Requests for federal disaster assistance need better evaluation report to the Congress

[From the U.S. Government Printing Office, www.gpo.gov]

BY THE COMPTROLLER GENERAL
0Report To The Congress
OF THE UNITED STATES

Requests For Federal Disaster Assistance
Need Better Evaluation

Since 1970 the President's Disaster Relief Fund has
provided over $3.8 billion to individuals and State
and local governments. The Federal Emergency
Management Agency, in assessing eligibility for
disaster assistance, has had difficulty in determin-
ing whether State and local governments are capable
of handling their own disasters and whether they
are committing a reasonable amount of their avail-
able funds for disaster relief. GAO also found that
the Agency has adopted a controversial cost-
sharing policy and has funded other than natural
disasters.
GAO recommends that the Congress clarify (1) the
extent of supplemental Federal assistance to be
given and (2) what kinds of disasters are eligible for
Federal assistance. The Federal Emergency Man-
agement Agency should analyze the impact po-
tential State inequities have on Federal disaster
assistance and submit a detailed plan and legis-
lative changes to the Congress to correct such
weaknesses.

~~ 551.3
. U55
: 1981 U .iV 0 : :CED-82-4
(W : -~~~~~~~~r~~~~~~ DECEMBER 7, 1981

!
,
5-
31 C .
.;'S

n ; .
Request for copies of GAO reports should be
sent to:
U.S. General Accounting Office
Document Handling and Information
Services Facility
P.O. Box 6015
Gaithersburg, Md. 20760
Telephone (202) 275-6241
The first five copies of individual reports are
free of charge. Additional copies of bound
audit reports are $3.25 each. Additional
copies of unbound re-port (i.e., letter reports)
and most other publications are $1.00 each.
There will be a 25% discount on all orders for
100 or mnore copies mailed to a single address.
Sales orders must be prepaid on a cash, check,
or money order basis. Check should be made
out to the "Superintendent of Documents".

~~~0LLER %~~~~~~6

COMPTROLLER GENERAL OF THE UNITED STATES
WASHINGTON D.C. 20548

B-i199765 .2

To the President of the Senate and the
Speaker of the House of Representatives
This report describes the difficulties the Federal Government
faces in assessing the eligibility of State and local governments
for supplemental disaster assistance. It recommends- that the
Congress'clarify the extent of disaster assistance provided to the
States and the Federal role in "non-natural" disasters. It also
recommends that the Federal Emergency Management Agency (FEMA)
take specific actions to improve its capability to evaluate
� ~disaster assistance requests.

This review addresses concerns arising from previous GAO
reviews regarding the disaster declaration process and FEMA's
.role in making recommendations to the President. This issue is
important because of the substantial Federal paymonts which can
result from this process. FEMA's proper assessment of requests
for disaster assistance is crucial to Presidential decisions to
grant or deny Federal disaster assistance.

we are sending copies of this report to the Director, Office
of Management and Budget; the Directpr, Federal Emergency Manage-
ment Agency; and interested congressional committees, subcommit-
tees, and individual Members of Congress.

.Comptroller General
of the United States

Li Sz. DEPARTMENT OF COMMERCE NOAA
COASTAL. SERVICES CENTER
rllr~l M I 2234 SOUTH HOBSON AVENUE
CHARLESTON, SC 29405-2413

-property Of csc Library

COMPTROLLER GENERAL'S REQUESTS FOR FEDERAL DIS-ASTER
REPORT TO THE CONGRESS ASSISTANCE NEED BETTER EVALUATION

D IG ES T

From fiscal year 1970 through April 1981, the
President provided individuals and State and
local governments with over $3.8 billion from
the President's Disaster Relief Fund. This
included 376 major disasters and 84 emergencies.
From the beginning of fiscal year 1979 through
April 198,1,, the President provided disaster
relief assistance of over $1.2 billion in 73
major disasters ar~id 17 emergencies. Since the
beg-inn ing ofifiscal year 1979 other Federal
agenci,es havle also authorized loans or provided
direc;t assistance totaling $1.6 billion.

GAO initiated this review to address concerns
arising from previous reviews regarding the dis-
aster declaration process and the Federal Emer-
gency Management Agency's (FEMA's) role in making
recommendations to the President. This issue is
important because of the substantial Federal
payment s which can result from this process.
GAO found problems existed in determining the
reasonableness of disaster assistance provided
by State and local governments. GAO also found
that FEMA has adopted a controversial cost-
sharing policy and has funded other than natural
disastets. (See pp. 7, 22, and 35.)

The Federal Disaster Relief Program is intended
to supplement the assistance States, their po-
litical subdi.visions, private relief organiza-
tions, and citizens provide for disaster relief.
If a catastrophe is of sufficient severity and
magnitude to warrant Federal assistance, the
President can declare a major disaster or emer-
gency. A broad range of assistance then becomes
available to individual disaster victims and
State and local governments. (See pp. I through
3.)

FEMA evaluates requests from State Governors for
assistance and recommends declarations or denials
to the President. Since the beginning of fiscal
year 1979, 90 major disaster and emergency re-
quests have been declared and 64 requests have
been denied. GAO undertook this review to deter-
mine the type and amount of information FEMA
obtains, the criteria it uses to evaluate the
requests, and the bases for its recommendations.
(See pp. 3 and 5.)
Tear Sheet CD8-
DECEMBER 7, 1981

GAO is making a number of recommendations to
the Director, FEMA, to ensure that each request
for major disaster or emergency assistance is
treated in a fair and equitable manner. GAO is
also recommending that the Congress clarify a
numbet of issues so that FEMA can better admin-
ister the law according to the Congress' intent.
(See pp. 19, 33, 34, 38 and 40.)
DETERMINING REASONABLENESS OF
STATE AND LOCAL GOVERNMENT ASSISTANCE
When FEMA receives a major disaster or emergency
request for assistance from a State Governor, it
needs to quickly evaluate the situation and
determine whether a declaration should be made.
These requests vary greatly as to form and con-
tent and do not always fulfill the requirements
of law or contain all of the types of information
FEMA uses to evaluate requests. (See pp. 8, 22,
24, and 28 to 31.)
FEMA assesses the severity and magnitude of the
situation and the capability and resource com-
mitments of the State and local governments and
makes its recommendation to the President. As-
sessment of severity and magnitude are reason-
ably consistent but those of capability and
commitment vary greatly because of the type,l
quality, and degree of information considered.
GAO's review Of 31 disaster requests disclosed
that FEMA uses a wide range of information in
arriving at its decisions. For example, FEMA
considered "State surplus" information in only
seven requests. The'lack of Consistency in the
quality and method of assessments and the lack
of knowledge by others,as to FEMA's methods of
evaluation can create doubt as to whether the
Federal Government is only providing supplemen-
tary assistance and whether each,request is
judged in a fair and equitable manner. (See
pp. 7, 22, and 28 through 32.)
FEMA's policies, procedures, and guidelines for
evaluating requests are not widely known. Dis-
closing internal assessment processes would help
State ;and local governments decide whether they
had a valid request to make, enable them to pro-
vide more complete and uniform ihformation, and
minimize doubts as to whether their requests are
treated in a fair and equitable manner. (See
pp. 32 and 33.)
ii

GAO is -recommendinci that the Director of FEMA
reevaluate and -improve its assessment criteria;
publish the policies, procedures, and guidelines
which FEMA uses to evaluate requests in the
Federa'l-Register;1 develop comprehensive, uniform
f-orms tor State and Federal officials to use;
require,Gov'ernors to provide the information
necessary for the evaluations; and make it clear
that future requests which fully comply with
Feder'al laws,,and regulations will help avoid
delays in processing the requests. (See pp. 19,
33," and 34..)

GAO is recommending that the Congress direct
FEMA to prepare a comprehensive analysis of the
impact potential State inequities have on Fed-
eral disaster assistance.and submit a detailed
plan and legislative changes to the Congress to
correct such weaknesses. GAO made this same
recommendation to FEMA in 1979. I/ (See p. 34.)

THE CONGRESS NEEDS TO CLARIFY
DISASTER ASSISTANCE ISSUES

Two recent actions have raised concerns that
FEMA may have changed the scope of the Disaster
Relief Act. First, in May 1980, FEMA adopted a
gene'ral policy of requiring State and local
governments to agree to pay 25 percent of the
eligible costs of public assistance programs.
This 'policy removes the administrative problems
ass~ociated with attempting to determine a
"reasonable" commitment for each disaster and
assures compliance with the statutory purpose
that the Federal aid be supplemental.

Although FEMA's cost sharing policy is consistent
with the act, it has created controversy among
the States. State officials contend that the
policy forces them to pay for disaster relief
costs which the States believe are beyond their
capability to assume or which constitute more
than a'reasonable amount of State and local
funds. Because of the controversy, GAO believes
that the Congress needs to clarify its intent on
this matter.

Second, the President has provided disaster
assistance for such events as the Love Canal
chemical contamination and the Cuban refugee

1/"Federal Snow Removal Reimbursement Policy:
Improvements Needed" (CED-79-97, Aug. 2, 1979).
Tear Sheet
l
iii

influx in Florida. This action has stirred
considerable controversy as to whether other
than natural catastrophes are within the pur-
view of existing law. Although congressional
intent is not clear on this matter, the act
does allow the President to make declarations
for other catastrophes. (See pp. 35 through
40.)

GAO is recommending that the Congress reevaluate
the act and clarify its intent with regard to
the extent to which supplemental Federal assis-
tance should be given in a major disaster and
the types of incidents that may receive disaster
assistance. (See pp. 38 and 40.)
AGENCY COMMENTS
The Director, FEMA, said this report has a
number of worthwhile comments regarding the
problems associated with State and local govern-
ment commitments, State and local government
capability, and cost sharing. He said most of
GAO's recommendations will be considered during
FEMA's current revision of a regulation relating
to the declaration process. The Director did
not agree to GAO's proposal that FEMA use com-
puter modeling as a tool for program decision-
making and evaluation. (See app. III.)
Shortly after receipt of the Director's comments,
FEMA established a plan for considering all of
GAO's recommendations, including a review of
computer applications, with final regulatory
action to be completed by September 30, 1982.
Until the regulation is issued and other actions
are completed, GAO has no basis for evaluating
the Director's response to its recommendations
or the FEMA plan. GAO continues to believe that
the.adoption and use of computer modeling would
aid FEMA in (1) developing more uniform data,
(2) refining assessment criteria, (3) providing
more uniformity in program evaluations, and (4)
giving FEMA officials better management control
over the operation of its disaster assistance
program.
The Director's comments and GAO's evaluations
are also contained on pp. 10, 13, 19, 20, 21,
27, 31, 32, 34, 38, and 39.
iv

I
C on te nt s

-. ~~~~~~~Page

DIGESTi
CHAPTER
INTRODUCTION,1
Disaster Relief Program 2
Organization and,management of the
Disaster Relief Program 4
Objectives, scope, and methodology 5
2 OPPORTUNITIES TO IMPROVE DATA GATHERING
AND DISASTER ASSESSMENTS 7
Assessing need for Federal assistance 7
Problems in reviewing FEMA assessments a
S-tatistical analysis indicates
consistency in FEMA decisions 9
Conclusions 19
Recommendations 19
Agency comments and our evaluation 19

3 IMPROVEMENTS SHOULD BE MADE IN EVALUATING
THE ABILITY OF STATE AND LOCAL
GOVERNMENTS TO HANDLE MAJOR DISASTERS AND
EMERGENC IES 22
Limited guidance for State and local
governments 22
Evaluation of the capability and commit-
ment of State and local governments not
conclusive 24
States' financial capabilities need to
be adequately assessed 31
Conclusions 32
Recommendations 33
Agency comments and our evaluation 34
Recommendation to the Congress 34

4 THE CONGRESS NEEDS TO CLARIFY
DISASTER ASSISTANCE ISSUES 35
FEMA cost-sharing policy is consistent
with the act 35
Conclusion 38
Recommendation to the Congress 38
Agency comments and our evaluation 38
Using the Disaster Relief Act for other
than natural catastrophes needs
clarification 39
Conclusion 40
Recommendation to the Congress 40

I
Page
APPENDIX
I

II
III

DR& R
FEMA-
GAO
Major disasters? emergencies, and denials
reviewed by GAO
Technical analyses

Letter dated September 29, 1981, from the
Director, Federal Emergency Management Agency
ABBREVIATIONS
Disaster Response and Recovery
Federal Emergency Management Agency
General Accounting Office
41
43

CHAPTER 1

INTRODUCTION
In 1950 the Congress enacted the first comprehensive disaster
relief program. This program authorized the President to provide
assistance for the temporary repair of local governments' public
facilities. Subsequent laws have extended this assistance to
include individuals and State governments and have broadened the
extent of coverage to include permanent repairs of public facil-
ities and other items. The present law contains a wide range
of grants and direct assistance programs for individuals and
State and local governments.
From the beginning of fiscal year 1970 through April 30,
1981, the President received 729 requests for major disaster or
emergency assistance. Federal assistance of $3.8 billion was
provided under the President's program for the 376 major disas-
ters and 84 emergencies that were declared. Figure 1 compares
requests with Presidential declarations of major disasters and
emergencies from calendar year 1974 through calendar year 1980.

FIGURE 1
PRESIDENTIAL DECLARATIONS OF MAJOR DISASTERS
AND EMERGENCI(ES
CALENDAR YEARS 1974 THROUGH 1980
I '7n
74 5 76l 77 78 79 80
Source: Federal Emergency Management Agency
1

DISASTER RELIEF PROGRAM
When a disaster threatens or occurs, local authorities
take immediate steps to warn and evacuate citizens, alleviate
suffering, anad protect life and property. if additional help
is needed, the Governor may exe-cute the State's emergency plan
or commit State resources.

The President's Disaster Relief Program is intended to
supplement the assistance provided by the States, their political
subdivisions, Private relief organizations, and citizens. When
Governors believe that supplemental Federal assistance is neces-
sary they may request that the Pvesident declare a "major dis-
aster" or an "emergency" to implement the provisions of Public
Law 93-288, the Disaster Relief Act of 1974. This act defines
major disaster as any:
11** *hurricane, tornado, storma, flood, high
water, wind-driven water, tidal wave, tsunami,
earthquake', volcanic eruption, landslide,
mnudslide, snowstorm, drought, fire, explosion,
or other catastrophe in any part of the United
States which, in the determination of the
President, causes damage of sufficient sever-
ity and magnitude to warrant major disaster
assistance under this Act, above and beyond
emergency services by the Federal Government,
to supplement the efforts and available re-
sources of States, local governments, and
disaster relief organizations in alleviating
the damage, loss, hardship or suffering caused
thereby."

in certain instances, the full range of assistance available
with a major disaster declaration may not be required. in such
cases, the Governor may decide to request the President to declare
an emergency, which would provide specialized assistance to meet
a specific need that the Federal Government is uniquely able to
provide. An emergency is any of.the various types of disasters
included in the definition of a major disaster which requires
Federal emergency assistance to-supplement State and local efforts
to save lives; protect property, public health, and safety; or
avert or lessen the threat of a disaster.

The Governor's request for a major disaster declaration by
the President must be 'based on a finding that the situation is
of such severity and magnitude that effective response is beyond
the State's and affected local governments' capabilities and that
Federal assistance is necessary. The Governor must certify that
a reasonable amount of State and local funds had been or will be
used (of which State commitments must be a significant propor-
tion) . The "Governor must also estimate the extent and nature of
Federal assistance required for each of the disaster-affected
areas. in conjunction with the Governor's request for a major
,disaster declaration, the Governor must take appropriate action
under State law and direct the execution of the State's emergency
plan.
2

A P.residential declaration of a major disaster makes a broad
range 6f Fed'eral assistanc-e available to individual: disaster vic-
tims and State and local governments. This help may include
temporary housing, minimum essential repairs to residences,
disaster unemployment assistance, food stamps, individual and
family grants to meet disaster-related necessary expenses, and
various other services. Assistance to governmental jurisdictions
may be directed at debris removal; emergency protective measures;
or repair or replacement of roads, streets, bridges, water
control facilities, and other public facilities.

After the President declares a major disaster or emergency,
the Federal Emergency Management Agency (FEMA), based upon the
Governor's request, designates the areas within the State that
will receive assistance and specifies the type of assistance that
will be provided.

Program statistics

The President may approve or deny a Governor's request for
a major disaster or emergency declaration. The following table
identifies the disposition of all requests for fiscal years 1979,
1980, and 1981 through April 30.

1979 1980 1981 Total

(7 months)
Major disasters:
Declarea 42 25 6 73
Denied 6 29 11 46

Total 48 54 17 119
Emergencies:
D Declared 11 5 1 17
Denied - 9 5 4 18

Total 20 10 5 35

Total requests 68 64 22 154

Once.the' President makes a declaration, FEMA has authority
to spend fundsifor disaster. assistance from the President's
Disaste.r..Relief Fund. In addition to the President's Disaster
Relief Fu'nd, FEMA can call upon other Federal agencies to provide
disaste'r assistance with or without reimbursement to the agencies
from the President's Fund. Federal agencies can also provide
assistance under their own authorities.
3

The table below identifies estimated President's,Fund
requirements for major disasters and emergencies declared for
fiscal years 1979, 1980, and 1981 through April 30.

1979 1980 1981 Total

(7 months)
�-------------- (000 omitted)�

Major disasters
Individual
assistance $199,162 $ 60,518 $ 3,110 $ 262,790
Public
assistance 418,788 239,439 29,846 688,073

Total 617,950 299,957 32,956 950,863

Emergencies 29,894 a/233,487 2,467 265,848

Total $647,844 $533,444 $35,423 $1,216,711
a_/Of this amount, $191 million is for the Cuban refugee emergency.
FEMA estimated that other Federal agencies, such as the Small
Business Administration, Farmers Home Administration, and the
Corps of Engineers, authorized loans or provided other assistance
totaling $1.6 billion. During this period, FEMA was authorized
to spend $21.4 million for administering the disaster relief
program.

ORGANIZATION AND MANAGEMENT
OF THE DISASTER RELIEF PROGRAM

FEMA's Office of Disaster Response and Recovery (DR&R) 1/
is responsible for implementing the President's Disaster Re-
lief Program. DR&R is headed by an Associate Director, who
reports directly to the Director, FEMA.

DR&R staff are in each of FEMA's 10 regional offices.
In May 1981, DR&R had 154 staff members, 64 at headquarters and
90 in the regional offices. When a major disaster or emergency
request is received, regional personnel assess the damage and
evaluate the need for Federal assistance. The region makes its
recommendation to headquarters, which reviews the data and makes
its recommendation to the President. When a request is granted,
regional personnel coordinate response and recovery activities.

I/On June 5, 1981, after the completion of our review, FEMA
announced the first step in reorganizing the agency. The
organization discussed in this report is that which existed
at the time of our review.
J

I
I
4

i
DR&R augments its permanent staff with "reservists" who are
employed by the region. These reservists are disaster assistance
specialists_ who, as.sist DR&R in carrying out its responsibilities
at the disaster site.

OBJECTIVES, SCOPE, AND METHODOLOGY
We> initiated this assignment to address concerns arising
from previous GAO reviews regarding the disaster declaration pro-
cess. and FEMA's role. in making recommendations to the President.
This is:sue i's important because of the substantial Federal pay-
ments wh'ich can, result fr.om this process. The review objectives
were to .'determine the type and amount of information FEMA obtains
on maj;or disaster and emergency requests and to evaluate the
criteria it 'used to make recommendations for declarations and
denials.

Our review was conducted from April 1980 through May 1981
primarily 'at FEMA headquarters in Washington, D.C. We also re-
viewed records and interviewed officials in FEMA's Atlanta Re-
gional Office. We interviewed the State Deputy Director of Civil
Defense in Atlanta, Georgia, and the State Director, Department
of.Civil Defense, in Montgomery, Alabama. We also discussed the
disaster. declar-ation process with officials of the National
Governors Association, Washington, D.C., and obtained their
views on establishing criteria for a declaration.

In,.February 1981, Congressman Steve Gunderson of Wisconsin
requested that our Office determine whether FEMA was administer-
ing-disaster assistance programs consistently, effectively, and
according. to congressional. intent. In this regard, the Admin-
istrator.,.Division of Emergency Government, Madison, Wisconsin,
provided us with certain documents and records. We discussed
with Congressman Gunderson his specific questions about the
disaster assistance program and advised him that this review and
our ongoing review of Federal disaster assistance provided to
State and local communities would address his concerns.

Senator Edward Zorinsky expressed his support for this review
and voiced his concern regarding FEMA's cost-sharing policy of
reimbursing State and local governments for only 75 percent of
public assistance costs. We advised Senator Zorinsky that this
question would be addressed in our report.

We reviewed the Disaster Relief Act of 1974 (Public Law
93-288), related FEMA regulations, and obtained legal opinions
from our Office of General Counsel on several issues.

We analyzed disaster records and statistics to identify
important areas and issues for further consideration. We made
comparative analyses by regions and States and by similar types
of disasters, locations, time periods, and dollar size. From
these analyses, we selected and reviewed 24 major disasters, 8
emergencies, and 9 requests which were denied. These reviews
involved 26 different States and 9 FEMA regions. Because large
5

major disasters, such as Mt. St. Helens, would need less
documentation to support a declaration, we concentrated our eval-.
uation on the smaller disasters. Although I major disaster we
reviewed required an estimated $129 million from the President'a
Disaster Relief Fund and 1 emergency about $21 million, the
remaining 23 major disasters averaged $3.2 million and the re-
maining 7 emergencies averaged about $262,000. (See app. I for
a list of major disasters, emergencies, and denials revliewed by
GAO.)
Some general criteria for evaluating requests are contained
in the act. However, FEMA advised us that it had developed no
additional specific criteria for evaluating requests for assis-
tance as each request is different and has to be subjectively
evaluated on the information available. We analyzed the data con-
tained in Governors' letters of request, regional assessments,
and headquarters' reviews, in an effort to identify th.e unwritten
policies, procedures, and criteria FEMA follows in making recom-
mendations to the President. Regional records pertinent to our
review were contained in FEMA's headquarters' files. We provided
written questions to FEMA on a number of its reviews of Governors'
requests and obtained written.responses for most of them.

In an effort to better understand the decisionmaking process,
we.did a statistical and computer analysis of the quantifiable
,data on 96 of the 102 major disaster requests that were declared
and denied in fiscal years 1979 and 1980. We made correlation
analyses to identify those disaster factors which had the most
influence on the final decision. we determined averages for sig-
nificant factors such as numbers of people injured, homes damaged,
estimated eligible costs, and others. We made numerous multiple
regression analyses to identify those declarations and.denials
which did not conform with the majority of-these decisions.
Appendix II describes our method and analyses in more detail.

CHAPTER 2

OPPORTUNITIES TO IMPROVE DATA GATHiERING

AND DISASTER ASSESSMENTS

Federal assistance is supplementary in nature and to be
provided only when the severity and magnitude of the incident is
beyond the capabilities of the State and affected local govern-
ments to effectively respond. It is to be given only when the
States have committed a reasonable amount of their funds to al-
l1eviate the damage, loss, hardship, and suffering resulting from
the incident.

Assessing the severity and-magnitude of disasters and evalu-
ating the need for Federal assistance quickly is a difficult task
which involves making decisions with incomplete information. FEMA
.decisions are consistent where the severity and magnitude of dis-
asters is concerned., But its decisions are not as consistent when
evaluating data related to States' and local governments' capa-
bilities and ~their. commitments to handle the disaster. (See ch.
3.)

Standard forms and computer models, such as those developed
by GAO, would improve FEMA's ability to gather data and assess
the need for Federal assistance.

ASSESSING NEED FOR

FEDERAL_ASSISTANCE

It was the Congress' intent to provide supplemental assis-
tance to State and local governments to help (1) alleviate the
damage, loss, hardship, or-suffering caused by a major disaster,
(2) save lives and protect property and public health, and (3)
maintain safety in the event of an emergency..

FEMA's internal guidelines state that the Disaster Relief
Act, prior disaster legislation, legislative history, and ex-
perience all provide some direc.tion but do not define the param-
eters of a major~ disaster. A principal component in FEMAI's
assessment is a determination that a disaster is of such severity
and magnitude that effective response is beyond the capabilities
of State and local governments and that Federal assistance,
authorized by the act or by other Federal programs triggered by
~the act, is required.

A key problem for FEMA. is measuring severity and magnitude.
For evaluation purposes, FEMA's objective is to quantify it in
terms of damages and1losses, which are further related to specific
actions required to alleviate hardship and suffering. FEMA per-
sonnel.were advised to measure severity and magnitude by consid-
ering (1) what the"State, local governments, and individuals will
have'to do if a declaration is made and (2) what the impact will
be on the State, local governments, and individuals if a major
disaster is not declared.
7

FEMA's ultimate objective is to determine the "unmet needs"
for which the Federal Government can provide assistance and the
impact which the disaster has on those affected by it. FEMA sees
each disaster as being unique and, therefore, needing to be
evaluated on its own merits.
PROBLEMS IN REVIEWINGI

A Governor's request to the President for a major disaster
or emergency declaration is sent to the appropriate FEMA Regional
Director. This request contains,' among other things,.the State'-s
assessment of damages and losses and the actions being taken to
overcome them. Regional personnel visit the scene of the incident
to make a first-hand evaluation of the damage. Based on this
review, the Regional Director makes a recommendation to FEMA head-
quarters and the FEMA Director recommends a course of,action to
the President.

For reporting to headquarters, FEMA has developed a regional
disaster/emergency summary for the gathering of data about the
incident and a separate regional analysis and, recommendation for
opinions and evaluations. The summary provides information on
the date and type of request and locations and probable incident
period and an assessment of severity, magnitude, and response
capabilities of State and local governments and private organiza-
tions. The assessment section includes such factors as numbers
of casualties (dead and injured), number of homes damaged and
destroyed, numbers and types of businesses and farms affected and
related damage estimates, and impact on public facilities. These
reports are reviewed by headquarters personnel in DR&R.

To better understand the damage assessment and evaluation
process performed by FEMA, we selected a sample of requests that
resulted-in 24 major disaster declarations, 8 emergency declara-
tions, and 9 denials. We made a detailed review of the requests
and analyses of the facts, conclusions, and recommendations
contained in FEMA's reports.

We found it difficult to evaluate the decisions on individual
requests or on different requests which appeared to us to be sim-
ilar in size, scope, time, or other circumstances. Each request
had different circumstances,,although the basic types of informa-
tion were similar. Governors' letters presented the information
in a variety of ways. many regional disaster assessments did not
always present the informatio.n in a logical, uniform manner, and
subjective opinions and conclusions were made on an overall basis
without specific reference to the data on which they were based.

Many cases that we reviewed resulted in questions for which
we could not find answers in the files. We requested in writing,
that FEMA supply written replies to our questions. Those which
FEMA did answer or explain were based on opinions, recollections,
and undocumented information which welcould not verify. The,
results of our analyses of individual requests are discussed on
pages 23 to 31.
8

I I
i

I
STATISTICAL ANALYSIS INDICATES
CONSISTENCY IN FEMA DECISIONS

Because of the difficulties we encountered in reviewing
requests, we Undertook a detailed statistical analysis of more
cases in an effort to discover the bases for FEMA's decisions in
granting or denying assistance-and to determine whether a more
objective approach could be taken in making disaster assessments.
If we could identify FEMA's criteria, we also wanted to evaluate
the extent they were consistently applied.

Method used

We selected 96 major disaster requests received in fiscal
years 1979 and 1980 as our data base. I/ We identified about 30
factors for these cases, such as number of people killed and in-
jured; number of homes damaged.and destroyed; and the dollar value
of homes, businesses, and farms damaged and destroyed. Data on
some of these factors was ofteni not available or was ambiguous
so we had to interpret some of the data. For example, when no
information existed-on the number of people injured or homes des-
troyed, we assumed none were. On the other hand, some obviously
large disasters had no data on the.dollar estimate of eligible
costs. For this type of missing data, we used recent FEMA cost
estimates. in addition to data from the files, we also used
published preliminary 1979 data on per capita income and State
operating funds as indicators of the State's capability to handle
the disaster on its own.

We were unab le to identify data on many individual factors
in enough cases to use in our analyses. Miany cases, for example,
did not provide estimates of the cost of business damage or the
cost of agriculture damage so we could not use them. As a result,
we were limited to using the estimat-ed cost of total damage and
the other factors discussed in this report. See appendix II for
more information on the data base and how we handled the problem
of missing data.

Two methods were used to.analyze the relationship between
FEMA's decisions and the factors identified for the disasters.
First, we attempted to identify differences between declared
and denied decisions in terms of different individual factors.
Second, we used more sophisticated statistical techniques to
develop models of the relationship between several factors
working in combination with FEMA's decisions. These are proba-
bility models that can identify past decisions that were not con-
sistent with the majority of the decisions made and can predict
how FEMA would decide a current request based on past decisions.
See appendix II for more details on these models.

I/During the period, a total of 102 major disaster requests were
received.. To preclude duplication in our analyses, we excluded
six requests for reconsideration of denials because they con-
tained no new information.
9

I
Factors influencing
FEMA decisions
Factors relating to disaster severity and magnitude, such as
the total estimated cost of eligible assistance, number of people
injured, and type of request, seem to influence FEMA's decisions.
The impact on FEMA decisions of other factors, such as States'
capability and commitment, are not as clear.
Estimated cost of
eligible assistance
The higher the estimate of eligible costs, the greater the
chance of a request being granted. The average cost in 96 appli-
cations was $7.2 million, ranging from about $100,000 to $177.5
million. The average cost in declared disasters was $9.83 million,
while it was only $1.12 million in denied disaster requests.
Table 1 shows the distribution of estimated eligible cost
by type of decision. For example, only 45 percent of the re-
quests involving up to $1 million were declared, while 100 per-
cent of the requests for $5 million or more were declared.
Table 1

Requested Decision Percent of total
amount Declared Denied Total Declared Denied
(millions)
$0-.9 13 16 29 45 55
1-1.9 9 7 16 56 44
2-2.9 8 4 12 67 33
3-4.9 13 2 15 87 13
5-8.9 10 0 10 100 0
9 and above 14 0. 14 100 0
Total 67 29 96
FEMA's Director contends that we are not entirely correct in
stating that the higher the estimate of eligible costs the greater
the chance of a request being granted. He said that the statement
would generally be true for the same State and type of catastrophe
and under similar circumstances.
FEMA suggests that we take more directly into account the
influences of several factors working in combination with each
other. It is true that table 1 analyzes only the influence of
estimated cost of eligible assistance. However, when we used
more sophisticated models to determine the relationship between
several factors working in combination with FEMA's decisions
(see p. 15.), the result was that the estimated cost of eligible
assistance still had a significant influence on the decision to
declare or deny a major disaster request.
10

Number of people injured

Generally, the more people injured, the greater the chance
of a disaster being declared. The average number of people in-
jured for all requests was 27, 37 for declared requests, and 1
for denied requests.

Table 2 shows the distribution by number of people injured.
There were 51 requests that did not report any injuries, 57 per-
cent of these were declared. On the other hand, all requests
reporting 50 or more injured were declared.
Table 2

I
I

I
Number
of people
inj ured
0
1-49
50-99
100-199
200 and
above
Total
Decision
Declared Denied
Percent of total
Declared Denied
Total
51
30
8
5
2
29
23
8
5
2
67
22
7
0
0
0
29
57
77
100
100
43
23
0
0
100
0
96
Homes damaged
The decision is not affected when the number of homes damaged
(major damage, minor damage, or completely destroyed) is less than
200. However, a disaster is more apt to be declared if the number
of homes damaged is greater than 200. Table 3 shows the distri-
bution of homes damaged by type of decision.

Table 3
Percent of total
Declared Denied
Number of
homes
Decision
Denied
Total
Declared
0
. 1-49
50-199
20 0-799
800 and
above
11
6
7
21
22
12
6
8
3
0
29
23
12
15
24
22
96
48
50
47
88
52
50
53
12
100
0
Total 67
11

Type of request
The chance of a request being declared or denied is closely
related to the type of assistance requested--individual assis-
tance, public assistance, or both. Requests for both types of
assistance have the greatest chance of being declared, while re-
quests fo r only public assistance have the least chance of being
declared. Requests for only individual or only public assistance
received only the kind of assistance requested. Requests for
both types of assistance received either both or individual
assistance only, but seldom received public assistance only.
Table 4 shows the distribution of type of assistance requested
by type of decision.

Table 4

Type of
assistance Decision Percent of total
requested Declared Denied - Total Declared Denied
Individual 3 5 a 38 62
Public I 11 12 8 92
Both A/63 13 76 83 17
Total 67 29 96 70 30
a/Of the 63 declarations that requested both types of assistance,
22 were declared for individual assistance only, 3 were de-
clared for public assistance only, and 38 were declared for
both individual and public assistance.

state cajpabilitv

We used State general operating funds and per capita income
as measures or indicators of a State's capability to financially
handle its disaster. We expected the chance of a major disaster
declaration to be greater for those States where these indicators
of capability were smaller. However, we found little difference
between average State operating funds and FEMA's decision to
declare or deny disasters.

Table 5 shows the distribution of%requests from States
according to their per capita income'. Nothing clearly indicates
how per capita income.influenced FEMA decisions. States in the
lowest per capita income category had a greater chance of their
request being approved than States with the highest income. But,
for the bulk of requests from States in the midranges, increases
in per capita income slightly increased the chances of approval.
12

I
Table 5

per capita Decision ___Percent of total
income Declared ]Denied Total Declared Denied
(thousands)
~$1-4.9 3 0 3 10 0 0
5-6.9 10 5 15 67 33
7-7.9 15 6 21 71 29
8-8.9 24 9 33 73 27
.9 and above 15 9 24 63 37
Total 67 29 96

State commitment
We looked at two indicators of a State's commitment to
disaster assistance. One measure was the amount of State and
local government funds the Governor said would be committed
to the disaster. The other was whether the Gove'rnor properly
certified that these funds would be available. We found little,
if any, difference between the averages for these factors for
declared or denied disaster request.s.

Table 6 shows the distribution of requests by commitment
statements and by type of decision. There were 63 requests
from States whose Governors did make the required commitment
statement and 33 requests from States whose Governors did not
make the required comnmitment statement. Equal percentages of
t-hese requests were declared. This indicates that the presence
or absence of a commitment statement had no definite impact on
the final decision.
Table 6

Made
commitmeint Decision Percent of total
statement Declared Denied Total Declared Denied

Yes 44 19 63 70 30
No 23 10 33 70 30
Total 67 29 96

FEMA's Director noted that for some disasters in the past,
a declaration was made with the expectation of working out the
specific State commitment and including it in the Federal-State
agreement. However, he said these instances were rare and gen-
erally limited to situations where immediate Federal assistance
was required.

t
FEMA's comment implies that our use of Governors' letters
as a basis for our analysis may have resulted in our overlooking
some State commitments. Our analysis was based on whether the
Governor fulfilled the statutory predeclarat!on requirement and
cextified as to State and local government commitments. If we
included those commitments made after-the-fact in our analysis,
the results should not change significantly because, as FEMA
indicates, the number of'instances was rare.

Time period
The relationship between when the decision was made and the
type of decision is not entirely clear. Figure 2 shows the num-
ber of disasters declared and denied for each quarter of fiscal
years 1979 and 1980. Toward the end of fiscal year 1979 decla-
rations increased, perhaps indicating more severe weather con-
ditions or a more liberal administration policy. Denials in-
creased thereafter, perhaps indicating less severe weather con-
ditions, a shortage of Federal funds, or a more conservative
administration policy.
In our regression analyses, time was used as a proxy for
inflation. See discussion in appendix II, page 47.

FIGURE 2

MAJOR DISASTER DECLARATIONS REQUESTED, DECLARED, AND DENIED

20 -

19 -

18 -

17- --
167- - REQUESTED
16 -

15-
14 .
LU
13 1

12-

10- /\~~~~~~~~~~~~DECLARED
10-
09- 1
/I /..-

6 I
5-~~~~~~~~~. DE.lE
12/78 ' 3/79 6/79 9/79 12/79 3/80 6/80 9180
TIME PERIOD
14

i
Political influence
We performed a.limited analysis of the relationship between
the decision and the political party i n power in each affected
area. We.compared the political parties of the Governors, Sen-
ator's, and'Representatives of the affected areas with the deci-
sions to declare and deny disaster requests.

Except for State Governors, we could find little, if any,
indication that political factors influenced the outcome. Table
7 shows the distribution of the Governors' political affiliation
by type of decision. on the surface it looks like the opposition
(Republican) party had a better chance of having its requests
granted.
Table 7

Governors'
political Decision Percent of total
affiliation Declared Denied Total Declared Denied

Democrat '41 23 64 64 36
Republican 22 6 28 79 21
Other or missing 4 0 4 100 0

Total 67 29 96

Influence of combining factors

The analysis up to this point has looked at the impact of
each factor individually. In practice, FEMA subjectively eval-
uates each request on the basis of combinations of factors. To
eval'uate this combined effect, we used a statistical technique
called regression analysis.

Using data from 96 requests and regression analyses, we de-
veloped mathematical models that estimate the relationship between
the decision and various combinations of factors,.such as people
injured, homes damaged, and State per capita income. The models
can also be used to assist in the evaluation of new requests.
New information could be entered into the formula, and it could
then be used to predict the probability of the request being
granted. Additional nonquantifiable data could then be used to
make the final decision. (See app. II.)

We developed six different models using different combina-
tions of factors, as shown in table 8. Some models appear to
have better predictive power than others, as noted by the per-
centage of decisions that the model agreed with the FEMA decision.
in general, the models support the inferences one would draw from
the previous tables. The models indicate that the estimated
eligible cost in dollars, in combination with numbet of homes
damaged and,number of peop,le injured, are positively associated
with the decisions. The higher their values the greater the
chance a request will be approved. Requests for combination of
I1
5

both individual and public assistance have the best chance of
being declared. Month of decision and per capita income are
negatively associated with FEMA's decisions. For example, as per
capita income increases, the chance of a request being approved
decreases. Also, there appears to be little, if any, relationship
between State commitment and the decision.
Table 8
Variables Used in
Regression Models
Model Number
1 2 3 4
5
6
Variables:
Type of assistance
requested
x
x x x
x
Estimated eligible
cost
x x x x x x
Number of people
injured
x
x x x x
x
x x x
Number of homes
damaged
x x x
State per capita
income
Decision month
State commitment

Interaction term with
several variables
Number of cases used
x x x x x x
x
x x
x
x
x x
x x

96 96 96 a/87 b/67 c/20
Percent of cases
agreeing with FEMA
decision
85 86 83 83 81 95
a/Deleted nine very large disasters with missing values for
selected variables.
b/Includes only requests for both individual and public assistance
and deletion of nine large disasters.

c/Includes only requests for individual or for public assistance.
16

The models were useful in providing insights about FEMA's
implicit decision criteria and the extent that FEMA applied the
criteria consistently.
Since FEMA seems to be more willing to grant certain types
of assistance than others, we tried to capture this additional
factor in other statistical models. These models focused pri-
marily on the type of assistance granted and were also used to
identify cases where FEMA may have made inconsistent decisions.
The six original models plus these additional models iden-
tified 20 decisions where FEMA may have applied the criteria
inconsistently. FEMA approved seven requests which our analyses
predicted would have normally been denied. Conversely, FEMA
denied 13 requests which our analyses predicted would have
normally been declared. Part of the reason for these differences
could be due to the limitations of our analyses or the data used.

our further analysis of the quantifiable data related to
these 20 decisions disclosed 8 decisions which were very incon-
sistent with the majority of FEMA's decisions. Table 9 lists
these decisions and c.ompares the values of key indicators for
each type of assistance granted or denied.. The first five re-
quests which were declared generally had less than average values
for three key indicators of disaster magnitude--estimated eli-
gible costs, homes destroyed, and people,,injured. In three of
these requests, the per-capita income--an indicator of financial
capability--is higher than the average. FEMA also denied three
requests for which the estimated eligible costs, homes damaged,
and people injured were generally higher than the average. Only
one of the denials had per capita income higher than the average.

We cannot judge whether FEMA made improper decisions in any
of the requests. Our review of the records for the 20 decisions
with which out models disagreed disclosed a variety of subjective
reasons for FEMA's'final decisions. While many FEMA decisions
appeared to be proper, some were not as clearcut because each
decision was based upon various criteria that were not uniformly
applied. One request for public assistance, for example, was
granted because the disaster had weakened State and local fiscal
resources so that it lacked the ability to respond to other
emergencies in the foreseeable future. While we do not question
the validity of this particular declaration, we did note that
such forward thinking was not applied to any of the other requests
that we reviewed.
t

I
1
7

Table 9
Inconsistent FEMA Decisions Identified by GAO Model
Number of homes Numb
Estimated eligible cost damaged and destroyed people
Average for For this Average for For this Average
fc
this type decision this type decision this type
..... (millions)-------
$ 5.90 $ .61 483 40 13
5.90 4.86 483 555 13
3.98 .44 1,092 450 22
14.09 2.00 858 312 50
5.90 10.80 483 0 13
1.12 1.35 73 334 1
1.12 1.43 73 251 1
1.12 1.67 73 93 1

bfr of
injured
or For this
e decision

4
0
0
0
0
3
0
3
Type of
assistance
Decision granted or
number denied

3 Public
6 Public
9 Individual
23 Both
24 Public
27 Denied
30 Denied
46 Denied

I
Per capita
Average for
this type

$7,775
7,775
8,105
7,728
7,775
8,227
8,227
8,227
income
For this
decision

$5,000
9,353
7,477
8,775
9,269
9,055
7,185
8,226
ao
.--- wwwwow - - - - - -- - - I - - ---

CONCLUS IONS
In assessing the impact of a disaster, FEMA seeks to quantify
severity and magnitude in terms of losses and damage, determine
what resources are available from others to cover the losses and
damage, and arrive at the unmet needs which would be eligible
for assistance from FEMA. If the unmet needs are causing undue
hardship and suffering to individuals or governments, a disaster
declaration is warranted.

For the most part, FEMA is consistent in determining the
severity and magnitude of a disaster, even though it does not
always have complete information documented in a uniform manner.
Complete information on damages and losses is understandably
difficult to obtain because of the nature of the circumstances
under which they are obtained. on the other hand, some of the
missing information may have been overlooked or ignored. The
lack of uniformity in Governors' letters and regional damage
assessments contribute to such a problem.

The best means for FEMA to assure that it receives all the
available information in the shortest possible time is to develop
,a uniform format to supplement the information now provided by
State and local governments and regional offices and require that
they be used consistently. This uniform format could include
provisions for all the specific damage and loss data which FEMA
needs. Since estimated eligible costs seem to be strongly as-
sociated with the decision, these estimates should be given par-
ticular attention. Also, the formats could provide for Governors
to support their statements on capability and commitment.

With more complete and uniform information, FEMA could de-
velop computer models such as those we developed in this review.
This tool would.be useful as a.guide to future decisionmaking and
serve as a record for management evaluation.

RECOMMENDATIONS

We recommend that the Director, FEMA:

--Develop comprehensive, uniform forms to be used by
Govetnors when submitting their requests and by regional
offices when performing damage assessments.
--Use computer models, such as those developed by GAO, as
a tool for program decisionmaking and evaluation.

AGENCY COMMENTS AND OUR EVALUATION

The Director, FEMA, informed us that FEMA would consider
developing comprehensive, uniform damage assessment forms during
its current revision of.the regulation relating to the declara-
tion process. Shortly afte-r receipt of the Director's comments,
FEMA established a plan for considering this recommendation which
it expects to complete by July 1, 1982. We have no basis for
19

evaluating the Director's response to our recommendations and the
FEMA plan until the regulation has actually been issued.
The Director did not agree to use computer models as a tool
for program decisionmaking and evaluation. He said that the use
of a computer model may have some merit in the evaluation of
certain aspects of the Disaster Relief Program, but that its use
in decisionmaking with regard to disaster requests is severely
limited. He said there is no proper way to quantify key factors
such as hardship, suffering, and impact for use in a decision-
making computer model. Subsequent to this reply, FEMA decided
to have a consultant review its computer applications and recom-
mend appropriate changes. It expects to complete this action by
October 1, 1982.

We do not suggest that the use of computers models would
supplant the judgment of agency officials in assessing the hard-
ship and suffering or the impact of a disaster. On the other
hand, we believe that using computer models can be an extremely
effective tool for program evaluatibn and management control.
Each disaster file is handled on an individual, ad hoc basis,
without formal reference to the circumstances or experiences and
lessons learned in the evaluation of prior disasters. Consequent-
ly, program evaluators only consider the factors associated with
the current incident. Problems discussed throughout this report
about missing data and the lack of consistency in the quality and
methodology of disaster assessments indicate a need for more uni-
formity in disaster evaluations. By using computer models, FEMA
could develop formulas which would enable it to predict whether
a current incident would be declared a disaster based on past
experience. Such information would help FEMA judge the current
incident. Also, if the data related to the disasters were com-
puterized, FEMA would have the ability to search its prior dis-
aster records to identify similar disaster situations and circum-
stances and enable it to compare the similarities and differences
before making judgments on a current disaster request. Such
model analyses and disaster comparisons could be presented to
the Director, FEMA, for consideration in making a recommendation
to the President.
Our use of computer models dealt primarily with the quanti-
fiable factors associated with the severity and magnitude of
disasters in an attempt to evaluate FEMA's past decisions and
identify the implicit criteria used in the decisionmaking proc-
ess. If FEMA used a similar approach, it may eventually find
that some of the data it gathers has little influence on its
decisions or that its decisions seem to be based on other factors
not included in the computer model. Thus, it could refine its
data-gathering techniques and more clearly establish the criteria
it implicitly uses in its decisionmakihg process. These models
could also be expanded to include information pertaining to the
capability of State and local governments and their related com-
mitments. And, as experience is gained, and the usefulness of
20

the models increases, some measures of hardship, suffering, and
impact might eventually be developed.

Finally, various management reports could be developed for
the Director, FEMA, and other officials which'would provide them
with better management control over the operations of the disas-
ter assistance program office.
21

CHAPTER 3

IMPROVEMENTS SHOULD BE MADE IN EVALUATING

THE ABILITY OF STATE AND LOCAL GOVERNMENTS

TO HANDLE MAJOR DISASTERS AND EMERGENCIES

Improvements in FEMA's evaluation process should be made
even though it is faced with a difficult task of obtaining ade-
quate information under the stressful conditions existing in a
natural disaster. FEMA should obtain certain uniform information
regarding a disaster to speed up and improve the quality and con-
sistency of its decisionmaking. Formal guidance to State and
local governments is minimal. FEMA's published regulations re-
garding procedures for requesting major disaster or emergency
assistance essentially restate the law; require little additional
information; and do not reveal the policies, procedures, and
practices used by FEMA to evaluate the information.

FEMA's methods of evaluating State and local governments'
ability to handle major disasters are not clearly defined and are
not uniformly followed. FEMA'assesses the severity and magnitude
of the event, the capability and commitmaent of State and local
governments, and then makes its recommendation to the President.
Th'ese assessments vary greatly as to the type, quality, and degree
of information used to reach the decision. Furthermore, FEMA does
not document oral information, such as telephone conversations and
evaluation meetings, which are important inputs to decisionmaking.

The lack of consistency in the quality and methodology of
disaster assessments and the lack of knowledge by outsiders as to
FEMA's methods of evaluation can create doubt as to whether the
Federal Government is only providing supplemental assistance and
whether each request is judged in a fair and equitable manner.
Better advanced planning and the issuance of uniform policies and
procedures would alleviate the problem.

LIMITED GUIDANCE FOR
STATE AND LOCAL GOVERNMENTS

Numerous Federal programs become available to States when a
major disaster or an emergency is declared. The Congress did not
intend for the Federal Government to assume a unilateral role in
responding to such occurrences. To the contrary, the act's purpose
is to assist "State and local governments in carrying out their
responsibilities to alleviate the suffering and damage which result
from such disasters." (42 U.S.C. 5121(b), (1976 ed.))

The procedures for obtaining Federal assistance, as outlined
in the act, are as follows:
--The Governor of the affected State shall request a major
disaster or emergency declaration by the President.
2
2

--The Governor's request shall be based on a finding that
the major disaster or emergency is of such severity and
magnitude that effective response is beyond the capability
of the State and the affected local governments.
--For major disaster requests, the Governor shall also:
1. Take appropriate action under State law and direct
execution of the State's emergency plan.
2. Furnish information on the extent and nature of State
resources which have been or will be used to alleviate
the condition of the disaster'.

3. Certify that for the current disaster, State and local
government obligations and expenditures will constitute
the expenditure of a reasonable amount of their funds for
alleviating the damage, loss, hardship, or suffering re-.
sulting from such disaster. Additionally, State commit-
ments must be a significant proportion of the State and
local government obligations and expenditures.

--The prior expenditure of funds is not a prerequisite for
emergency requests, but the Governor is still required to:

1. Furnish information describing State and local efforts
and resources which have been or will be used to alleviate
the emergency.

2. 'Define the type and extent of Federal aid required.
For major disasters, FEMA's regulations require that Gover-
nors' requests include (1) an estimate of the amount and severity
of the damage, broken down by type, such as private nonag.ricul-
tural, agricultural? and,public, (2) a statement of actions
pending or taken by the State and local legislative and governing
authorities with regard to the disaster, (3) a certification as
to the amount of State and local government expenditures and ob-
ligations that will be used which will not require Federal reim-
bursement, (4) an estimate of the extent and nature of Federal
assistance and funds needed for each disaster-affected area, and
(5) other appropriate justifications in support of the request.
For emergency requests, the Governor is required to describe
the State and local efforts and resources which have been or will
be used fot which no Federal funding will be requested. The re-
quest must also define the particular type and specific extent
of Federal aid required.
The Governor of the affected State is responsible for com-
plying with these provisions of the law and implementing regula-
tions. Likewise,' FEMA is responsible for obtaining sufficient,
physical and financial information so that it can make an informed
recommendation to the President. Unfortunately, FEMA's regula-
tions do little to clarify and explain the law not do they
23

provide information to Governors as to the factors FEMA uses and
considers when it evalua-tes the States', capabilities and commit-
ments. Consequently, Governors are unable to predict the outcome
of their requests with any degree of certainty. This is partly
evidenced by the fact that 64 Governors' requests, or 42 percent,
have been denied since the beginning of fiscal year 1979.
EVALUATION OF THE CAPABILITY AND
COMMITMENT OF STATE AND LOCAL
GOVERNMENTS NOT CONCLUSIVE
By law, Federal disaster assistance is intended to be supple-
mentary in nature. it should be provided only when the severity
and magnitude of a major disaster or emergency is beyond the capa-
bilities of the State and affected local governments to effec-
tively respond and when they have committed, for major disasters,
a "reasonable" amount of their funds to alleviate the damage,
loss, hardship, or suffering. Capclbility is a crucial determinant
because many States may be abl-e to respond to very severe disas-
ters, whereas other States may not be able to respond to lesser
disasters. Capability can al so be a crucial determinant in
judging the reasonableness of the dollar value of commitments
which State and local governments are required.to make..
The bases for, and supporting documentation of, FEMA's deter-
minations of State and local capability and the reasonableness
of their commitments are often not conclusive. FEMA does not
require strict adherence to the requirements of the act and it
does not require complete and uniform information fot measuring
capability or for determining the reasonableness of commitments.
The type and quantity of information obtained on capability and
commitment varies from request to request and, thus, leads to
considering different information for decisionmaking.
FEMA does not always require Governors
to comply with legal requirements
The act requires Governors to follow certain specified
procedures as a prerequisite to obtaining Federal disaster assis-
tance. These procedures were established to ensure that Federal
assistance was supplementary to State and local disaster assis-
tance and that States were making good faith efforts to do what
they could to alleviate the damage, loss, hardship, and suffering.
Governors in many cases are not providing sufficient information,
and FEMA is not requiring such information to make basic deter-
minations relating to capability and commitment.
Our analysis of 96 major disaster requests in fiscal years
1979 and 1980 disclosed that almost one-half of the requests did
not contain (1) the required statement of finding by the Goverhor
thAt the disaster was of such severity and magnitude that it was
beyond the capability of State and local governments, (2) a cer-
tification by the Governor that State and local commitments were
.a reasonable amount of available funds, or (3) either statement.
The requests granted and denied are as follows:
24

Governor did not make required statement
Beyond Reasonable Both
capability commitment missing Total

Major disasters
granted
Fiscal year
1980 4 6 1 11
Fiscal year
1979 6 -8 6 20
Total 10 14 7 31
Major disasters
denied
Fiscal year
1980 3 2 7 12
Fiscal year
1979 1 1 0 2

Total 4 3 7 14

Total 14 17 14 45

Our computer analysis as to the importance of the presence
or absence of-these statements revealed that they did not have a
significant influence on FEMA's final decision. This was further
supported by our indepth review of 31 major disaster declarations
and denials, 8 of which lacked a capability statement and 5 of
which lacked a commitment statement.
Capability
These requests usually contained no comments whatsoever
regarding State and local capability or they only emphasized the
need for Federal assistance. For example, in one request that
was approved, the Governor said "It is my judgment that Federal
assistance is necessary to supplement the State and local efforts

FEMA officials told us that the fact that State and local
governments could not adequately respond to the situation was im-
plicit in the Governor's statement. FEMA told us that although
the Governor did not employ the precise wording of the law, his
request certainly evinced an understanding of the spirit of law
and legislative intent.
25

in another request, the Governor said:
"In my judgment, Federal assistance is necessary to
effectively alleviate the conditions caused by the
disaster. Legislative bodies of the affected areas
(4 counties) are considering all ways of coping with
the financial problems involved."

The Governor made no statement whatsoever regarding State
capability, although it is one of the most affluent States in
the Nation. In this case, FEMA concluded that one county did
not have the capability to handle the disaster and it was granted
Federal assistance. FEMA told us that, based on telephone con-
versations between headquarters and the region, it determined
the State would not be able to provide an effective response to
the unmet needs in~ one of the four counties involved, but the
rationale for this decision was not documented.

An Assistant General Counsel, FEMA, said it was impermisi-
sible to declare a major disaster if the Governor does not certify
that the situation is beyond the caLbabilities of the State and
local governments.

Commitment

In these requests the Governors generally described State
and local efforts being made and actions being taken to alleviate
the disaster situation. And, even when a dollar commitment was
mentioned, the Governors did not always certify that it was a
reasonable amount of available State and local government funds.
For example, one Governor said that 200 State employees and nu-
merous crews from the county, the district, and contractors were
working on the disaster and that certification of expenses in-
curred for this disaster would be transmitted as soon as possible.
FEMA's regional office reported that the State legislature had
recently enacted the fiscal year budget without providing funds
for disaster assistance and that the State's expenditures had
been insignificant. The required certification by the Governor
was not in FEMA's files so we asked FEMA to explain the bases
for recommending this major disaster request which was declared.
FEMA told us that its
If* * * declaration letter stipulated 75/25 cost
sharing of eligible public assistance and this
was considered an adequate commitment on the part
of the State and local governments."

State commitments must be
significantly proportional

The act requires that State commitments be a significant
proportion of State and local government obligations and expendi-
tures. FEMA cannot possibly determine whether State commitments
are significantly proportional when commitments (1) are only
26

described as significant, (2) do not specify obligations and
expenditures of local governments, or (3) include a total amount
not identified to either the State or local government. For
example, one Governor who requested Federal assistance only cer-
tified that State and local governments would make reasonable
expenditures and obligations consisting of State and local efforts
for public safety, rescuer removal of debris, and the State's
share of the individual and family grant program. He included no
dollar estimates for these expenditures.
Regarding this Governor's certification, the FEMA region
reported that it had requested additional information estimating
the monetary value of State assistance. The region said this
information was not available from the State at that time ande
if required by headquarters, it would be provided later. How-
ever, based on its inspection and past experience with State
operations, the region concluded that the State would continue
to provide all available State'resources in recovery operations.
Federal assistance was provided even though FEMA did not have
the required information on State and local commitments.

One year later *this same Governor made similar statements
regarding State and local government commitments for another
disaster without specifying details on the commitment. FEMA
estimated that this commitment would amount to $115,000 based
on the estimated cost for debris removal and emergency protec-
tive measures. Although FEMA did not determine the proportion of
State and local government commitments, Federal assistance was
granted.

The Director said that corrective action has been taken and
that FEMA is insisting that legal requirements are complied with
when a request is made. In this regard, the weekly Director's
Memorandum dated February 27, 1981, issued to headquarters and
regional personnel, contained the following policy statement:

"I. On receipt of requests from Governors for
Presidential declarations under PL 93-288, care-
fully review the requests to ascertain that the
request complied with the requirement for certain
essential elements of information such as State
and local commitment or implementation of the
State emergency plan (for major disasters). Where
such required actions are not included in the
request, it should be noted in the acknowledgement
to the Governor and the State advised to furnish
the information in order to preclude a delay in
processing."
This matter will also be considered when FEMA revises its
regulation on the Declaration Process.
27

Analysis and evaluation of
capability and commitment
data is inadequate'

FEMA does not use the same type of information for (1)
measuring the capability of State and local governments.and (2)
assessing the reasonableness of their commitments. Each request
is analyzed and evaluated on diffLerent bases and, thus, there is
no assurance that each major disaster and emergency request is
considered in a fair and equitable manner. We recognize the need
for FEMA to act quickly when a disaster occurs. However, this
situation could be remedied if FEMA makes States aware in advance
of the kinds of information it needs to perform complete and uni-
form evaluations and consistently requires the States to provide
the informiation.

Governors generally provide much of the information required
by the law and regulation, but they do not provide much of the
information FEMA uses in its evaluation. They usually describe
State and local efforts by identifying the various departments and
agencies whose personnel'and equipment have been used to respond
to the incident and assist individuals, evacuate citizens, remove
debris, and perform emergency'protective measures. These efforts
are known as."soft match," since the salary and equipment costs
would normally not have exceeded the cost of regular operations.
Soft match is not usually quantified in terms of dollars spent.
Governors usually identify the type of State and local work which
has been done or will be done and the amount of funds spent or to
be spent for which no Federal reimbursement will be requested.

Some Governors also voluntarily provide additional informa-
tion regarding State and local financial conditions, general
economy of the affected area, and prior disaster history.

Although FEMA believes each disaster request must be
subjectively evaluated on its own merit, it did develop an inter-
nal paper in 1975 on the "Logic of Evaluating Requests for Dec-
laration of a Major Disaster." To measure capability, FEMA's
objective is to determine what State and local governments can do
or have the ability to do. The paper suggests considering a
States' (1) authority to spend funds, (2) financial conditions,
(3) general economy--income, unemployment, etc., and (4) prior
disaster history. The paper also suggests considering location
and concentration of damage. To evaluate the reasonableness of
commitments, FEMA's objective is to determine whether the State
and local government response is commensurate with the severity,
and magnitude of the incident and whether the State is playing
an important role in the recovery effort. In this regard, the
paper suggests considering (1) general economy, (2) impact of the
incident on fiscal outlays, (3) fiscal outlays in recently de-
clared and non-declared disasters, (4) authority to spend funds,
(5) debt and borrowing limitations, and (6) bottom line or forced
contr ibution,.
28

In June 1981, headquarters advised the regions that their
damnage assessments were not uniform. Regions were instructed to
review the "Logic" paper for guidance and were offered copies of
the paper if it were not available in the regions.
We used these guidelines to determine the type and quantity
of information FEMA actually uses to arrive at its decisions.
FEMA uses the Governor's request and obtains whatever additional
information that is readily available to make a decision. The
Director informed us that for marginal cases, the FEMA national
office requests additional information which may require field
trips and.research. Nevertheless, the following schedule shows
the type.of information FEMA considered in the 31 disaster
requests we reviewed.

Requests
Information information
Type of information Provided by added by
considered Governors FEMA Total

Prior disaster history 14 10 24
General economy 9 12 21
Fiscal outlays in
recent disasters 13 7 20
Authority to spend'funds 6 9 15
Financial condition 7 7 14
Impact of disaster
on fiscal outlays 6 6 12
Surplus 0 7 7-
Ability to spend funds 4 0 4
Debt and borrowing
capacity 0 1 1
Note: FEMA declared 24 and denied 7 of these disaster requests.
The weight or influence which these factors have on FEMA's
subjective decisions are impossible to determine. in some re-
quests FEMA considered information on only one factor and in other
requests FEMA considered information on several factors.

We believe that if FEMA is to evaluate requests on a fair
and equitable basisF it must obtain the same type of information
in every case and uniformly evaluate them. FEMA's decisions for
declarations or denials may have been different if additional
information factors had been considered in individual cases. For
example, in one case FEMA based a denial, in part, on the fact
that the State and local governments had debt and borrowing
capacity. Had this factor been Uniformly considered in all re-
quests, other FEMA declarations might have been denials.

The financial conditions of States vary, with many States
having large surpluses. Few Governors provide information on
surplus funds in their requests. if they do, they usually advise
that they do not have the authority to spend surplus funds for
disaster relief. Some States also advise that they have no
I
29

I
authority or programs to aid their own citizens or their local
governments. FEMA provides Federal assistance -to States not
having the authority to spend surplus funds or provide other
assistance.
For example, FEMA denied a major disaster request because
the State and at least one of the two affected counties had large
surplus balances. The Governor protested FEMA's decision. He
explained that he did not have the authority to spend surplus
funds without approval from the State legislature. FEMA then
recommended and the President declared an emergency and Federal
assistance was provided. A week later the State legislature,was
to convene and could have authorized State assistance.

Surplus does not appear to be a valid and reliable measure
of State capability since FEMA does not know whether i t is acces-
sible. Nevertheless, if State surplus is a factor considered by
FEMA in reaching its subjective decision, then FEMA should uni-
formly consider this factot for all requests.

FEMA denies Federal assistance to States when it determines
the request is primarily for "budgetary relief." In these re-
quests, FEMA generally finds that much of the damage has already
been repaired and that Federal assistance would essentially re-
'place budgeted funds spent. Whether these budgeted funds had
been earmarked for other essential public services that were fore-
gone is not considered. Had the State and local government not
proceeded and made the repairs, they would have had an unmet need
and Federal assistance may have been provided. Such a practice
penalizes those who take care of their needs in anticipation of
Federal assistance. If this.practice is valid, it should become
a formal policy and made known to all States.

FEMA may deny part of a request for Federal disaster assis-
tance if it believes the State or local government commitment is
inadequate. For example, one Governor identified six departments
of the State that had committed personnel and some equipment.. He
said,
"State and local efforts in response to this disaster
have been and are presently continuing to be signifi-
cant in the areas of debris removal, damage assessment,
and recovery efforts. * * * Costs for these items and
the indirect costs of managing the disaster recovery
effort from the State and local funds are considered
to be a significant contribution toward disaster costs
for which no Federal reimbursement will be requested."

FEMA considered the State commitment to be totally unaccept-
able, but still granted Federal assistance to individuals and,
ultimately, for limited public assistance. FEMA explAined to us
that it concluded that correcting the damage and loss to public
facilities was within the capability of the State and local
.governments. Consequently, its denial of public assistance would
force these governments to repair their own public facilities.
.3
0

As such, this would constitute an increase in commitments which
FEMA then considered reasonable. This practice is known as forc-
ing a commitment. In our view, the Governor should be advised
that the commitment is inadequate and be given an opportunity to
increase it. otherwise, the entire request should be denied.
FEMA's inability to obtain adequate and reasonable State
and local government commitments is acknowledged by its adoption
of a 75 percent/25 percent cost-sharing policy.with regard to
eligible public assistance costs. We found that this policy
is consistent with the act, as discussed more fully in chapter 4.

STATES' FINANCIAL CAPABILITIES
NEED TO BE ADEQUATELY ASSESSED
In August 1979, we reported on FEMA's reimbursement policies
for snow emergencies (CED-79-97). 1/ We identified problems in
FEMA's reimbursement policies and made several recommendations
to correct these policies. FEMA has not resolved all of the
problems. One unresolved problem concerns inequities in the
Disaster Relief Act of 1974. Our report recognized that the way
the act is implemented unfairly penalizes or rewards States
applying for Federal reimbursements because of differences in
State laws and States' preparedness to deal with major disasters
and emergencies. We recommended that FEMA prepare a comprehen-
sive analysis of the impact of these potential inequities on
Federal disaster assistance and submit to the Congress a detailed
plan and legislative changes to correct such weaknesses.
FEMA agreed that inequities exist in disaster relief because
of differences in State laws,r fiscal procedures, and flexibility
in allocating funds. It agreed to conduct a study if directed
by the Congress to do so.
The recommended analysis has not been undertaken. if this
analysis had been undertaken, FEMA would now be in a better posi-
tion to evaluate the capability of State and local governments
and the reasonableness of their commitments. And, the problems
discussed in this report may have been minimized.

FEMA said that our comments on budgetary relief (p. 31)
are not consistent with the comments and recommendations made
by GAO in its August 1979 report on FEMA's reimbursement policies
for snow emergencies.

I/The GAO report was actually directed to the Administrator of
the Federal Disaster Assistance Administration, one of FEMA's
predecessor agencies.
31

IWe requested that FEMA identify the specific comments and
recommendations to which they referred. They identified several
pages in our August 1979 report and informed us that throughout
these pages we were insisting that States use all available
resources before Federal assistance be provided.

We believe that FEMA has misinterpreted the message in
both of our reports. That message is that FEMA is not adequately
assessing the capability of State and local governments to respond
to their own disasters and emergencies. Out August 1979 report
contained comments, such as the following:
--"The act intends that Federal funds should not be dis-
bursed to the extent that the State can use its legally
,available financial and physical resources."
--"The act requires that State and local governments will
do all they are capable of doing in responding to snow-
storms before Federal assistance may be provided."1
---" To the extent that States have the financial capacity
to meet the demands of the snowstorm, they should not be
eligible to receive financial assistance under the act."

Our August 1979 report also made the following recommendation:

--That FEMA "revise the snow removal reimbursement policy
to better safeguard the President's Disaster Relief Fund
by (1) ensuring that Federal reimbursements are not paid
to States that can afford their own cost of snow removal,
(2) reimbursing only the cost of those activities which
the State could not have performed without Federal finan-
cial assistance, and (3) -discontinuing to reimburse States
for costs that would have been incurred without a snow-
storm."

In discussing "budgetary relief" in this report, we referred
to those situations where Federal assistance is denied to States
who have, as an expediency, used budgeted funds from other
sources for disaster assistance. This raises a question as to
whether the States have used all "available" financial resources
or whether they have gone beyond their own financial capability.
A major issue in both reports is that FEMA does not know
what State and local government funds are available for disaster
assistance and are, therefore, not adequately assessing the
capability of State and local governments to handle their own
disasters.

CONCLUS IONS

When a major disaster or an emergency occurs, FEMA needs to
quickly evaluate the situation and determine whether a declara-
tion should b'e made. FEMA reacts to the situation, obtains
and considers readily available information, and makes its
32

recommendation to the President. However, F'EMA does not always
document the relevant factors considered in their decisions.
Better advanced planning by FEmA and uniform policies, procedures,
And guidelines would enable FEMA to speed the process while, at
the same time, enable it to obtain more information and perform
better.analyses. This should also minimize unwarranted requests
and the expenditure of Federal resources to evaluate the requests.
Not all Governors comply with the requirements of the act
and regulations. As a result, FEMA frequently finds it neces-
sacy to work with incomplete information. FEMA should advise the
StAtes in advance as to the specific nature of these requirements
and make it clear 'that future requests which fully comply with
Federal laws and regulations will help avoid delays in processing
their requests.

FEMA has made little effort'to inform States fully on the
policies, procedures, methods, and guidelines which it uses to
evaluate requests. The disclosure of such information would
enable State and local governments to better decide whether
they had a valid requ.est to make and provide them with know-
ledge 'of the types and amounts of information FEMA needs to make
its decision.

Differences in State laws, fiscal procedures, and funding
levels make it difficult for FEMA to assess States' capability
and create inequities among the States in administering disaster
relief.

RECOMMENDATIONS
We recommend that the Director, FEMA:
--Reevaluate and improve FEMA's assessment criteria for
evaluating major disaster and emergency requests.
--Establish written policies, procedures, and guidelines
to use when evaluating major disaster and emergency
requests and publish them in the Federal Register. This
should include (1) an explanation of FEMA's basic philos-
ophy for evaluating capability and commitment, (2) an
explanation of the use of evaluation factors, such as debt
and borrowing capacity, surplus funds, prior disaster
history, and (3) FEMA's positions on budgetary relief,
forced commitments, and similar matters.

--Require that Governors' requests include comprehensive
information on the financial capability of the State,
the availability of such resources under State law, per
capita income, disaster trends, and similar factors to
expedite FEMA's assessment of the level of capability each
State could attain and to aid FEMA in evaluating the
reasonableness of State commitments.
33

--Make it clear that future requests which fully comply with
Federal laws and regulations will help avoid delays in
processing the requests.
--Require the documentation of all substantive discussions
and evaluation meetings held by FEMA.

AGENCY COMMENTS AND OUR EVALUATION
The Director informed us that the first four recommendations
are being given consideration in the development of a revised
regulation on the Declaration Process. Shortly after receipt of
the Director's comments, FEMA established a plan for consiaering
all o f the foregoing recommendations, with final regulatory action
to be completed by September 30, 1982. We have no basis for
evaluating the Director's response to our recommendations and the
FEMA plan until the regulation has actually been issued.

The Director agreed that documentation of the basis for
their final recommendation to the President, including a statement
of relevant factors considered~, is needed. However, he would not
want the documentation process to impede the expeditious process-
ing of the requests. We had spoken to the former Associate Di-
rector, DR&R, regarding this matter and he did not,believe that
such documentation would slow down the decisionmaking process.

RECOMMENDATION TO THE CONGRESS
As recommended to FEMA in August 1979, we recommend that the
Congress direct FEMA to prepare a comprehensive analysis of the
impact of potential State inequities on Federal disaster assis-
tance and submit to the Congress a detailed plan and legislative
changes to correct such weaknesses.
34

CHAPTER 4

THE CONGRESS NEEDS TO CLARIFY
DISASTER ASSISTANCE ISSUES
Two recent FEMA actions have raised serious concerns that it
may have changed the scope of the Disaster Relief Act. First, in
May 1980 FEMA adopted a general policy of requiring State and
local governments to pay 25 percent of eligible public assistance
costs. Second, disaster assistance has been provided for man-made
events, such as the Love Canal chemical contamination and the
Cuban refugee influx in Florida. Such assistance has stirred
consiLderable controversy because many feel that these situations
were not natural disasters covered by the act.
FEMA COST-SHARING POLICY
IS CON51STENT WITH THE ACT
FEMA has encountered numerous problems, as explained earlier,
in trying to assess the "1reasonableness" of State and local gov-
ernment commitments. To overcome these problems, FEMA unilater-
ally adopted in May 1980 a policy of providing 75 percent Federal
funding for eligible public assistance costs and requiring the
State and local governments to finance the remaining 25 percent.
FEMA advises us that the President ultimately decides the actual
percent which,State and local governments must assume.

Problems in obtaining
a reasonable State and
local commitment
To ensure that Federal disaster assistance is supplemental
to State and local efforts, FEMA regulations require each State
applying for Federal disaster relief to execute a Federal-State
agreement as a prerequisite to receiving Federal disaster assist-
ance (44 C.F.R. 205.44 (1980)). Under these agreements, the
Governor must assure that a reasonable amount of State and local
government funds will be committed to disaster relief.
FEMA has difficulty in ascertaining the reasonableness of
the Governors' commitments. FEMA's experience is that these
commitments extended to circumstances where the Governors have on
occasion agreed to assume (1) the expense of accomplishing work
which is necessary for recovery purposes but which would not be
eligible for Federal reimbursement undet the disaster assistance
program, (2) admin-istrative and other costs, sometimes including
the cost of prudent hazard mitigation to reduce or eliminate the
threat of similar future damage and hardship, or (3) the entire
cost of particular projects or categories of projects which would
otherwise be eligible for Federal reimbursement. Governors have
also, at times, agreed to a contribution based on a percentage of
disaster-r-elated costs which would otherwise be eligible for
Federal reimbursement.
35

FEMA often found, after the fact, that,the State and local
contributions were not substantial or significant. Furthermore,
States were not always prudent in their requests for public
assistance. The Associate Director, DR&R, told us that States
were providing limited co-mmitments so far as public assistance
was concerned and that FEMA attempted to negotiate greater con-
tributions from States. The Deputy Associate Director, DR&R,
told us that commitments were often soft, in that State and local
governments often used normal programs such as police and fire
departments as their required commitments.

At a Regional Directors' conference in December 1980, the
Director, FEMA, cautioned that "without restraints the Disaster
program could becomDe a welfare program."

Adoption of cost-sharing
policy for public assistance

in May 1980 FEMA adopted a policy of requiring all States to
agree to pay 25 percent of the costs of eligible public assist-
ance programs unless such States could show, or the President
determines, that 25 percent was not a reasonable commitment.
FEMA's adoption of the cost-sharing approach was an effort to
obtain reasonable commitments of funds and uniformity among the
States.

FEMA justified this new policy primarily on the grounds that
it is in accordance With congressional intent to make Federal
disaster assistance supplemental to State efforts and that such
a policy readily ensures that States will pay a "reasonable
amount" of the costs of disaster relief. FEMA believes that such
a policy would also provide greater equity among the States.
Some States, for example, have set aside funds to use in the
event of a disaster while other States have no fund and, there-
fore, depend on 100 percent Federal funding. One State with a
voter-imposed limitation on taxes argues that this limitation is
grounds for more generous Federal disaster assistance.

FEMA also believes that such a cost-sharing arrangement has
the advantage of reducing Federal expenditures for the disaster
relief program--estimated at $30 million in fiscal year 1982--and
that it will reduce the time previously required to deliver
disaster relief assistance to affected States by reducing the
time it takes to evaluate the extent of State and local contri-
butions to the disaster relief effott.

FEMA policy creating
opposition and concern

FEMA's cost-sharing policy has created much opposition and
concern on the part of Governor's! legislators, and organizations
of State officials. The Governors who have had to agree to thie
cost-sharing in order to receive Federal disaster assistance feel
the 75/25 percent cost-sharing policy is an arbitrary FEMA de-
cision that should not have been made without consulting with
36

them and the Congress. They contend that State and local
governments already bear a major role in disaster relief because
of their predeclaration commitment of resources and their assump-
tion of ineligible public assistance costs. They also contend
that FEMA's new policy further burdens the State and local gov-
ernments by requiring them to pick up, in addition to the above
mentioned cost, 25 percent of the eligible public assistance
costs.
Legislators have expressed the belief that statutory
authority does not support this policy, and several Governors
have written directly to the President making such statements as:
--Iam signing the Disaster Agreement under protest."

--"I must object strongly to your limitation of Federal
reimbursement for public facilities."
in view of the foregoing, FEMA has submitted proposed legis-
lation to incorporate the 75/25 cost-sharing policy into present
law.
Cost-sharing policy
consistent with the act
FEMA'has not formalized its policy by promulgating a regula-
tion or by publishing a notice in the Federal Register. However,
our analysis of the act and its legislative history indicates
that FEMA can determine that previous levels of sharing have been
unacceptably low and that 25 percent represents a reasonable level
of State and local government participation in all foreseeable
circumstances. The policy would not be proper if FEMA were un-
willing to deviate from it no matter what the circumstances. How-
ever, FEMA informs us that it remains willing to modify its cost-
sharing ratio if the circumstances surrounding a disaster warranted
a different cost-sharing ratio. Thus, we cannot say that FEMA's
policy is an abuse of discretion.
While this policy has not been embodied in a formal rule,
FEMA has provided us with a draft of such a rule. The draft rule
justifies this approach primarily on grounds that it is in ac-
cordance with congressional intent to make Federal disaster
assistance supplemental to State and local efforts and that such
a policy readily ensures that States and local governments will
pay a reasonable amount of disaster relief costs.
The draft of FEMA's proposed rule expressing its 75/25
percent cost-sharing policy indicates that exceptions might be
made when there are:
`* * * such extraordinary and compelling
circumstances, unique to a particular disaster-
affected State, that State and local contribu-
tions at this level or of this type would be
clearly precluded. Such circumstances might
37

include the occurrence of an overwhelmingly
catastrophic event or repeated disasters
affecting the same jurisdiction.":I
FEMA's policy of requiring States to bear 25 percent of the
cost of public assistance, if applied without exception, could
result in State and local governments having to pay for disaster
relief costs which they believe are beyond their capability or
which they believe constitute more than a reasonable amount of
State and local funds. Conversely, States which can afford.to
pay more than 25 percent could be allowed to contribute less than
a reasonable amount.

We are not able to conclude that FEMA is acting improperly,
as long as it does not refuse to make exceptions to the policy.
In terms of the statutory goal that costs be shared between
Federal and State and local authorities, FEMA can legitimately
seek assurances that the State and local governments'are, in fact,
contributing a reasonable amount of their funds, as required by
section 301(b) of the act. As a consequence, based on its ex-
perience with the program, FEMA concluded that 25 percent was a
reasonable proportion for St.ate and local governments in virtually
all conditions.

CONCLUS ION
FEMA's apparently strict adherence to the 75/25 cost-sharing
ratio is based on a discretionary assessment that this cost-
sharing approach results in a reasonable level of State and local
government participation in disaster relief costs in almost
all cases. It obviates the administrative problems associated
with attempting to determine a reasonable commitment for each
disaster and also assures compliance with the statutory purpose
that the Federal aid be supplemental. Moreover, FEMA has left
itself free to recommend providing more or less than 75 percent
of the ass-Istance should it be convinced in a particular situa-
tion that an exception is warranted. Under these circumnstances,
FEMA's cost-sharing policy is consistent with existing law. How-
ever, the policy has created controversy on the part of Gover-
nors, legislators, and organizations of State officials.
RECOMMENDATION TO THE CONGRESS
Because of the controversy surrounding FEMA's cost-sharing
policy, we recommend that the Congress reevaluate the present law
and clarify the extent to which supplemental Federal assistance
should be given in major disasters and emergencies.
AGENCY COMMENTS AND OUR EVALUATION
The Director informed us that FEMA had not adopted a
general policy of requiring State and local governments to pay
25 percentlof eligible public assistance costs. He said such
adoption would have had to be promulgated as a regulation in
the Federal Register. Accordingly, he said all references to
FEMA's policy on cost sharing should be corrected.
38

We see to need to change our terminology. We are using
the word "policy" in a general sense as described by Webster's
dictionary: "a principle,, plan,,or course of action, as pursued
by a government, organization., individual, etc." We did not ad-
dress the issue of whether this policy should have been promul-
gated.in the Federal Register as it was not germane to our
review.
USING THE DISASTER RELIEF
ACT FOR OTHER THAN NATURAL
CATASTROPHES NEEDS CLARIFICATION

In 1980 the President declared emergencies for the Cuban
refugee influx into Florida and the chemical contamination of
Love Canal in New York. This provoked co'ngressional conc ern that
the act was drafted too.broadly,land,was being applied to catas-
trophes.other th.an'nat,ural disasters as the Congress original-
ly intended.
Because other than natural1 catastrophes do not generally
fall within the definition of those catastrophes specifically
enumerated-.bby section 102 of the act, major disaster or emergency
declarations issued in response to such catastrophes must be
supported by reference to the general phrase in the act's defin-
itional section "other catastrophe." While a literal reading
of this phrase indicates that the President possesses consider-
~able latitude in selecting the type of catastrophes which may
qualify as major disasters or emergencies, the legislative his-
tory of the act indicates that the Congress intended it to al-
leviate State and.local conditions caused primarily by natural
catastrophes.

In 1980, the Senate Committee on Environment and Public
Works criticized the President's declaration of an emergency for
the Cuban refugee influx in Florida. The committee stated that
the Congress historically intended Federal disaster assistance to
apply to a fixed range of responses to particular kinds of events
which were spelled out in the act's definitional sections. Fur-
thermore, the committee stated that the reauthorization act
(Public Law 96-568) should not be applied to extraordinary occur-
rences not specifically enumerated in the act insofar as it is
not intended to provide assistance for occurrences of strictly a
social or economic nature.
The Deputy Associate Director, DR&R, told us that he
considers the Cuban refugee influx to be a man-made emergency.
He contends that FEMA assistance was provided primarily to al-
leviate the hardship and suffering of Florida citizens and not
to help the Cuban refugees. He'contends FEMA should continue to
provide such assistance because FEMA is the only Federal agency
with existing authority to readily provide funds, direct other
Federal agencies to act, and reimburse State and local govern-
knents for their expenditures.
39

CONCLUSION

Until the Congress amends the definition section of the
act, by deleting the pDhrase "or other catastropDhe" or includes
an additional provision limiting the scope of the act to occur-
rences resulting from natural catastrophes, Presidential authority
to declare a major disaster or an emergency in response to other
than natural disasters exists. While Presidential declarations
premised on this approach seem to be based on a broad interpre-
tation of the act, nonetheless, as pIresently worded, the act gives
discretionary authority to the President to determine which occur-
rences amount to other catastrophes.

RECOMMENDATION TO THE CONGRESS

We recommend that the Congress clarify the act by spelling
out as clearly as possible the type of incidents which may re-
ceive disaster assistance. It could define more precisely the
intended coverage of the act and specify that disaster assistance
should be provided if there are no other Federal programs avail-
able and the State is unable to cope with the situation. FEMA
and the administration would then be better able to administer
the act as the Congress intends.
i
I

I
I
40

APPENDIX I
APPENDIX I
MAJOR DISASTERS, EMERGENCIES,

AND DENIALS REVIEWED BY GAO
- ; -, ,~~~~~~~~~~~~~~�

FEMA
declaration Date of
number declaration

Major disasters:
-
....
Total estimated
requirements
(note a)

$ 1,618,553
1,328,294
1,335,866
1,909,482
1,255,000
8,249,808
23,462,600
133,000
1,618,555
2,475,:098
2,967,000
219,341
444,000
1,650,000
510,000
287,292
3,732,946
12,024,918
128,783,242
3,389,000
373,600
566,000
547,034
2,607,376
FEMA
region

9
9
6

4
6

: 7
6
7
4
3
1
9
10
9
9
9
6
6
7
5
9
State

Calif.
La.
Hi.
Ark.:
Okla.
Ala .
Ill.
Tex.
Kans.
N. Meex.
Iowa
N.C.
Va.
Conn.
Am. Sam.
Wash.
Hi.
Ariz.
Calif.
La.
Ark.
Mo.
Wis.
Calif.
566
567
573
574
576
578
583
587
588
589
590
605
606
608
610
612
613
614
615
616
617
620
626
633
Oct. 9, 1978
Dec. 6, 1978
Mar. 7, 1979
Apr. 11, 1979
Apr. 13, 1- 9:'79
Apr. 18, 1979
Apr. 30, 1979
Junez 14, I979-
June 15, 1979
June 23,1i979I
July 1, 1979
Sept. 29, 1979
Sept. 29, 1979
Oct. 4, 1979
Nov. 9, 1979
Dec. 31, 1979
Feb. 6, 1980
Feb. 19, 1980
Feb. 21, 1980
Apr. 9, 1980
Apr. 16, 1980
May 15, 1980
July 24, 1980
Oct. 2, 1980
Emergencies:
3070
3071
3072
3073
3074
3075
3078
3081
Mar.
Mar.
Mar.
Mar.
Mar.
Mar.
Feb.
June
12, 1979
12, 1979
13, 1979
15, 1979
17, 1979
24, 1979
1, 1980
13, 1980
Wash.
Mo.
Ga.
N.H.
Ala.
Fla.
Calif.
Pa.
10
7
4
1
4
4
9
3
30,000
100
41,000
319,168
657,338
784,506
21,294,311
0
41

APPENDIX I
APPENDIX I
FEMA
declaration
number
Total estimated
requirements
(note a)
FEMA
region
Date of
declaration
State
Denials:
$ 3,053,960
913,804
976,000
1,227,100
2,465,500
1,461,560
3,570,000
3,570,000
963 .211
Mar. 12, 1979
Mar. 30, 1979
Apr. 23, 1979
Nov. 26, 1979
Dec. 7, 1979
Apr. 26, 1980
May 1, 1980
May 31, 1980
June 10. 1980
Fla.
Miss.
Wash.
Kans.
Hi.
Hi.
N.Y.
N.Y.
Ala.
4
4
10
7
9
9
2
(Appeal) 2
4

a/Estimated requirements for major disasters and emergencies were
obtained from FEMA reports available at the time of our review.
Amounts related to denials were obtained from regional disaster
summary reports.
42

APPENDIX 11
APPENDIX 11
~TECHNICAL ANALYSES-
This appendix provides details for the technical reader on
the analyses and models used in this report'.
THE DATA BASE
we analyzed 96 of 102 requests for disaster assistance from
fiscal years 1979 and 1980. Our primary source of information
was the FEMA decision files on each disaster. The files contained
information on FEMA's decisions with respect to the Governors'
requests-,-i.e., complete or-parti,al assistance granted and assis-
tance denied. The files contained statistics describing the
magnitude of the disasters, in terms of number of lives lost,
people injured, number-of homes damaged, and similar data. The
files also contained miscellaneous data on a State's capability
to handle the disaster and some information on the local dollars
committed to dealing with it. We identified a total of about 30
pertinent variables.contained in the FEMA files.
The major,problem with the data was that much of it was
missing or ambiguous. While a few of the 96 requests had infor-
mation on most of the 30 variables., most did not, and it was
not always clear as to why the data was missing. For example,
if there were no statistics on number of homes damaged, we did
not know whether any homes were damaged in the disaster or whether
no good esti-mates to report existed.
A review of some of the requests'and discussions with FEMA
officials led us to believe that often requests with missing
values fell into two categories: (1) minor disasters where a
missing value usually meant zero.or a relatively small number
and (2) obviously large disasters like Mt. St. Helens where a
missing value meant a large,number and officials, in their hurry
to process.the re'quest, did not take the time to estimate the
number.

We tried to reconcile the missing data problem in several
ways. We used only the important variables with the least
amount of missing data and combined several variables. For
example,w we combined homes with minor damage, major damage,
and those completely destroyed into one measure of homes damaged.
We also added the estimated amount of eligible costs for indi-
vidual assistAnce and public assi stance to have one figure as
the total amount of eligible assistance. This combination of
variables was also desirable to reduce them to a manageable
number for-the regression analysis discussed later.
Another way we tried to solve the problem was to substitute
what-we believed were reasonable estimaltes for missing values.
For example, there were 16 requests that had mis'sing values
for the total estimated eligible costs--an essential variable
for the analysis. We found that 9 of the 16 requests were large
and declared as disasters. For these, we used current estimated
43

APPENDIX II APPENDIX II
requirements as a substitute. The other seven requests were
denied assistance and appeared to have total damage less than
the smallest request received ($138,000). We arbitrarily assIumed
$100,000 as the total eligible,costs for these denied requests.
We did not have good substitute values for missing dat-a on homes,
damaged or people killed'or injured so we assumed they were zero
for some analyses. While the assumption may be adequate for the
small disasters, it is probably inadequate for the large disas-
ters. Consequently, we dropped the nine large disasters with
missing values from some of our analyses.
In addition to the FEMA files, we collected da ta from the
statistical abstract on,per,capita income in the States where.
the disasters occurred--we used this as an indicator of a State's
financial capacity. For.jurisdictions in the Pacific, such:as
the Trust Territories with no published data on per capita in-
come, we assumed a value of $1,000.
Such assumptions and approximations may appear to make anyz
analyses of questionable value. The reAder should certainly
keep these limitations in mind in reviewing the results. How-
ever,.a major focus of our analysis was to assess FEMA's de-,
cisionmaking criteria and this application data with all the
missing values is the primary data FEMA had on which to base
its decisions, although FEMA probably supplemented the decision
,file data with some information from phone calls and other in-
formed contacts. While our assumptions may be crude, we expect
that they simulate, to some degree, the assumptions FEMA made.
BASIC ANALYSIS
For selected variables, we developed a separate frequency
distribution for both granted and denied requests. These vari-
ables are total estimated eligible costs, people injured, homes
damaged, State per capita income, month the event occurred,I
political party of the Governor, and commitment statement on the
part of the State in which its disaster occurred. These dis-
tributions were compared to see if there was a difference between
granted and denied requests. These distributions are shown in
the tables in the text of the report.
DECISION MODELS
To further estimate the basis of FEMA's decisions we
developed several statistical models. We attempted to see how
FEMA used several variables in concert to make decisions. For
example, if some measure of disaster severity were extremely
high, would this compensate for other variables that wer'e low?
We also used the models for leads on cases that FEMA may have
classified inconsistently. This was done by determining.what
individual cases the models classified the outcome, granted
or denied, differently than FEMA.
44

APPENDIX II
APPENDIX II
The models were basically statistical multiregression
models. But instead of using the traditional straight line re-
gression, we used an "S" shaped curve to try to better fit the
data to the dichotomous dependent variable, i.e., 0 (denied) or
1 (granted). See figure 3, which is a hypothetical example
showing the cumulative probability of granting assistance for
one independent variable.

FIGURE 3
CUMULATIVE PROBABILITY CURVE
1.0

/
/
I
/
/
/
/
/
/
I

- /
I
I
I
I

/
I
I

/ /
I
I
a
w
2
z

0
o
'U

n
a
I-
'4

0.5
0

o
0
m -
A.
./

, _

_ _ _ - -
i II rr IL .00goo so
t-
I
I
I I I
0
30
10
20
40
: TOTAL ESTIMATED ELIGIBLE COST IN MILLIONS OF DOLLARS
r I ;
45

APPENDIX II APPENDIX II

We used a logit model based on, the cumulative logistic
probability function which is often used to give this k,ind of.
probability curve. The equation for the logistic model is:

P = 1 + EXP 1-(Bo + BiX1 + B-2X2...+BXj + u)] ,

Where: P is the probability a request is granted.

X1, X-...Xnare the influencing variables, i.e., size

of disaster in dollars, homes damaged, etc.

Bo is the constant term in the regression equation.

B1, B2... Bn are the regression coefficients for the

variables, size of- disaster in dollars, homes damaged,
etc.

EXP is the exponential function.

u is the error term.

The logit equationocan be written as: 1/

Z Log (1_ = Bo + B1X] + B2X2...B1nXXn + u
u yt-p /

Where: Z is a probability index that a request
is granted.

While some authorities recommend estimating the logit
models by Maximum Likelihood or simulation, we used the simpler
Ordinary Least Squares (OLS) method using P=.999 for P=l and
P=.001l for P=0. These approximations were necessary because
otherwise Z will be undefined.. The more sophisticated method
may be more mathematically rigorous, but in many cases will,
give the same results as OLS.

1/If the estimated probability of approval, as -derived from the
estimated value of Z, is equal to or greater than .5, the
request is considered approved, otherwise it is denied.
46

APPENDIX
IIAPEDXI
APPENDIX 11
Another analytical problem was that requests for assistance
can be decided in several ways. For example, a request for a
combination of both individual and,public assistance could be
decided in four ways: (1) both-granted, (2) only public assis-
tance granted, (3) only individual assistance granted, or (4)
both denied. But, the models were designed only to classify
decisions in one of two ways--granted or denied. Ideally we
would have liked to develop separate models for each .type of
assistance but there were not enough observations to do this.
To partially accommodate this problem we grouped all the requests
together and included the categorical variables-type of request.
We also ran separate models for certain categories.

To more thoroughly analyze F'EMA's decisionmaking process we
*experimented with different variables.. In some cases we also
combined several variables into one term, an.interaction term.
This is an attempt to capture differential responses for some
of the continuous variables, e.g., total el-igible cost, number of
homes damaged, etc. As a consequence, several models were devel-
*oped. Six of these-are shown in table 8 of the text. The
discussion that follows is a summary of the results of the six
models.

R~jessonresults

For most of the six models a positive relationship exists
between the Z score and the total eligible costs, number of people
injured and number of homes damaged. I/ This indicates that the
greater the.values of total eligible costs, the more people in-
jured, and the more homes damaged, the greater 'the probability
of a request being granted.
The impact of a Governor making the required commitment
statement is not entirely clear since the coefficient for the
commitment variable is not significant at the 80 percent confi-
dence level except in one case. Also, the sign of the coef-
ficient is negative in two of the models.

The coefficient of the month variable is negative and
is usually significant. It suggests that the probability of
getting a request granted-is less the more months that pass
starting from October 1978 if all other variables remain the
same. This may be partly because of inflation decreasing the
real.value of total dollars requested. Also, while the month
variable was not explicitly designed to estimate fiscal year
influences, it probably picked up some of the impact of budget
problems towards the end of fiscal year 1980.

l/While eligible cost and people :injured were statistically
significant in the six models, homes damaged was significant
in only three of the models.
47

APPEINDIX II
APPENDIX II
The sign of the coefficient of per capita income is
negative but is not significant at the 80-percent confidence
level in most models. This negative sign indicates that the
chance of a request from a State with a low per capita income
being granted is higher if everything elseremains the same.

Table 10 shows the regression coefficients for the Z score
for two different models provided for illustration purposes.
These numbers show the magnitude and direction of the relations
between the variables and Z scores. The significance of the
coefficient is indicated by the "t" shown in parentheses. The
larger the absolute number for "t" the more confident we are in
the significance of the coefficients. I/

Summary statistics on the table are indicators of how
accurately the regression equation, as a whole, expresses the
relationship among variables. The R2 expresses the percentage
of the variation in Z score that is explained by the influencing
variables. The closer R2 is to 1.0 the better the equation fits
the data.

The R2 we achieved for some of the models might be con-
sidered somewhat low. However, in this kind of model with qual-
itative dependent variables, the goodness of fit of these equa-
tions is better measured by the percentage of correct classifi-
cation of decisions (i.e., cases the model classified the same
as FEMA). Most of the models classified more than 80 percent
of the decisions as FEMA did. Overall, we believe the regression
equations express the relationship between the decision and the
other variables reasonably well. However, we feel FEMA needs
to do additional work on those models before they can be used
for assessing requests.

We also used the six models to find leads on cases that
FEMA may have classified inconsistently. But they seemed to
overclassify cases granted because of certain technical problems.
There seemed to be a different probability of a case being de-
clared for the different types of assistance. For example, if
one requests both types of assistance FEMA seems more likely to
grant either both or individual assistance in comparison with
granting public assistance. We did not try to evaluate FEMA's
underlying rationale for this but we tried to capture these in-
fluences by adding the categorical variable-type of assistance
granted. Therefore, some of the models have not only the vari-
ables that characterized the disaster, e.g., people injured, but
also the two categorical variables--type of request as well as
type of assistance granted. The combination of these two
categorical variables in a sense measures the propensity of FEMA

I/Some authorities would argue that all variables would be
significant in our analysis since we are dealing with the
universe of requests in fiscal year 1979 and 1980 and not a
sample.
48

APPENDIX II APPENDIX II
to grant certain types of assistance given a certain type of
request. We also used these additional models to look for FEMA
inconsistencies. Table 11 lists the cases that FEMA classified
differently from any of our models. To further evaluate if these
20 cases were inconsistent FEMA decisions, we compared the values
of key variables, e.g., number of homes damaged, for these cases
with the average value for the particular type of assistance
granted. In those cases denied, the values of key variables were
compared with the average of those turned down. Table 9 in the
text shows this comparison for eight cases.

APPENDIX II
APPENDIX II
Table 10

Selected Regression Equations
Model number
3 . 5
Independent variables
5.7987*
(2.06)
8.4649*
(2.58)
Constant
Request for individual
assistance only

Request for public assistance
only
-5.8968*
(-3.05)

-9.0600*
(-5.54)
Total estimated eligible
cost, in $100 millions
3.6828*
(1.37)
27.343*
(1.66)
Homes damaged and destroyed,
in thousands
0.8259*
(2.08)
0.2074
(0.43)
People injured, in thousands

Per capita income in the State
in which the disaster
occurred, in dollars
9.9469*
(1.51)

-0.0004
(-1.08)
29.7749*
(2.30)

-0.0003
(-0.94)

I
If Governor made required
commitment statement

Month of decision
0.4873
(0.43)
-1.3144
(-0.92)
-0.2458*
(-2.38)
Summary statistics
0.40
96
0.24
67
Number of decisions
Number of decisions classified
differently than FEIMA
Percent of decisions classified
the same as FEMA
16
13
83
8
1
* Significant at least at the 80-percent confidence level.
50

APPENDIX II
APPENDIX II
Table 11
Decisions that Models Predicted
Differently from FEMA
* D

iL.

I. 1
FEMA
Decision decision
number (note a)
Model
decision
(note a)
3
6
9
23
24
27
30
31
32
34
35
37
40
44
46
52
71
1
1
1
1
1
0

0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
0
0
1
1

1 indicates approval.
_1
73 1
97 1 0
98 0
a/0 indicates disapproval and

APPENDIX III
APPENDIX III
Federal Emergency Management Agency
Washington, D.C. 20472

SEP 29 1981

Mr. Henry Eschwege
Director
Community and Economic
Development Division
United States General
Accounting Office
Washington, D.C. 20548
Dear Mr. Eschwege:
I appreciate the opportunity to review the draft report entitled
"Improvements Should Be Made in Evaluating Requests for Federal
Disaster Assistance." As the a�ency charged with the responsibility
for administering the President s Disaster Relief Fund, FEMA has
an interest in the entire subject matter of this report.
This report has a number of worthwhile comments regarding the
problems associated with State and local government commitment,
State and local government capability, and cost sharing. These
are interrelated and must be considered in requests for Federal
disaster assistance under Public L~w 93-288. FEMA has been
seeking improvements in this area and currently is in the process
of revising 44 CFR 205, Subpart C, the Declaration Process. The
comments and recommendations in this draft report will be con-
sidered in this revision.
FEMA, on behalf of the Administration, has submitted an amendment
to the Disaster Relief Act by including 75/25 cost sharing for
Public Assistance. We believe that our proposal would result in a
clear identification of funds expended on the part of a State and
its local governments as required by the Act for a commitment.
I am enclosing specific comments on this draft report.
Sincerely,

Director
Enclosure
52

APPENDIX III APPENDIX III

GAO DRAFT REPORT ENTITLED "IMPROVEMENTS SHOULD
BE MADE IN EVALUATING REQUESTS FOR FEDERAL DISASTER ASSISTANCE

Specific Comments by FEMA to this report are:
1. Page 6, second paragraph. This paragraph refers to FEMA's
evaluation of requests wherein it states that "FEMA advised us
that it had no specific criteria for evaluating requests for
assistance..." There are criteria established -in Sections 201 and
301 of the Act that FEMA 'uses for its evaluation of requests.
2. Page 8, first 'paragraph. This paragraph dealing with "unmet
needs" and "impact" is accurate, but this paragraph also fails to
recognize the statutory criteria contained in Sections 201 and 301
of the Act. (See GAO note 2, below.)
3. Page 8, first paragraph under Problems. In responding to a
request, regional personnel visit the scene of the incident to
make a first-hand evaluation of the damage and requirements for
Federal assistance. The report then states, "They then make their
recommendation to the Director of FEMA..." It would be more
correct to revise'the last sentence to read, "Basedon this
review, the Regional Director then makes a,recommendation..."
4. Page 10,seco�nd paragraph. The first sentence reads, "It is
clear that the higher the estimate of eligible costs, the greater
the chance of a request being granted." This is not entirely
correct. Generally, it would be true for the same State, the same
type of catastrophe and under similar circumstances.
5. Page 13, State Commitment. It should be noted' that for some
disasters in the past, a declaration was made with the;expectation
of working-out the specific State commitment and including it in
the Federal-State Agreement. Section 44 CFR 205.44 provides that.
the Federal-State Agreement shall contain a statement of the
commitment. Since no FEMA.funding is approved until this agreement
is signed, the specific commitment is sometimes obtained after the
declaration. Instances of this are rare and are generally limited
to those situations where immediate Federal assistance is required
to prevent a delay in initiating Federal disaster response activities.
6. Page 19-, first recommendation. 'That FEMA develop comprehensive,
uniform forms to be used by Governors in submitting their requests
and by regional offices in performing damage assessments.
FEMA is currently in the process of revising 44 CFR 205, Subpart
C, the Declaration Process. This recommendation will be considered.
7. Page 19, second recommendation. That FEMA use computer models
such as those developed by GAO as a tool for program decision-
making and evaluation.

GAO notes: 1. Some page references have been changed to agree with the
final report.
2. This paragraph refers to FEMA's internal guidelines.
Statutory criteria is set forth on pages 23 and 24.
53

APPENDIX III APPENDIX IllI

Nonconcur. The use of a computer model may have some merit in
thle evaluation of certain aspects of the Disaster Relief Program
and has been discussed. However, its use in decision-m aking with
regard to disaster requests is severely limited. Key factors in
the evaluation of a request include hardship, suffering and impact.
Thete is no proper. way that these can be quantified for use in a
decision-making computer model.
8. Page 23, line 18: "Prior" should be added before "expenditure,"
so t'S sentence will read, "The prior expenditure of funds is not
a prerequisite for emergenQy requests, but the Governor is still
required to..." in the definition section (102(l)), an "emergency"
is a catastrophe ". ..which requires Federal emergency assistance
to supplement State and local efforts to save lives and protect
property.. ." Further, in requesting an "emergency," a Governor
must describe the "State and local efforts and resources which
have been or will be used to alleviate the emergency..." (Section
301(a).) lie therefore conclude that some expenditure of State
and local funds must be made either before or following an e'mergency.
9. Page 23, last paragraph. These'comments with regard to the
capability and commitment of State and local governments will be
considered in the current revision of Subpart C.
10. Pag e24, fourthparagraph. The report states that FEMA does
not aTw'ays r`equire Governors to comply with legal requirements.
FEMA has taken corrective action in this regard. FEMA it insisting
that these legal requirements are complied with when a request is
made.
11. Page 29, second paragraph. The statement that FEMA bases its
decision on the "Governor's request" and "whatever additional
information is readily available" infers that FEMA may not adequately
investigate each request. it should be noted that for marginal
requests, FEMA national offfee does request additional information
which may require field trips and research. These cases go beyond
"readily available" in the scope of assessments.
12. Page 30, fourth paragraph. With reference to "budgetary
relief," the comment in this report is not consistent with the
comments and recommendations made by GAO in its August 2, 1979
report, "Federal Snow Removal Policy: Improvements Needed."
13. Pages 33 and 34, Recommendations. GAO recommends that the
Director of FEMA take the following actions:
Reevaluate and improve FEMA's assessment criteria.
Establish written policies, procedures, and guidelines that
are used by FEMA to evaluate major disaster and emnergency
requests and publish them in the Federal Register. This
should include (1) an explanation of FEMA's basic philosophy
,for evaluating capability and commitment, (2) ani expression
of the use of evaluation factors such as debt and borrowing
54

APPENDIX'- I S-I
APPED IX I I I

I
. ... .
capacity;. s-urpiu, fundsd-. pr hler saterhistory;-ad (3) -
its positions on .b,udgetary relief', fobrcid Ccommftments d an :d
similar matters. ' " - ' - -
Require. that G:overno,rs!. r,equests include comprehensive
information on S.tate laws aindfunding' provfisfions, Sta:te
revenues, per capita incom:e, -dias:ter ttinds, and sinfilar ?
factors to ex.pedite', FgMA's assessm,nt tf'the'l level' of i.
capability eeach State coduld a,ttd'in and'"to aid-FEMA 'in. :. :' -:
evaluating ithe-te'asonableess of' State 'ommitments:. -i .
Make it clear.t that .future reques't's whic'h fully comply with
Federal laws and reg'ulation"s wi il' help avoid delays in..
processing the requests. '- -
Require the documentation of all substantive discussions
and evaluation meetings held by FEMA.
The first four recommendations are being given consideration in
the development of Subpart C. As to the final recommendation, we
agree that documentation of the basis for the final recommen-
dation, including the statement of relevant factors considered, is
needed. However, we would not want the documentation process to
impede the expeditious processing of the requests. It is FEMA's
position that such documentation is not available under the Freedom
of Information Act.
14. Page 35, first paragraph: FEMA has not "adopted a general
policy of requiring State and local governments to pay 25 percent
of eligible public assistance costs. Such adoption would have
had to be promulgated as a regulation in the Federal Register.
Rather, the FEMA Regional Directors, with Governors of disaster-
affected States, have entered into Federal-State Agreements for
each disaster starting with Mount St. Helens in May 1980, which
provide that the costs of public assistance shall be met 75 percent
by FEMA and 25 percent by State and local governments. For a
vast majority of such disasters, the President has required this
cost-sharing directly in his declaration letters. All similar
references to FEMA's "policy" on 75/25 cost sharing at pages 35-
39, and elsewhere in the report, should be corrected.
15. Page 35, third paragraph. The first sentence reads, "To
ensure that Federal disaster assistance is supplemental to State
and local efforts, FEMA's regulations require each State applying
for Federal disaster relief to execute a cost sharing agreement as
a prerequisite to receipt of Federal disaster assistance..." This
is misleading and technically inaccurate. FEMA's regulations do
not require that Federal-State Agreements be "cost sharing agree-
ment[s]." However, the report correctly notes that the Governor
must assure that a reasonable amount of State and local government
funds will be committed. The first sentence can be corrected by
the elimination of "cost-sharing" in the third line.

APPENDIX III
APPENDIX III
16. Page 38, Recommendation to the Congress. That the Congress
reevaTu-Tt� t-he present law and FEMA's cost sharing policy and
clarify the extent of supplemental Federal assistance to be given
in major disasters and emergencies.
Concur. FEMA, on behalf of the Administration, has submitted
an amendment to the Disaster Relief Act by including 75/25 cost
sharing for Public Assistance. Our proposal applies to restoration
work and debris removal but not to other emergency work. We believe
that this proposal would result in a clear identification of funds
expended on the part of a State and its local governments as
required by the Act for a commitment. Also, it would provide a
more consistent approach among States with regard to their commitment
on public sector repairs.

(068109)
56