to facilitate the closure of funding activity for
disasters occurring from fiscal years 1989 through 1997 increased FEMA's
obligations and recoveries for past disasters considerably. At the end of
September 1997, 419 old disasters, with a projected remaining cost to FEMA
of over $3 billion, remained open. Over the next 2 years, FEMA closed out
its funding activities on 382 of the disasters that had occurred from 1989
through 1997, obligating $2.9 billion in disaster relief funds. As a result,
the amount that FEMA records as remaining costs for these disasters was
reduced to approximately $296 million. It is less clear what impact other
FEMA initiatives, such as setting goals to provide disaster funding more
quickly, are having on the Fund's requirements because sufficient time has
not elapsed to measure their effectiveness.

We provided FEMA with a draft of this report for its review and comment.
FEMA acknowledged that data inconsistencies have occurred in reporting to
the Congress on the status of the Disaster Relief Fund. It stated that it is
currently revamping its process for data collection and reporting and will
add controls to ensure that disaster data are updated and reported in a
timely, accurate, and consistent manner. However, FEMA expressed concern
about three areas of the report that it believes needs further
clarification, including our recommendation aimed at improving how FEMA
projects the timing and cost of future disasters. We modified our
recommendation but did not modify the report in other areas. FEMA also
suggested technical clarifications, which we incorporated into the report as
appropriate.

FEMA's Disaster Relief Fund supports a wide range of programs providing
grants to assist state and local governments and certain private nonprofit
organizations. Its largest funded program, the Public Assistance program, is
intended to help communities repair or replace roads, bridges, utilities,
and public buildings after a disaster occurs. Similarly, its Hazard
Mitigation Grant program assists communities in implementing long-term
hazard mitigation measures, such as the purchase of flood-prone buildings,
following a disaster. The Fund also provides people in disaster areas with
temporary housing assistance and, through its Individual and Family Grants
program, with grants to meet other needs.

While FEMA's budget for a given fiscal year averages approximately $2.8
billion, the Congress caps the agency's annual appropriation at $320 million
or at the President's budget request, whichever is lower.2 The remainder
comes from emergency supplemental appropriations made by the Congress, as
needed to respond to large disasters when they occur. Annual budget requests
are based on current Fund balances and estimates of funding requirements for
past and future disasters. 3 Money is appropriated to the Fund on a "no-year
basis"; that is, the dollars remain available until expended. It is
difficult, however, to determine funding requirements because of the
uncertainties in forecasting the number, magnitude, and types of disasters
that will occur in any year. Once a disaster occurs and the President
officially declares that it is eligible for federal assistance, joint FEMA
and state teams visit the affected communities, survey damaged and destroyed
facilities, determine eligibility, and develop initial estimates of damage
costs. This information makes up the Project Worksheet (formally the Damage
Survey Report). On the basis of those Project Worksheets, FEMA obligates
funds for approved projects; that is, funds are made available to the state
to begin repair and restoration work. However, because the time required to
identify and approve projects varies (some require historical and
environmental reviews, or the amount of required funding is in dispute), as
does the time required to contract and perform the needed work, it is
difficult to accurately forecast the fiscal years in which payments will be
made for projects related to specific disasters.

FEMA provides monthly reports to the House and Senate Appropriations
Committees on the status of the Fund. These reports contain information on
funds available for the current year, actual obligations to date, remaining
costs for disasters that have occurred to date, an estimate of funding
requirements for these disasters for the remainder of the fiscal year, a
projection of the costs of disasters that might occur in the remaining
months of the fiscal year, and an estimate of funding requirements for the
next fiscal year.

During fiscal years 1998 and 1999, FEMA implemented two program initiatives
that could affect how it estimates budget requirements in the future. First,
FEMA's Director chartered three closeout teams in November 1997 to expedite
the closeout of the agency's funding activities for past disasters--many
dating back to 1989. The teams are composed of staff from FEMA's Office of
Financial Management who focus on resolving program issues and obligating
funds for approved projects to close out work on specific disasters. Second,
in 1998, FEMA redesigned its Public Assistance program to provide money to
applicants more quickly and to simplify the application process. For
example, FEMA established a goal to approve projects and obligate funds
under the Public Assistance program for 90 percent of disasters within 2
years of the date on which they are officially declared disasters by the
President.

Disasters

To produce estimates of the Disaster Relief Fund's current requirements,
FEMA needs accurate information on the remaining costs for past disasters.
FEMA relies on its regional offices for these cost estimates. To estimate
the remaining costs for past disasters, staff from the regional offices take
their estimates of the total projected costs of past disasters and subtract
data provided by FEMA headquarters staff on the obligations to date. Thus,
the accuracy of the regions' estimates of remaining costs is dependent on
the quality and timeliness of the obligation data FEMA headquarters provides
to the regions. We found that, largely because of a mistake in the process
FEMA uses to extract data from its Integrated Financial Management
Information System (IFMIS), errors exist in the obligation data that are
provided to the regions. These errors, in turn, contribute to errors in the
regions' data on remaining costs. In addition, because of the time needed to
complete this process, the cost estimates produced by the regional offices
for FEMA headquarters are based on obligation data that are as much as 4 to
6 weeks old. These flawed data are ultimately used by FEMA to prepare its
monthly report to the Congress on the projected obligations for the
remainder of the year for past disasters. FEMA has taken steps to correct
the weaknesses identified in its data, such as the development of a new
automated database system to replace its current system for tracking
disaster costs.

Costs for Past Disasters

FEMA relies on its regional offices to periodically update cost information
on past disasters. Consequently, staff in FEMA's 10 regional offices
maintain a Disaster Financial Status Report on each disaster within their
areas of responsibility to develop this information when it is needed for
budgeting purposes.

Quarterly, FEMA's headquarters asks regional office staff to update
estimated cost data, by program area, for each disaster in their database of
Disaster Financial Status Reports. To assist the regional offices with the
quarterly update, FEMA's Office of Financial Management provides each with a
spreadsheet of obligations to date by disaster, extracted from its IFMIS.
IFMIS is FEMA's official record of the Disaster Relief Fund's budget,
obligation, and expenditure transactions. This system does not, however,
have the capability to produce standard reports on obligations by disaster
to date by individual programs (e.g., Public Assistance and Hazard
Mitigation Grant programs). Thus, FEMA uses a special process to extract the
obligation to date data from the IFMIS system.

Once updated, the regions forward the spreadsheets of updated cost data to
the Office of Financial Management, where they are consolidated into a
single national database that is subsequently used to develop FEMA's budget
submissions and monthly reports to the Congress. To allow time for the
regional offices to prepare their quarterly reports, the extraction is made
from 4 to 6 weeks prior to the date of the monthly report. As a result,
regional staff often must forecast a disaster's total estimated expenditures
based on obligation data that are at least a month old. Compounding this
timing problem, some obligation amounts are inaccurate because of an error
in the process FEMA uses to extract the data from IFMIS and because
obligation amounts were either missing from or incorrectly recorded in
IFMIS.

During our review of Disaster Financial Status Report data provided by FEMA,
we found that the regions' obligation amounts often differed from the
obligation amounts reported by IFMIS. Using data submitted for the August
1999 Disaster Financial Status Report, we found that the regions identified
individual differences in over a third of the disasters, collectively
reaching nearly $250 million nationwide. No pattern in the differences
between the regions' and headquarters' obligation amounts was
distinguishable. Some of the differences in obligation amounts reported by
the regions were higher than the amounts provided by headquarters, and some
were lower. Obligations reported by headquarters were $18 million higher
than those reported by regional staff. FEMA officials were unable to tell us
which numbers were correct. According to FEMA officials, several of the
regional offices do not regularly update the data sent to them by FEMA
headquarters. If all of the regional offices updated the data, total
differences would likely have been larger. Although regional offices
questioned the accuracy of the obligation data, headquarters staff
responsible for reconciling data discrepancies initially failed to determine
the cause and make the needed corrections.

The fact that differences exist between regional data and IFMIS data raises
concern about the accuracy of the remaining cost estimates used by the
regions to produce quarterly Disaster Financial Status Reports. Any
difference in the data can contribute to incorrect cost estimates because
the total estimated cost of a disaster equals the obligations to date plus
the amount of the estimated remaining obligations. If the initial obligation
amounts were incorrect, the regional cost estimates FEMA uses to produce its
monthly reports on funding requirements for past disasters would also be
incorrect.

for the Remainder of the Fiscal Year for Past Disasters

FEMA's approach to projecting obligations for the remainder of the year for
past disasters appears to be reasonable. To produce the projection, FEMA
multiplies its estimate of the remaining costs for past disasters, as of the
beginning of the fiscal year, by an assumed obligation rate to arrive at
total obligations for the fiscal year. Actual obligations to date are then
subtracted from the current year estimate to obtain FEMA's monthly
projection of obligations for the remainder of the year for past disasters.
Based on our work, we believe that this approach, if used with accurate
estimates of remaining costs for past disasters, would produce reasonable
estimates of remaining obligations for these disasters. Most importantly,
the obligation rates currently used by FEMA are based on the most recent
data available and, in our opinion, are reasonable. However, in producing
its estimates of remaining obligations for past disasters, FEMA uses
estimates of remaining costs for past disasters that are provided by its
regional offices through its quarterly update process, as already discussed.
Because of the discrepancies in these estimates, the accuracy of FEMA's
projections of obligations for the remainder of the fiscal year for past
disasters is questionable.

Estimate Funding Requirements for Past Disasters

FEMA, in researching the errors found during our review, identified two
possible problems. According to FEMA officials, the first involved a
systemic error in the computer program used to extract obligation data from
IFMIS. The second involved errors in the IFMIS database resulting from
obligation amounts that were missing, or incorrectly recorded, in the
system. FEMA is in the process of correcting those problems.

The first problem FEMA officials noted involved projects that ultimately
cost less than the amount FEMA initially approved. FEMA found that if a
state returned the unexpended funds for a particular disaster, the method
FEMA was using to extract data from IFMIS did not recognize and account for
the unexpended funds. According to FEMA officials, the computer program has
since been modified to recognize differences between obligations and
budgeted amounts and to look for the transaction crediting the budget
amount. If the credit transaction is found, the extraction program
deobligates that amount. According to FEMA officials, at the end of March
2000, FEMA used the new extraction method and found that collective
differences between obligation amounts reported by its 10 regional offices
and IFMIS obligation amounts had been reduced to about $41 million.
According to these officials, in April and May 2000, headquarters staff
worked with regional staff to resolve the remaining differences. By June,
FEMA had identified and corrected about $37 million in errors--leaving a
total of about $3 million in differences between the obligation amounts
reported by the regions and headquarters. We did not independently verify
these data. FEMA is continuing to research the remaining differences.

According to FEMA officials, of the $41 million in differences between the
regions' and headquarters' obligation amounts that existed after the new
extraction method was employed, about $36 million, or 90 percent, was due to
data errors in IFMIS.4 According to FEMA, for various reasons, valid
obligation amounts were either not recorded or were incorrectly recorded in
IFMIS. For example, FEMA officials stated that while researching the cause
of the differences between the regions' and headquarters' obligation
amounts, they found that some refunded payments were not recorded in the
IFMIS database. These involved cases in which states had made refund
payments directly to FEMA. When FEMA converted from its previous financial
system to the current system (IFMIS), the regional office staff manually
reconciled all data between the two systems as well as the system FEMA uses
for tracking the transfer of disaster funds to the states. Any unmatched,
and therefore unsupported, transactions were removed from the IFMIS
database.5 In researching the errors identified in our review, FEMA staff
determined that the removed transactions were actually valid credit
transactions representing state refunds submitted directly to FEMA, thus
bypassing the electronic funds transfer system. According to a FEMA Office
of Financial Management official, a review of the records in question
revealed that, at most, $1.4 million in transactions, covering disasters
that occurred between 1989 and 1996, was missing from the IFMIS database.
FEMA is in the process of verifying and reinstating the removed
transactions.

In an effort to develop more timely and accurate cost estimates for
disasters, FEMA is developing a new database system to replace its current
system for tracking disaster costs. The new system comprises individual
database tables for the Public Assistance, Individual and Family Grants, and
Hazard Mitigation Grant programs. The new system is expected to provide
complete cost information for a disaster. To ensure data integrity, the new
system is expected to restrict access to only those program officials who
are in the region where a disaster is managed and who have responsibility in
the program area. FEMA plans to have the new automated monthly Disaster
Financial Status Report system--in combination with the new data extraction
method--operational beginning in August 2000.

Timing and Cost of Future Disasters

Improvements are needed in the methodology FEMA uses to estimate the timing
and costs of future disasters. In its monthly report on the status of the
Fund, FEMA includes an estimate of the total costs of disasters anticipated
in the current and forthcoming fiscal year.6 Beginning with its first
monthly report in the fiscal year, FEMA allows the estimate of the total
cost of new disasters to decline each month by a constant rate--about 8
percent. This approach results in a constant decline over the course of the
year in the estimated costs for new disasters. However, August and September
represent the height of the hurricane season, and for the past 5 years,
these months have been the most costly months, on average, for FEMA. As a
result, FEMA's method for distributing costs for future disasters does not
adequately account for the timing and severity of disasters.

Figure 1 compares, in 1999 constant dollars, the 5-year average cost of
disasters (from the month the disaster was declared to the end of the fiscal
year) with FEMA's estimate of the cost of disasters to be declared during
the remainder of fiscal year 2000. As shown, FEMA's estimate of the costs of
disasters to be declared from December 1999 through the end of fiscal year
2000--representing the period covered by FEMA's first monthly status report
for the fiscal year--is about $100 million higher than the 5-year average.
Both the 5-year historical average and FEMA's estimate decline over time
because there is less time left in the year for disasters to occur. However,
beginning with the March 31, 2000, report, FEMA's 8-percent rate of decline
causes its estimate of the cost of future disasters to become increasingly
lower than the 5-year average. This result is due to the fact that FEMA's
approach fails to account for the historically high costs of disasters
occurring in the last quarter of the fiscal year. At the end of July, FEMA
estimates that disasters costing about $500 million will be declared during
August and September. However, the 5-year average cost for disasters
declared during this period is about $1 billion--twice as large as FEMA's
estimate.

Sources: GAO's analysis of FEMA's data on estimated total costs, as of
January 31, 2000, for disasters declared from fiscal year 1995 through
fiscal year 1999 and FEMA's monthly report on the Disaster Relief Fund.

We believe a more accurate way to estimate the costs of future disasters
would be for FEMA to base its estimate on the inflation-adjusted 5-year
estimated average costs of disasters--as presented in figure 1. For each
monthly report to the Congress, FEMA could calculate the 5-year average cost
of disasters for the remaining months of the fiscal year. FEMA could then
estimate the required amount of funding for the remainder of the fiscal year
by simply multiplying the 5-year average by the assumed spend-out rate for
the remainder of the fiscal year.

of Obligating Disaster Relief Funds

In early 1999, FEMA discovered that its rate of obligating disaster relief
funds had increased dramatically. According to FEMA officials, much of the
increase in the rate of obligations was the result of the FEMA Director's
recent emphasis on expediting the closure of disasters,7 particularly older
disasters. FEMA officials also cited other recently implemented program
initiatives intended to identify and fund projects faster than in the
past--such as a more efficient grant delivery process--as another factor
contributing to the increase in the obligation rate. It is less clear,
however, what effect these more recent changes are having on FEMA's
obligation rate because sufficient time has not elapsed to quantify their
impact.

Rate

In November 1997, FEMA's Director chartered three teams of Office of
Financial Management staff, referred to as closeout teams, and directed them
to assist FEMA regional office staff and state emergency management
personnel in closing out funding activities for all past disasters.8 The
three closeout teams focused on disasters that had remaining costs for
public assistance and hazard mitigation because projects under these
programs often require longer planning and approval periods. Primarily, the
teams collected grant information from the states and assembled the
documentation needed to facilitate the projects' closure. According to a
FEMA Inspector General report, the teams focused on resolving long-standing
issues and obligating funds for approved projects, thus reducing the amount
reported as remaining costs.9

The closeout teams, which were fully operational in January 1998, originally
focused on disasters declared between fiscal years 1989 and 1994. After
completing the closeout of these disasters, the teams were directed to
assist in the closeout of disasters declared between 1995 and 1997. Their
primary goal was to eliminate all remaining costs for these disasters by the
end of fiscal year 1999 by obligating or recovering the funds. A total of
533 disasters were declared in fiscal years 1989 through 1997. When the
teams began their work, almost 80 percent (419) of these disasters remained
"open," representing over $3 billion in estimated remaining costs. Although
the closeout teams did not completely achieve their goal of closing out all
of the disasters, 382 of the 419 disasters were closed by September 1999,
representing 91 percent. This resulted in reducing the estimated remaining
costs for these past disasters by about $2.9 billion--leaving approximately
$296 million in remaining costs.

Another factor contributing to the reduction in the number of open disasters
was FEMA's February 1999 policy that established separate decision points
for "programmatic disaster closure" and "financial closure." Under the new
definitions, a disaster is "programmatically closed" when all decisions on
the eligibility and funding of the projects dealing with it have been agreed
on and the estimated costs are equal to the obligations for the disaster, in
other words, when all funds for the disaster have been obligated. The
"financial closure" decision point occurs when all funds are disbursed and
both FEMA and the state have reconciled their costs for the disaster, which
could be several years after the programmatic disaster closure occurs.

According to FEMA officials, the closeout teams had a considerable impact on
the rate at which FEMA was obligating funds. The rate of obligations for
disasters declared between fiscal years 1989 and 1997 increased from 42
percent in fiscal year 1998 to 71 percent in fiscal year 1999. This increase
in the rate of obligations for past disasters contributed to FEMA's March
1999 estimate of a potential $900 million shortfall in the Fund by the end
of the fiscal year.

The current appointments for the closeout teams will expire September 26,
2000. FEMA plans to replace the closeout teams with a Chief Financial
Officer's Field Support Team, which would be operational for 2 years. The
role of this new team would be, among other things, to support the regions
in disaster closeout, Disaster Relief Fund reporting and financial
reconciliation, and disaster grant administration.

Initiatives

In addition to the establishment of closeout teams, in September 1997, FEMA
set a goal to close out the Public Assistance program for 90 percent of all
disasters within 2 years of their declaration date. Also, FEMA's Director
approved a redesign of the Public Assistance program in October 1998. The
goals of the new Public Assistance program are to obligate 50 percent of
emergency work funding within 30 days of the declaration and to obligate 80
percent of permanent work funding within 180 days of the declaration.

According to FEMA officials, these changes have also increased the rate at
which FEMA is obligating funds for new disasters. However, our examination
of available data failed to reveal any noticeable increase in obligation
rates for disasters declared since fiscal year 1996. This may be due to the
fact that sufficient time has not elapsed to measure the effects of these
changes. Over time, FEMA will need to assess the effectiveness of these
programmatic changes and make the necessary changes to ensure that disasters
continue to be closed out in a timely manner.

To ensure that FEMA receives adequate resources to respond quickly when
disasters occur, the Congress needs accurate and timely information on the
Disaster Relief Fund's current requirements and projected needs for future
disasters. We recognize that it is difficult to precisely predict when
disasters will occur or their severity. The methodology FEMA currently uses
to estimate the timing and cost of future disasters tends to underestimate
the cost of disasters that occur late in the fiscal year. Through better use
of historical data, FEMA could provide more reasonable estimates of the
timing and costs of future disasters.

To improve the method for projecting the timing and cost of future
disasters, we recommend that the Director of FEMA base FEMA's estimate of
the costs of disasters to be declared during the current and forthcoming
fiscal year on the inflation-adjusted 5-year average cost of declared
disasters. For each monthly report to the Congress, FEMA should use the
5-year average cost of disasters, from the date of the report to the end of
the fiscal year, and should present a range of end-of-year balances based on
assumptions about the timing and cost of future disasters.

We provided FEMA with a draft of this report for its review and comment.
FEMA's comments and our detailed responses are in appendix I. FEMA
acknowledged that some data inconsistencies have occurred in its reporting
to the Congress on the status of the Disaster Relief Fund. FEMA stated that
it is currently taking steps to revamp and streamline its process for data
collection and reporting and will add controls to ensure that disaster data
are updated and reported in a timely, accurate, and consistent manner. We
support FEMA's actions taken to improve data accuracy and commend it for its
quick response.

However, FEMA had three primary concerns with the report that it believes
need clarification. First, the agency believes that the report overstates
the magnitude of the data discrepancies and, by inference, the problems with
FEMA's IFMIS database. Concerning the data discrepancies, FEMA took issue
with our reporting the absolute value of the discrepancies in obligations,
arguing that it is the net amount that affects the monthly reports to the
Congress. This is only true, however, if it is known which numbers are
correct. FEMA could not tell us which numbers were correct--a fact it
acknowledges in its response. Therefore, we believe it is appropriate to
report both sets of numbers, and we did not modify our report.

Second, FEMA believes that the report implies that there are problems with
the official IFMIS database. We disagree. The report clearly states that the
data discrepancies or errors were due largely to the process FEMA used to
extract data--not the raw data in the IFMIS database.

Finally, FEMA took issue with our recommendation that it base its estimate
of the costs of disasters to be declared during the remaining months of a
fiscal year and the forthcoming fiscal year on the inflation-adjusted 5-year
average cost of declared disasters. We continue to believe that our
recommendation will help FEMA improve how it projects the timing and cost of
future disasters, and we did not revise our recommendation to reflect this
issue. FEMA offered an alternative recommendation that it be required to
present a range of available balances in its monthly report to the Congress.
We agree that it might be prudent to provide a range, as this approach could
better inform congressional budget decision-making, and we modified a
portion of our recommendation accordingly. FEMA also provided technical
clarifications, which we have incorporated into the report as appropriate.

To address each of our objectives, we reviewed the appropriate laws and
regulations and FEMA's policies and procedures. We also interviewed
officials in (1) FEMA's Office of Financial Management, Response and
Recovery Directorate, and its Office of Inspector General; (2) FEMA's
Disaster Finance Center in Berryville, Virginia; and (3) FEMA's regional
office in Denton, Texas.

To determine how accurate and timely FEMA's estimates of funding
requirements for past and future disasters are, we (1) examined FEMA's
methodology for accounting for past and future disasters and (2) analyzed
data from FEMA's Integrated Financial Management Information System. We
interviewed officials in FEMA's Office of Financial Management to discuss
their methodology and assumptions used for estimating funding requirements.
We also tested a sample of financial data for several disasters to determine
the accuracy of the obligation data that are ultimately used for determining
fund requirements. We did not, however, conduct a reliability assessment of
FEMA's Integrated Financial Management Information System because a recent
assessment by FEMA's Office of Inspector General highlighted the strengths
and weaknesses of the system.10

To determine what impact recent FEMA initiatives have had on the rate of
obligating disaster relief funds, we (1) reviewed summary reports prepared
by each of the closeout teams, (2) interviewed officials assigned to the
closeout teams to ascertain their methodology for assisting in the closeout
of old disasters, and (3) interviewed officials in FEMA's Office of
Financial Management to identify changes in their assumptions used for
determining Disaster Relief Fund requirements. We reviewed FEMA's guidance
to its field offices and state emergency management offices concerning
recent programmatic changes. We also obtained and analyzed data on the rate
at which FEMA incurred obligations.

We performed our work from August 1999 through July 2000 in accordance with
generally accepted government auditing standards.

We are sending copies of this report to the appropriate congressional
committees; the Honorable James L. Witt, Director of the Federal Emergency
Management Agency; and the Honorable Jacob J. Lew, Director of the Office of
Management and Budget. We will also make copies available to others upon
request.

If you have any questions about this report, please contact Pat Moore or me
at (202) 512-7631. Key contributors to this assignment were Pat Valentine,
Rick Smith, and Thom Barger.
Stanley J. Czerwinski
Associate Director, Housing and
Community Development Issues

Comments From the Federal Emergency Management Agency

1. We disagree with FEMA's statement that our report overstates the
magnitude of the data discrepancies we uncovered in the Disaster Financial
Status Reports database. The difference between our views and FEMA's centers
on the significance of reporting collective numbers to indicate the absolute
value (total dollar amount) of differences between headquarters and regional
offices. FEMA questions the significance of this reporting, stating that
collective numbers do not affect the reports to the Congress. This is true
only if the collective numbers are derived from numbers that are correct. As
we stated in our report, FEMA officials could not tell us which amounts were
correct, and the agency acknowledges this. It is one of the reasons for
reporting the collective differences. Another reason is that we believe that
it is important to give some perspective as to the size of the total
differences in obligations--differences that affect a third of all past
disasters and range from as little as $1 to almost $25 million.

As a result of our work, FEMA used a new data extraction technique at the
end of March 2000 to produce a new set of IFMIS obligation amounts, by
disaster. In April of 2000, we requested that FEMA provide us with two
Disaster Financial Status Reports, one using the old extraction method and a
second using the new extraction method. A comparison of obligation amounts
in each of the reports would have provided an indication of how errors
caused by the old extraction method affected the reports to the Congress.
FEMA declined to provide the reports.

2. We disagree with FEMA's statement that a footnote in the draft report
suggests that the reports to the Congress are not materially wrong. The
footnote refers to the $36 million in corrections to obligation errors in
IFMIS that were made after FEMA developed a new method for extracting data
from IFMIS. Once FEMA identified and corrected these errors, it was
determined that the net amount in error did not materially affect FEMA's
financial statements. However, as FEMA acknowledges, the majority of errors
we uncovered were caused by problems that existed in the original method
FEMA used to extract data from IFMIS. As noted above, we have no way of
knowing the total amount of errors caused by the old extraction method, and
FEMA declined to provide us with the information needed to make this
determination.

3. We agree with FEMA that it can be both labor-intensive and time-consuming
to resolve differences in amounts reported by regional and headquarters
staff, especially given the large number of disasters involved. However,
FEMA's own guidelines for submission of the quarterly Disaster Financial
Status Report states that regions should work with FEMA's Office of
Financial Management and the respective Territorial Closeout Office to
ensure consistent reporting and to resolve any discrepancies in data errors.
The intent was to resolve the discrepancies in a timely manner. We found
that regional staff repeatedly reported many of these differences to FEMA's
Office of Financial Management, but no action was taken to resolve them
until we inquired about the discrepancies. If FEMA had followed its
guidelines and taken steps to correct the errors when they were first
reported, fewer differences would have been carried over from prior Disaster
Financial Status Reports, and the number of transactions and records needing
research would likely have been smaller.

4. As discussed in the report, we recognize that FEMA is developing a new
system for tracking disaster costs that is intended to reduce or eliminate
the types of discrepancies that occur in the current system.

5. We did not intend to imply, and did not state in our draft report, that
there are problems with standard IFMIS reports. Our report indicates that
FEMA determined in March 2000 that most of the differences in obligation
amounts were attributed to the extraction of data from IFMIS. The report
also discusses the fact that FEMA made $36 million in corrections to IFMIS
that were not, however, related to the extraction process. FEMA identified
the need for these corrections as a result of its use of a new extraction
method.

We disagree with FEMA's statement that "With the exception of the minor
discrepancies which related to refunds from States, which the report
references, the errors identified strictly relate to the previously-employed
logic behind the extraction of data from IFMIS for the DFSR." As we state in
our report, the problem with the state refunds not being reported in IFMIS,
which totaled $1.4 million, was just one example of errors found in the
IFMIS data. The $36 million in errors in IFMIS--representing differences
between headquarters and regional data--have now been corrected, according
to FEMA.

6. FEMA provides two reasons for why it believes our recommendation for
improving the method for projecting the timing and cost of future disasters
would not necessarily result in better forecasts of the costs of future
disasters. First, FEMA believes that the timing and cost of disasters are so
unpredictable that there is no discernible pattern from year to year.
Second, FEMA believes that it is better to use the 5-year average of annual
obligations, as opposed to the 5-year average of projected cost for
disasters as we recommend, because disaster cost data tend to fluctuate in
the year preceding the current one. While we agree that the timing and cost
of natural disasters are unpredictable, historically most hurricanes have
occurred during August and September. FEMA's methodology basically assumes
that no major hurricanes will occur in these months. This approach does not
seem reasonable when one considers that, as FEMA points out, major
hurricanes have occurred during this period in 5 of the last 11 years and in
3 out of 4 years since 1995. We do not agree, however, with FEMA's use of
the 5-year average of annual obligations to estimate future disaster costs.
While total cost projections stabilize after the first year, annual
obligations are influenced by obligations occurring for disasters that are
as much as 10 years old. In addition, as pointed out in our report, annual
obligations for fiscal years 1998 and 1999 have been significantly
influenced by the recent activities of the closeout teams. The teams focused
on a large backlog of old disasters, which have now been mostly closed out.
However, the effect on FEMA's 5-year average of obligations will be felt
over the next 4 years. It is also important to point out that FEMA
consistently uses disaster cost projection data in producing its monthly
reports to the Congress.

7. We agree that, given the uncertainty about the timing and cost of natural
disasters, it may be prudent for FEMA to provide a range of available
end-of-year balances in its monthly reports. The balances would be based on
different assumptions about if and when a disaster may occur and how costly
it might be. This approach could better inform congressional budget
decision-making. We modified our recommendation to reflect this issue.

(385815)

  

1. FEMA uses the term "remaining costs" to refer to estimates of the total
federal projected costs anticipated to fully fund a disaster's damages minus
obligations made to date for that disaster. Because of the time it takes to
identify and approve projects, remaining costs may be outstanding for
several years after a disaster occurs.

2. Public Law 102-229, enacted in 1991, limits FEMA's regular Disaster
Relief Fund appropriation to $320 million or the President's budget request,
whichever is lower.

3. For more information on federal funding for disasters, see Budgeting for
Emergencies: State Practices and Federal Implications (GAO/AIMD-99-250,
Sept. 30, 1999).

4. According to FEMA's Office of Inspector General, the $36 million would
not materially affect FEMA's financial statements because the net change to
IFMIS from correcting the errors was about $490,000--some errors were higher
and some were lower than the amount that should have originally been
recorded.

5. In reconciling the data, the staff found transactions in IFMIS that were
not recorded in the system FEMA uses to electronically transfer funds to the
states (i.e., the "Smartlink" system). The Department of Health and Human
Services operates this system.

6. This estimate is based on the Disaster Relief Fund's 5-year average level
of obligations (excluding the Northridge, California, earthquake), adjusted
for inflation--$2.991 billion for fiscal year 2000.

7. A disaster is considered "closed" when all of the funds for the disaster
have been obligated, thus reducing remaining costs to zero.

8. Originally there were three closeout teams (Eastern, Central, and
Western) whose "territories" were defined by regional boundaries. However,
because of its reduced workload, the Central Territorial Closeout team was
disbanded in May 1999 and the Eastern and Western Teams absorbed its
workload.

9. Review of the Territorial Closeout Teams, FEMA OIG, Audit Report H-02-00
(Dec. 22, 1999).

10. Review of FEMA's Integrated Financial Management Information System
(IFMIS), FEMA OIG, H-10-99 (July 15, 1999) and Management Letter on FEMA's
Fiscal Year 1999 Financial Statements, FEMA OIG, H-07-00 (June 5, 2000).
*** End of document. ***