Crop Insurance: USDA Needs a Better Estimate of Improper Payments To
Strengthen Controls Over Claims (Letter Report, 09/22/1999,
GAO/RCED-99-266).

Pursuant to a congressional request, GAO provided information on crop
insurance, focusing on: (1) the extent to which crop insurance claims
are paid in error--either unintentionally or fraudulently--and, to the
extent practical, a comparison of the rate at which claims are paid in
error with rates for other types of insurance; (2) the insurance
companies' and the Department of Agriculture's Risk Management Agency's
(RMA) quality controls to ensure that accurate claims payments are made;
and (3) the proposals being considered to reduce insurance companies'
administrative requirements and the potential impact of these proposals
on the operations of the crop insurance program.

GAO noted that: (1) there are no precise estimates of the extent to
which crop insurance claims are paid in error; (2) while the RMA
estimated that about 5 percent of claims were paid in error in 1997, the
agency's methodology for estimating errors was questionable in several
respects; (3) specifically, the estimate was based on an inadequate
sample size and did not include the results of timely, on-site reviews
to detect errors resulting from fraud; (4) although information on
payment errors for other types of property and casualty insurance is
limited, a recent insurance industry study reported higher rates of
fraud-related payment errors than RMA reports for crop insurance; (5)
RMA and the insurance companies revised the process for examining the
accuracy of paid claims in 1998; (6) previously, the agency has reviewed
the claims of a few companies every year for accuracy, but available
resources limited the number of claims that could be examined; (7) under
the new process, the agency is able to get much broader coverage of
claims activity by relying on the companies themselves to review an
agency-selected statistical sample of their claims to detect erroneous
payments; (8) the companies use agency guidance for ensuring that the
sampled claims were properly paid; (9) the agency then reviews a sample
of these same claims to determine whether the companies' review
processes are adequate; (10) while it is too early to evaluate the
effectiveness of this approach, success will depend heavily on how well
the companies implement this approach and the quality of the RMA's
oversight of the process; (11) RMA and the companies are considering
proposals to simplify administrative requirements in three principal
areas: (a) developing alternatives to producers' actual production
histories, which are used to determine the insured value of a crop; (b)
simplifying the administration of one type of crop
insurance--catastrophic; and (c) changing other administrative
requirements, such as allowing farmers to self-certify claims below
certain dollar amounts; (12) the agency and the companies do not agree
on how these simplification proposals would affect program operations;
and (13) for example, while some simplification proposals could reduce
the companies' administrative costs, these proposals could also increase
claims payments, which would increase government costs.

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

 REPORTNUM:  RCED-99-266
     TITLE:  Crop Insurance: USDA Needs a Better Estimate of Improper
	     Payments To Strengthen Controls Over Claims
      DATE:  09/22/1999
   SUBJECT:  Insurance companies
	     Agricultural programs
	     Insurance claims
	     Comparative analysis
	     Erroneous payments
	     Insurance cost control
	     Internal controls
	     Claims processing
	     Property damages
IDENTIFIER:  Federal Crop Insurance Program

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed product.                         **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************

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

Report to the Honorable Charles W.  Stenholm, ranking Minority
Member, Committee on Agriculture, House of Representatives

September 1999

CROP INSURANCE - USDA NEEDS A
BETTER ESTIMATE OF IMPROPER
PAYMENTS TO STRENGTHEN CONTROLS
OVER CLAIMS

GAO/RCED-99-266

Crop Insurance

(150087)

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

  FCIC - Federal Crop Insurance Corporation
  FFMIA - Federal Financial Managers Integrity Act
  FMFIA - Federal Managers Financial Integrity Act
  HHS - Department of Health and Human Services
  OIG - Office of the Inspector General
  RMA - Risk Management Agency
  USDA - U.S.  Department of Agriculture

Letter
=============================================================== LETTER

B-283433

letter date goes here

The Honorable Charles W.  Stenholm
Ranking Minority Member
Committee on Agriculture
House of Representatives

Dear Mr.  Stenholm: 

Federal crop insurance protects participating farmers against crop
losses caused by perils such as droughts, floods, hurricanes, and
other natural disasters.  This multibillion-dollar program,
administered by the U.S.  Department of Agriculture's Risk Management
Agency (RMA), provides subsidized insurance through private insurance
companies that assume a portion of the risk associated with claims
payments.  Since 1981, when the current crop insurance program was
established, RMA has provided $14.1 billion to farmers for insured
crop losses.  The program's loss experience is a major factor in
determining the cost of federal crop insurance to farmers and to the
government. 

Concerned about RMA's effectiveness in managing the process to
minimize erroneous claims payments, you asked us to (1) review the
extent to which crop insurance claims are paid in errorï¿½either
unintentionally or fraudulentlyï¿½and, to the extent practical, compare
the rate at which claims are paid in error with rates for other types
of insurance; (2) examine the insurance companies' and RMA's quality
controls to ensure that accurate claims payments are made; and (3)
describe the proposals being considered to reduce insurance
companies' administrative requirements and the potential impact of
these proposals on the operations of the crop insurance program. 

   RESULTS IN BRIEF
------------------------------------------------------------ Letter :1

There are no precise estimates of the extent to which crop insurance
claims are paid in error.  While the Risk Management Agency estimated
that about 5 percent of claims were paid in error in 1997, the
agency's methodology for estimating errors was questionable in
several respects.  Specifically, the estimate was based on an
inadequate sample size and did not include the results of timely,
on-site reviews to detect errors resulting from fraud.  Although
information on payment errors for other types of property and
casualty insurance is limited, a recent insurance industry study
reported higher rates of fraud-related payment errors than the Risk
Management Agency reports for crop insurance.  We are making a
recommendation to the Secretary of Agriculture to improve the
agency's methodology for estimating the program's error rate for
claims payments. 

The Risk Management Agency and the insurance companies revised the
process for examining the accuracy of paid claims in 1998. 
Previously, the agency had reviewed the claims of a few companies
every year for accuracy, but available resources limited the number
of claims that could be examined.  Under the new process, the agency
is able to get much broader coverage of claims activity by relying on
the companies themselves to review an agency-selected statistical
sample of their claims to detect erroneous payments.  The companies
use agency guidance for ensuring that the sampled claims were
properly paid.  The agency then reviews a sample of these same claims
to determine whether the companies' review processes are adequate. 
While it is too early to evaluate the effectiveness of this approach,
success will depend heavily on how well the companies implement this
approach and the quality of RMA's oversight of the process. 

The Risk Management Agency and the companies are also considering
proposals to simplify administrative requirements in three principal
areas:  developing alternatives to producers' actual production
histories, which are used to determine the insured value of a crop;
simplifying the administration of one type of crop
insurance--catastrophic; and changing other administrative
requirements, such as allowing farmers to self-certify claims below
certain dollar amounts.  The agency and the companies do not agree on
how these simplification proposals would affect program operations. 
For example, while some simplification proposals could reduce the
companies' administrative costs, these proposals could also increase
claims payments, which would increase government costs. 

   BACKGROUND
------------------------------------------------------------ Letter :2

Although the federal crop insurance program was established in 1938,
it was substantially changed and expanded by the Federal Crop
Insurance Act of 1980.  Currently, RMAï¿½through the Federal Crop
Insurance Corporation (FCIC)--relies on private companies to sell and
service crop insurance policies under their own name and to adjust
losses when a claim is made.  Federal crop insurance is currently
available for 75 crops on a county-by-county basis.  In addition, the
number of farmers participating in the program has increased to over
400,000, out of about 2 million farmers nationwide. 

The financial soundness of the crop insurance program depends on
substantial participation by the nation's farmers.  Premium rates and
cost to the government are determined largely by the program's loss
experience.  Generally, the higher the crop losses, the higher the
premiums in future years.  From 1981 through 1998, FCIC paid farmers
$14.1 billion for insured crop losses, and in 1998 alone, FCIC paid
$1.7 billion. 

Federal crop insurance offers farmers two primary levels of coverage,
catastrophic and buyup, which are available for major crops. 
Catastrophic insurance, created by the Federal Crop Insurance Reform
and Department of Agriculture Reorganization Act of 1994, was
designed to provide producers with a minimum level of protection for
a small processing fee.  Buyup insurance protects against more
typical and smaller crop losses in exchange for a farmer-paid
premium.  In addition to these two primary types of coverage, federal
crop insurance offers a number of products that protect farmers'
revenues against declining market prices for their production. 

Federal crop insurance is sold to farmers by insurance agents
representing one or more insurance companies that have a standard
reinsurance agreement with the U.S.  Department of Agriculture (USDA)
to sell and service crop insurance.  The agent and the farmer work
together to determine the level of coverage for expected yield and/or
price and the number of acres to be planted.  A farmer's expected
yield is based on 4 to 10 years of production history.  In the
absence of at least 4 years of production history, the county's
averages for each crop can be used.  The agent and the farmer sign a
completed application attesting that the information is correct prior
to transmitting the application to the insurance company. 

Upon receiving the application, the insurance company (1) verifies
the information by comparing it with any prior records, (2)
recalculates the computations on the applications, (3) determines if
the completed application meets all RMA requirements, and (4)
calculates the premium.  If the application meets all company and RMA
requirements, the company issues the policy to the farmer. 

If farmers incur crop losses, they file a claim with their insurance
agent or company.  The company assigns an adjuster who visits the
farm and, using RMA guidance, determines the percentage of loss of
the acres planted.  The adjuster forwards the claim to the insurance
company, which verifies and recalculates the claim.  If all company
and RMA requirements are met, the company pays the claim to the
farmer.  All paid claims are subject to review by the companies and
various government agencies, including RMA, USDA's Office of the
Inspector General (OIG), and GAO. 

The Federal Managers Financial Integrity Act (FMFIA), and the Federal
Financial Management Integrity Act (FFMIA), hold agency managers
responsible for ensuring that adequate systems of internal controls
have been developed and implemented to, among other things, minimize
improper payments.  An adequate system of internal controls should
provide reasonable assurance that an agency is effectively and
efficiently using resources, preventing unauthorized disposition of
agency assets, and producing reliable financial reports. 
Furthermore, the Government Performance and Results Act (Results Act)
encourages federal agencies to more effectively use their resources
by requiring annual reporting on the extent to which agencies meet
their annual performance goals. 

In the last 5 years, the OIG has reported that the insurance
companies' quality control reviews were superficial and did not
provide an independent verification that claims were properly paid.\1
To strengthen the quality control process, the Inspector General
recommended that RMA develop an estimate of claims paid in error,
including a detailed definition of what constitutes an error.  In
1998, in response to a series of internal reviews and to the
requirements of the Results Act, RMA initiated a study to estimate
the number of claims paid in error.  This estimate is discussed in
the following section. 

--------------------
\1 Report To The Secretary On Federal Crop Insurance Reform, USDA
Office of the Inspector General, Audit Report No.  05801-2-At (April
1999); Risk Management Agencyï¿½Reinsured Companies' Actual Production
History Self-Reviews, USDA Office of the Inspector General, Audit
Report No.  05099-1-Te ( Sept.  1997); and Risk Management
Agencyï¿½Federal Crop Insurance Claims, USDA Office of the Inspector
General, Audit Report No.  05601-3-Te (Feb.  1998). 

   EXTENT TO WHICH CROP INSURANCE
   CLAIMS ARE PAID IN ERROR IS
   UNCERTAIN
------------------------------------------------------------ Letter :3

The extent to which crop insurance claims are paid in error has never
been precisely calculated.  USDA reported in its 1998 financial
statements that an estimated 5 percent of crop insurance claims were
paid in error in 1997.  However, this estimate may not be accurate
because RMA's methodology for estimating errors is questionable in
several respects.  Information on error rates for other types of
insurance that could be used to compare with the RMA estimate is
limited. 

      RMA'S ESTIMATE OF ERRORS IS
      QUESTIONABLE
---------------------------------------------------------- Letter :3.1

RMA's estimate of errors may not be accurate because it is based on a
small sample and did not include the results of timely field work to
detect fraud.  In order to determine a rate of claims paid in error,
RMA used a statistical sample of 200 claimsï¿½out of more than 121,000
claims paid in 1997.  However, because the sample size was relatively
small, the estimated error rate could be inaccurate by as much as 80
percent.  Therefore, RMA's 5-percent estimate could be as low as 1
percent or as high as 9 percent.  Because of the small sample size
and the fact that only part of a claim may have been paid in error,
these error rates cannot be used to estimate the percentage of
dollars paid in error.  However, according to GAO and OIG reports
completed in 1987 through 1991, incorrect payments totaled 10 percent
or more of the total dollar value of claims paid in the sample
reviewed.  Neither the GAO nor the OIG reports contained national
estimates of error rates for payments. 

According to RMA and OIG officials, a larger sample size would have
generated a more accurate error rate to serve as a baseline for
judging future performance.  These officials said that a minimum
sample size of between 500 and 800 claims would be needed to produce
an accurate error rate.\2 However, RMA officials told us they did not
have the resources to review a larger sample size. 

RMA estimated that one-tenth of 1 percent of the claims paid in 1997
were fraudulent.  To arrive at this estimate, RMA reviewed the
paperwork associated with claims to uncover fraudulent statements
about the amount of acreage planted, production history, or crop
produced.  However, RMA's Deputy Administrator for Compliance told us
that this type of review is not always effective in identifying fraud
and that a more timely and thorough field evaluation is needed to
develop accurate estimates of fraud.  More specifically, claims would
have to be examined when they are adjusted.  In this way, a reviewer
could review the condition of crops for which losses are claimed and
interview individuals who might have knowledge about the farmer's
production, such as elevator operators or field office officials in
USDA's Farm Service Agency.  This kind of inspection is common for
other types of insurance claims.  For example, the Federal Emergency
Management Agency checks for fraud by quickly inspecting flood sites
for damage after insurance company inspections and before issuing
claims payments. 

Even with a timely, thorough field visit, identifying rates for
fraudulent claims payments is very difficult, according to RMA's
Deputy Administrator for Compliance, because it is hard to determine
whether an individual has purposefully misrepresented claims data. 

--------------------
\2 A larger sample size will also provide a basis to estimate the
dollar value of claims paid in error. 

   INFORMATION ON ERROR RATES FOR
   OTHER LINES OF INSURANCE IS
   LIMITED
------------------------------------------------------------ Letter :4

We could not obtain information on overall (unintentional and
fraudulent) payment error rates for other types of private insurance
because companies consider such data proprietary.  However, other
federal insurance programs periodically measure payments made in
error.  For example, the Department of Health and Human Services
(HHS) made such an estimate for improper Medicare benefit payments to
providers in 1998.  While the Medicare payments made in error are not
directly comparable to crop insurance payments made in error,
improper Medicare payments were about 7 percent of the $176 billion
paid to health care providers, according to HHS estimates. 

Although information on overall error rates for other types of
insurance was limited, we identified a recent industry study of
payment errors associated specifically with fraud.  According to this
study, the fraud rate for property and casualty insuranceï¿½the type of
insurance most comparable to crop insuranceï¿½was estimated to be
between 10 and 25 percent in 1996.\3 The study's estimates were based
on surveys of fraud prevention associations and insurance companies. 

--------------------
\3 Insurance Fraud:  The Quiet Catastrophe 1996, Conning Insurance
Research and Publications. 

   RMA IS RELYING ON COMPANIES'
   QUALITY CONTROL EFFORTS TO
   REDUCE ERRONEOUS PAYMENTS
------------------------------------------------------------ Letter :5

Under a recently instituted process to minimize the number of
erroneous payments, RMA relies principally on the insurance
companies' reviews of an RMA-selected statistical sample of claims
payments.  This process is also designed to ensure broader coverage
of claims reviews by having all the companies review a portion of
their claims payments each year.  To prevent and detect erroneous
claims, RMA has also instituted procedures that increase training
requirements for agents and loss adjusters. 

      RMA'S NEW PROCESS FOR
      REVIEWING CLAIMS DEPENDS ON
      THE ACCURACY OF THE
      COMPANIES' EFFORTS
---------------------------------------------------------- Letter :5.1

Until 1998, RMA conducted its own annual reviews of claims payments
to assess companies' handling of claims, including error detection. 
However, because of limited resources, RMA did not review enough
claims or insurance companies to adequately determine if claims were
properly paid.  Instead, these reviews focused on a sample of claims
from 3 to 4 of the 18 companies that sell and service crop insurance. 
In addition, RMA relied on referrals from farmers, insurance
companies, and USDA's Farm Service Agency to identify erroneous
payments.  A number of the insurance companies told us that they
supplemented RMA's reviews by internally reviewing the accuracy of a
portion of the claims' payments they made each year. 

Starting in 1998, RMA and the insurance companies agreed to change
the way claims are reviewed for erroneous payments.  This new process
provides a basis for the companies to determine if their internal
controls are working effectively and if their employees, including
agents and adjusters, are properly trained in order to minimize
erroneous claims payments.  RMA's role is to oversee the companies'
efforts to review claims payments for errors.  To implement the new
process, RMA provided the companies with guidance on the type and
scope of reviews to be conducted.  It also provided each company with
a statistical sample of at least 50 (and in many cases up to 150) of
its policies that had claims paid for crop year 1998.  RMA's
compliance staff provided oversight by reviewing about 600 selected
claims from the same sample.  The random samples were designed to
cover a percentage of crop insurance policies written and crop
insurance claims paid for each company. 

Under RMA's new review process, for each policy in the sample, the
company must verify the accuracy of the information reported by the
policyholder, the agent, and the loss adjuster, including the planted
acreage reported for the policy, the certification of the producer's
actual production history, and the summary of insurance coverage.  In
addition, each company is required to review all relevant claim
information and supporting production information, such as the
production worksheets, appraisal reports, and settlement sheets.  RMA
completes the review process by selecting a sample of the companies'
reviews and auditing them for quality and completeness.  To implement
these new review requirements, several participating companies
instituted new procedures and hired and trained additional staff. 

In 1998, the first year of implementing the new review process, RMA
and the companies worked together to clarify technical and reporting
requirements and placed less emphasis on completing all the steps in
the review process.  For example, RMA reviewed procedural matters
with companies and clarified supporting documentation requirements
for future submissions rather than evaluating the effectiveness of
the companies' reviews. 

In addition to the new review process, RMA has established a working
group with the participating crop insurance companies to review and
revise compliance processes as needed.  This working group hopes to
find ways RMA and the companies can cooperate to better use limited
compliance resources and to reduce the number of erroneous payments. 

      REVISED TRAINING
      REQUIREMENTS FOR AGENTS AND
      ADJUSTERS
---------------------------------------------------------- Letter :5.2

In order to reduce the potential for erroneous claims payments, RMA
requires the companies to provide a minimum number of hours of
training per year for their agents and loss adjusters.  In 1998, to
improve the skills and knowledge of agents and adjusters, RMA revised
these requirements.  For new agents, annual training requirements
increased from 8 hours to 12 hours.  For experienced agents, these
training requirements decreased from 6 hours to 3 hours.  For new
loss adjusters, annual training increased from 32 classroom hours and
24 field training hours to a total of 60 hours, of which 24 must be
in the classroom.  For experienced loss adjusters, training
requirements increased from 16 hours to 18 hours.  Also, by 2000, all
agents and adjusters will be required to pass certain competency
tests in order to continue to sell and service crop insurance
policies.  Among other things, an agent must be able to review
actuarial documents, know how to complete and distribute forms and
materials used in sales and service activities, and understand
reporting requirements and other procedures and regulations. 

   CONSENSUS LACKING ON THE
   EFFECTS OF ADDITIONAL
   SIMPLIFICATION PROPOSALS ON
   PROGRAM'S OPERATIONS
------------------------------------------------------------ Letter :6

In response to the Federal Crop Insurance and the Department of
Agriculture Reform Act of 1994, RMA and the industry are considering
proposals in three areas to reduce administrative effort and cost: 
(1) changing the way producers' actual production histories are
determined, (2) simplifying the administration of catastrophic
insurance, and (3) increasing the amount of allowable overstatement
of acreage and production on claims formsï¿½referred to as the
allowable tolerance for erroneous claims payments.\4 However, RMA and
the companies do not agree on the savings associated with these
changes and their potential effect on the program's vulnerability to
losses. 

--------------------
\4 The proposals we refer to were submitted by a working group of the
National Crop Insurance Servicesï¿½an industry association representing
the majority of the companies participating in the federal crop
insurance program. 

      CHANGING THE WAY ACTUAL
      PRODUCTION HISTORY IS
      CALCULATED
---------------------------------------------------------- Letter :6.1

Currently, a crop insurance policy's insured value and associated
risk is based primarily on the producer's actual production history
for 10 years.  Developing this history is time-consuming and somewhat
complex because it is based on a rolling 10-year crop production
average, and a farmer often cannot adequately document production. 

Questions are frequently raised by farmers and their associations
about the number of crop years required for determining this history,
the reliability of the farmer-provided information, and the
reliability and fairness of the formula used.\5 In addition, some
farmers have raised questions about the validity of using a 10-year
average crop production history because recent technical improvements
in herbicides, pesticides, and other farming practices have allowed
farmers to greatly increase production in the past few years.  To
deal with these questions, insurance companies have proposed two
alternative approaches that the industry is now considering.  These
proposals, if adopted, would probably be pilot-tested in a number of
locations for each crop before they are instituted nationwide. 

One industry proposal would replace farmers' individual production
history with an average county yield for all farmers in that county. 
By adopting this approach, the insurance companies and the farmers
could eliminate extensive documentation and calculations and make it
easier for the companies to compute losses.  However, farmers with
above-average production are generally opposed to this change because
it would benefit farmers with lower than average historical yields
and penalize farmers with higher historical yields.  Such a change
could discourage farmers with above-average production from
participating and encourage greater participation from farmers with
below-average production.  The change could benefit the federal
government by reducing the possibility of fraud associated with false
or inaccurate production data, but it could increase claims payments
to producers with below-average production. 

The second industry proposal would eliminate the use of production
history and average county yield altogether.  Instead, the proposal
would rely heavily on the combined judgment of the farmer and
insurance agent to establish expected production and value of
coverage, and on the integrity of the loss adjuster to calculate the
amount of loss.  It is therefore a more subjective approach and could
expose the government to greater risk.  In addition, this alternative
could eliminate the need to collect actual production data, making it
difficult to return to the current method for determining insurable
value and loss. 

Changes in the way the production histories are calculated could be
costly to insurance companies because of the need to create new
software programs and revise thousands of individual histories. 
According to some insurance company officials, these revisions could
undermine their companies' investments in databases. 

--------------------
\5 In addition, USDA's OIG has consistently reported problems
associated with calculating and applying production history data for
determining claims.  Risk Management Agencyï¿½Crop Insurance Claims in
Virginia 1995 and 1996 Crop Years, Audit Report No.  05601-1-Hy (Dec. 
1997); Risk Management Agencyï¿½Reinsured Companies' Actual Production
History Self-Reviews Washington, D.C.; USDA Office of the Inspector
General, Audit Report No.  0599-1-Te (Sept.  1997); and Risk
Management Agencyï¿½Crop Insurance Claims in California 1995 and 1996
Crop Years, USDA Office of the Inspector General, Audit Report No. 
05601-1-SF (May 1997). 

      SIMPLIFYING CATASTROPHIC
      INSURANCE COVERAGE
---------------------------------------------------------- Letter :6.2

This industry proposal would simplify catastrophic insurance in three
ways.  First, it would eliminate coverage for different acreages
(fields), which is now allowed when a farmer has different
ownership-operator agreements (shares) within a county.  This
proposal would require farmers to buy a farmwide policy by crop,
regardless of ownership arrangements, and could reduce the companies'
administrative costs for selling and servicing catastrophic crop
insurance.  Insuring on a farmwide basis by crop would reduce the
opportunity for farmers to create or enhance claims by shifting
reported production from one insured field to another.  (According to
agency officials, some producers shift reported production from one
insured field to another in order to understate production on one
acreage, resulting in a fraudulent claim for that field.) This
practice is difficult to discover and document.  However, according
to farmers and their associations, requiring them to buy farmwide
policies would reduce participation in the program because this
change could reduce their coverage.  For example, assume a farmer
insured two 100-acre fields with actual production histories of 80
bushels of corn per acre.  Under a farmwide policy, the farmer would
not receive a claims payment if one field produced 40 bushels per
acre and the other produced 120 bushels per acre since the average
farmwide production would be 80 bushels per acre, which is equal to
the farmer's actual production history.  Under a field-by-field
policy, the farmer could receive a claims payment on the field that
produced 40 bushels per acre because its yield was below the actual
production history. 

The two remaining administrative simplification ideas under this
proposal include easing reporting requirements on production history
and allowing farmers to self-certify claims.  For example, several
companies said that allowing farmers to report acreage by telephone
and to self-certify claims below certain dollar limits could simplify
their administrative processing of claims.  However, according to an
RMA official, acreage reporting by telephone could create legal
problems if the farmer disputed the unsigned acreage report.  In
addition, self-certification of claims could increase the number and
amount of fraudulent claims.  Since this proposal was first made, RMA
has permitted self-certification of claims on a trial basis. 

      CHANGING TOLERANCES
---------------------------------------------------------- Letter :6.3

Other industry proposals to reduce administrative costs are under
consideration by RMA and the industry.  One proposal, increasing the
tolerances for acreage reporting and claims dollars paid, could
affect the program's financial soundness. 

Currently, insurance companies are required to seek reimbursement for
overpayments from farmers who incorrectly reported planted acres if
the overpayment exceeds $250.  The proposed change would not require
reimbursement for an error less than $500.  This change responds to
crop insurance companies' concerns that it costs more than $250 to
collect small overpayments. 

This proposal to increase tolerances could increase the amount of
claims filed and indemnities paid, according to an RMA official, once
farmers become aware of the higher tolerance.  This could adversely
affect the soundness of the crop insurance program.  However, it
could also reduce administrative costs and provide faster claims
service.  The costs and savings associated with this proposed change
have not been estimated. 

   CONCLUSIONS
------------------------------------------------------------ Letter :7

Recent changes to RMA's program for ensuring the accuracy of claims
payments appear to offer a good opportunity to maximize the limited
resources available to both the crop insurance companies and RMA. 
However, until RMA has sound information on the level of improper
claims payments, it cannot evaluate, among other things, the
effectiveness of the new quality control program or the impact of
these changes on the program effectiveness.  Accurate information on
improper payments is an important part of the reporting required by
FMFIA, FFMIA, and the Results Act.  Such information is also
fundamental for basic program management and proper control over the
millions of federal dollars spent each year for claims payments. 

   RECOMMENDATION
------------------------------------------------------------ Letter :8

We recommend that the Secretary of Agriculture require the
Administrator of the Risk Management Agency to evaluate the costs of
alternative methods for developing more accurate estimates of error
rates for claims payments and implement an alternative that would
improve the estimate at a reasonable cost to the federal government. 
Alternatives that could be considered include (1) having the Risk
Management Agency sample and analyze a sufficient number of claims to
make an estimate and (2) using the claims sampling done by the
insurance companies under the quality control program to make the
estimate. 

   AGENCY COMMENTS
------------------------------------------------------------ Letter :9

We provided copies of a draft of this report to USDA's Risk
Management Agency for review and comment.  We met with officials from
the Risk Management Agency, including the Agency Administrator.  USDA
generally agreed with the information provided in our report and the
report's conclusions and recommendation.  The agency provided a
number of technical changes and clarifications to the report, which
we incorporated as appropriate. 

   SCOPE AND METHODOLOGY
----------------------------------------------------------- Letter :10

To determine the extent to which crop insurance claims are paid in
error and compare this rate with the experience of other types of
insurance, we reviewed agency documentation and discussed with RMA
and insurance company officials their efforts to determine error
rates.  We also contacted a number of insurance companies to
determine how federal crop insurance claims paid in error compare
with other types of insurance, such as auto and casualty. 
Furthermore, we discussed with program officials other federal
programs, including Medicare and federal flood insurance, which
periodically measure their claims paid in error. 

To review RMA's internal controls over claims payments for crop
losses, we reviewed agency documentation and discussed with agency
and insurance company officials their efforts to improve their
quality control programs.  We also discussed with USDA's OIG its
recommendations for improving USDA's efforts in reducing crop
insurance claims paid in error.  Furthermore, we reviewed a series of
OIG reports dating from 1983 that have commented on the extent to
which crop insurance claims have been paid in error and on the
weaknesses in federal and insurance companies' controls over claims
payments for crop losses. 

To identify opportunities to reduce administrative requirements that
may contribute to erroneous claim payments and/or increase costs for
the insurance companies, we contacted National Crop Insurance
Services, Inc., an industry association for crop insurance companies;
several participating crop insurance companies; RMA; several private
insurance agents and farmers representing different regions of the
country; and associations representing major commodities insured by
federal crop insurance.  We discussed with these officials how a
proposed change might affect program simplicity, savings, and
soundness.  To identify changes already made, we contacted RMA
officials to obtain a summary of simplification. 

We performed our review from January 1999 through September 1999 in
accordance with generally accepted government auditing standards. 
Although we did not independently assess the accuracy and reliability
of USDA's computerized databases, we used the same files USDA uses to
manage the crop insurance program, which are the only available data. 

--------------------------------------------------------- Letter :10.1

We are sending copies of this report to Senator Richard Lugar,
Chairman, and Senator Tom Harkin, Ranking Minority Member, Senate
Committee on Agriculture, Nutrition, and Forestry; Representative
Larry Combest, Chairman, House Committee on Agriculture; and other
appropriate congressional committees.  We are also sending copies to
the Honorable Dan Glickman, the Secretary of Agriculture; the
Honorable Kenneth Ackerman, Administrator, the Risk Management
Agency; and the Honorable Jacob Lew, Director, Office of Management
and Budget; and other interested parties.  Copies will also be made
available to others upon request. 

Please contact me at a (202) 512-5138 if you or your staff have any
questions about this report.  Key contributors to this report were
Ronald E.  Maxon, Jr.; Sheldon H.  Wood, Jr.; Robert G.  Hammons; Jay
Scott; David A.  Rogers; and Carol Herrnstadt Shulman. 

Sincerely yours,

Robert E.  Robertson
Associate Director, Food
 and Agriculture Issues

*** End of document. ***