Crop Insurance: Additional Actions Could Further Improve Program's
Financial Condition (Chapter Report, 09/28/95, GAO/RCED-95-269).
Since the 1930s, the federal government has offered subsidized crop
insurance to farmers. After the program was expanded in 1980 to include
more crops and locations, however, it paid out about $3 billion more in
claims through 1994 than it received in premiums from farmers and the
federal government. To correct this imbalance, Congress required that,
by October 1995, the program reduce its projected ration to at least $1
in premiums to $1.10 in claims paid. In other words, insurance rates
were to cover at least 91 percent of the anticipated claims--termed
"91-percent adequacy." The Agriculture Department (USDA) estimates that
the government's costs for the program will total $1.5 billion for
fiscal year 1996. This report examines whether USDA (1) set the
insurance rates to achieve the requirement of 91-percent adequacy, (2)
reduced the losses caused by high-risk farmers, (3) based payments to
farmers for claimed losses on their actual production history, and (4)
set deadlines for farmers to buy crop insurance before planting their
crops.
--------------------------- Indexing Terms -----------------------------
REPORTNUM: RCED-95-269
TITLE: Crop Insurance: Additional Actions Could Further Improve
Program's Financial Condition
DATE: 09/28/95
SUBJECT: Insurance premiums
Insurance claims
Farm income stabilization programs
Agricultural production
Insurance losses
Farm subsidies
Insurance regulation
Grain and grain products
Financial management
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. *
* *
* A printed copy of this report may be obtained from the GAO Document *
* Distribution Facility by calling (202) 512-6000, by faxing your *
* request to (301) 258-4066, or by writing to P.O. Box 6015, *
* Gaithersburg, MD 20884-6015. We are unable to accept electronic orders *
* for printed documents at this time. *
**************************************************************************
Cover
================================================================ COVER
Report to the Ranking Minority Member, Committee on Agriculture,
Nutrition, and Forestry, U.S. Senate
September 1995
CROP INSURANCE - ADDITIONAL
ACTIONS COULD FURTHER IMPROVE
PROGRAM'S FINANCIAL CONDITION
GAO/RCED-95-269
Crop Insurance Program's Financial Condition
(150825)
Abbreviations
=============================================================== ABBREV
GAO - General Accounting Office
USDA - U.S. Department of Agriculture
Letter
=============================================================== LETTER
B-265825
September 28, 1995
The Honorable Patrick J. Leahy
Ranking Minority Member
Committee on Agriculture, Nutrition,
and Forestry
United States Senate
Dear Senator Leahy:
This report responds to your request that we examine the adequacy of
premiums to cover the claims projected to be paid in the crop
insurance program and the U.S. Department of Agriculture's actions
to reduce the losses caused by high-risk farmers, insure farmers on
the basis of their actual production history, and set deadlines for
farmers to purchase crop insurance before planting their crops. The
report contains a matter for consideration by the Congress and makes
a number of recommendations to the Secretary of Agriculture to
improve the management and financial condition of the crop insurance
program in these areas.
We are sending copies of this report to interested congressional
committees; the Secretary of Agriculture; and the Director, Office of
Management and Budget. We will also make copies available to others
upon request.
If you or your staff have any questions, I can be reached on (202)
512-5138. Major contributors to this report are listed in appendix
VI.
Sincerely yours,
John W. Harman
Director, Food and
Agriculture Issues
EXECUTIVE SUMMARY
============================================================ Chapter 0
PURPOSE
---------------------------------------------------------- Chapter 0:1
Since the 1930s, the federal government has helped farmers mitigate
the risks of farming by offering subsidized crop insurance. However,
after the crop insurance program was expanded in 1980 to include more
crops and locations, it paid out about $3 billion more in claims
through 1994 than it received in premiums from farmers and the
federal government. The ratio of income from premiums to claims paid
was thus about $1 to $1.40. The Congress required that, by October
1995, the program reduce its projected ratio to at least $1 in
premiums to $1.10 in claims paid. Stated differently, insurance
rates were to be set to generate income from premiums to cover at
least 91 percent of the anticipated claims payments--termed
"91-percent adequacy" in this report. The U.S. Department of
Agriculture (USDA), the agency responsible for administering the crop
insurance program, estimates that the government's costs for the
program will total $1.5 billion for fiscal year 1996.
Concerned about the financial condition of the crop insurance
program, the Ranking Minority Member of the Senate Committee on
Agriculture, Nutrition, and Forestry asked GAO to examine whether
USDA (1) set the insurance rates to achieve the legislative
requirement of 91-percent adequacy, (2) reduced the losses caused by
high-risk farmers, (3) based payments to farmers for claimed losses
on their actual production history, and (4) set deadlines for farmers
to purchase crop insurance before planting their crops.
BACKGROUND
---------------------------------------------------------- Chapter 0:2
Under the federal crop insurance program, restructured by the Federal
Crop Insurance Reform Act of 1994, farmers purchase insurance against
crop losses from private insurance companies with whom USDA has
contracted. Farmers choose both the proportion of the crop to be
insured and the unit price (e.g., per bushel) at which any loss is
calculated. They can choose to insure as much as 75 percent of their
normal production or as little as 50 percent. With respect to the
unit price, farmers choose whether to value their insured production
at the full USDA-estimated market price or at a lesser percentage of
the full price. USDA sets different premium rates for the various
coverage and production levels.
These rates vary by crop, location (county), farm, and farmer.
Consequently, hundreds of thousands of premium rates are in effect.
To set premium rates, USDA calculates a basic rate for each crop in
each county for the farmers who buy insurance at the 65-percent
coverage level and whose normal production level is about equal to
the average production in the county. From this basic rate, USDA
makes adjustments to establish rates for other coverage levels and
for those farmers whose production levels are higher or lower than
the county's average.
GAO examined the crop insurance program for six major crops insured
by USDA--barley, corn, cotton, grain sorghum, soybeans, and wheat.
These six crops have historically accounted for about 75 percent of
the claims paid.
RESULTS IN BRIEF
---------------------------------------------------------- Chapter 0:3
USDA has improved the overall financial condition of the crop
insurance program for the six crops reviewed by raising the premium
rates, but the basic rates, on average, still do not meet the
requirement of 91-percent adequacy set by the Congress. Moreover,
the adjustments to the basic rates for other coverage and production
levels are inaccurate. These problems in the rate structure are
compounded by USDA's recent decision to increase the benefits to
farmers who could not plant crops because of adverse weather
conditions.
USDA now sets higher rates for high-risk farmers, which will help to
reduce the government's losses. These higher rates will result in
savings for the program, although not as much as USDA has estimated.
USDA's estimate assumed a greater number of high-risk farmers than
have actually been charged the higher rates.
USDA has also made changes to more accurately calculate farmers'
production levels on the basis of historical experience. These
changes should result in basing insurance on more accurate levels of
production. However, two practices by USDA--limiting any reduction
in the farmers' insured production level to no more than 10 percent
annually and not routinely verifying the production history provided
by farmers--reduce confidence that USDA pays claims on the basis of
actual production levels.
Revised purchasing deadlines should better ensure that farmers buy
crop insurance before the planting season is under way. However,
USDA generally sets the same deadline for an area covering several
states rather than considering local growing conditions. As a
result, some farmers can more precisely evaluate growing conditions
at planting time and therefore are more likely to purchase crop
insurance only when growing conditions are poor.
PRINCIPAL FINDINGS
---------------------------------------------------------- Chapter 0:4
CHANGES IN PREMIUM RATES
HAVE IMPROVED PROGRAM'S
FINANCIAL CONDITION, BUT
SOME RATES REMAIN TOO LOW
-------------------------------------------------------- Chapter 0:4.1
On average, the basic premium rates for the six crops reviewed are 89
percent adequate for crop year 1995. However, the rates for some
crops and locations and for some coverage and production levels are
still too low. For the crops, the adequacy of the basic rate is the
lowest for corn at 81 percent. Corn is also the largest crop insured
under the program, accounting for over $250 million in annual
premiums. For locations, the adequacy of the basic rates varies.
For example, the basic rates in about half of the locations GAO
reviewed--accounting for about 24 percent of the premiums--were less
than 80 percent adequate. Even though the Congress allows rate
increases of 20 percent annually, USDA has not increased the basic
rates sufficiently to achieve the legislative requirement. USDA has
acted primarily out of its historical concern that higher rates would
drive farmers from the program. While this concern is legitimate,
inadequate increases in the basic rates result in continuing losses.
Although the basic rates are approaching 91-percent adequacy, the
majority of crop insurance is purchased at rates that are, with some
exceptions, too low. For example, at the 50-percent coverage level,
rates were about 11 percent lower than required. The rates were
inadequate for all crops at the higher production levels and
excessive for some crops at the lower production levels. For
example, at the higher production levels, the rates for cotton were
about 25 percent lower than required. In response to GAO's analysis
of the rates for various production levels, USDA said that it would
have an actuarial consultant review the rate-setting process.
While USDA has made improvements to the program's rate structure, in
June 1995 it undermined these efforts by providing a higher level of
benefits under the program's "prevented planting" provision. Under
this provision, farmers who could not plant crops because of adverse
weather conditions were formerly paid at 50 percent of the coverage
they purchased. USDA raised this level to 75 percent without first
adjusting the premium rates to account for this increase. USDA
estimates that this change will cost about $135 million in additional
claims. The Department had planned to recover the money paid on
these claims through future rate increases, but its Office of General
Counsel has since stated that the governing legislation does not
allow it to do so. Therefore, USDA cannot recover these funds.
USDA IS TAKING ACTION TO
IDENTIFY HIGH-RISK FARMERS
-------------------------------------------------------- Chapter 0:4.2
USDA has instituted a program to identify farmers who make frequent
and substantial claims so that it can increase their premiums and/or
reduce the production levels at which they are insured. USDA's
program for targeting high-risk farmers for rate increases is
generally sound and will reduce the government's outlays for crop
insurance. However, GAO estimates that the program will produce
initial savings of $33 million annually; this amount is less than
half of USDA's estimated savings of about $70 million. The
difference occurs because USDA assumed that it would have about
double the number of farmers in the high-risk program than it has
actually included.
CHANGES TO ESTABLISH
ACCURATE PRODUCTION LEVELS
ARE UNDERMINED BY LACK OF
VERIFICATION OF FARMERS'
PRODUCTION HISTORY
-------------------------------------------------------- Chapter 0:4.3
To ensure that the claims paid for losses are based on farmers'
normal production levels, USDA began in crop year 1994 to require
farmers to purchase insurance at actual production levels or on the
basis of a reduced estimate of production levels. This change will
help ensure that farmers do not purchase insurance for production
levels higher than they are likely to achieve and, as a result, make
claims for production losses that are not real. GAO estimates that
USDA's action should reduce the program's outlays initially by about
$44 million annually; this amount is about 60 percent of USDA's
minimum estimate of $75 million annually. The difference occurs
largely because USDA's estimate did not take into account the
decision to limit any reduction in farmers' insurable production
levels to 10 percent annually.
Furthermore, in implementing this change USDA has not addressed a
long-standing problem that GAO\1 and USDA's Inspector General have
previously identified. That is, USDA does not require that a loss
adjuster verify the accuracy of the production history supplied by
the farmer when adjusting the claim.
--------------------
\1 Crop Insurance: FCIC Should Strengthen Actual Production History
Program Controls (GAO/RCED-89-19, Dec. 15, 1988).
REVISED PURCHASING DEADLINES
REDUCE RISK TO THE
GOVERNMENT
-------------------------------------------------------- Chapter 0:4.4
By moving the deadlines for purchasing crop insurance 30 days earlier
in the year, as the Congress mandated, USDA has helped reduce the
chance that farmers will buy crop insurance only when they determine
that growing conditions are poor. However, USDA has not adjusted
these deadlines for local growing conditions. Consequently, for
about 12 percent of the deadlines GAO reviewed, farmers are still
able to purchase crop insurance close to or during the planting
period, improving their chances of predicting poor production.
MATTER FOR CONGRESSIONAL
CONSIDERATION
---------------------------------------------------------- Chapter 0:5
If the Congress wants to ensure the financial viability of the crop
insurance program, it may wish to prevent USDA from making policy
decisions about the program that are not funded under the crop
insurance program's rate structure. To do so, the Congress would
need to amend the Federal Crop Insurance Reform Act of 1994 to
specifically prohibit the Secretary of Agriculture from making policy
decisions that increase benefits without first increasing the rates
to cover the anticipated claims.
RECOMMENDATIONS
---------------------------------------------------------- Chapter 0:6
GAO is making a number of recommendations to the Secretary of
Agriculture to help improve the financial condition of the crop
insurance program. In particular, GAO recommends that the Secretary
direct the Deputy Administrator for Risk Management to annually raise
premium rates up to the 20 percent authorized by the Congress, if
needed to ensure that the rates meet the legislative requirement of
91-percent adequacy and cover future claims. As part of this
rate-setting process, the Deputy Administrator should develop an
annual report that shows the expected adequacy of premium rates each
year, by crop and by state, so that USDA's management and the
Congress can be kept informed of the program's financial condition.
If the premium rates are not raised as required, the Deputy
Administrator should include in the annual report the additional
amount by which it has subsidized farmers' purchase of crop insurance
when the rates are inadequate. GAO is also making recommendations on
verifying farmers' production history and setting purchasing
deadlines before the planting season.
AGENCY COMMENTS AND GAO'S
EVALUATION
---------------------------------------------------------- Chapter 0:7
GAO requested comments on a draft of this report from the Secretary
of Agriculture. GAO then met with officials of the Department,
including the Deputy Administrator for Risk Management, who provided
USDA's comments. Overall, USDA agreed with GAO's conclusion that the
basic premium rates for the 1995 crop year are 89 percent adequate.
However, USDA believes that the program's financial soundness has
been improved even more than these rates suggest when the other
changes, such as increasing the premiums for high-risk farmers and
improving the calculation of farmers' insured production levels, are
taken into account. GAO recognizes that the changes USDA has made
are improving the program's financial condition. GAO also recognizes
that these changes may offset many of the shortfalls in premiums
identified in this report. However, when the $135 million shortfall
resulting from the decision about prevented planting is included, the
net shortfall for the program as a whole is substantial. Although
USDA generally agreed with the recommendations to the Secretary of
Agriculture, it disagreed with GAO's recommendation that USDA
increase rates up to the 20 percent authorized by legislation. It
cited its concern that abrupt increases may discourage farmers from
purchasing more than the minimum mandatory level of crop insurance.
While GAO also recognizes this possibility, the premium rates for
many crop programs will continue to fall short of the legislative
requirement unless rates are raised as much as allowed. USDA's
comments and GAO's evaluation of them are discussed in chapters 2, 4,
and 5.
INTRODUCTION
============================================================ Chapter 1
The risks associated with natural disasters have always been a part
of farming. Historically, farmers assumed these risks as part of the
hazards of doing business. Since the 1930s, many farmers have been
able to transfer part of the financial losses from these risks to the
federal government through subsidized crop insurance. Before 1980,
the crop insurance program was smaller, covering fewer crops and
locations, and its premiums were generally adequate to pay the
claims. Since the program was expanded in 1980 to cover more crops
in more locations, it has not been financially stable, paying out
more in claims in most years than the premiums the farmers and the
government had paid in. To reduce the government's cost for the crop
insurance program, the Congress required that, by October 1, 1995,
the U.S. Department of Agriculture (USDA) lower the program's
projected losses from over $1.40 in claims paid for every $1 of
premiums taken in to $1.10 or less. In March 1994, USDA issued a
plan explaining how it expected to achieve the desired improvement.
HOW CROP INSURANCE WORKS
---------------------------------------------------------- Chapter 1:1
Federal crop insurance is a program that is relatively simple in
concept but highly complex in implementation. Farmers who buy crop
insurance can file claims\1 for part of the money that they would
otherwise lose when droughts, floods, infestations of insects, or
other natural disasters keep them from harvesting their normal
expected crop. The size of the claim depends on the extent of the
crop loss and the amount of insurance coverage\2 the farmer has
purchased.
Two types of coverage--catastrophic and additional--are available for
most major crops\3 under changes made by the Congress in 1994. Under
catastrophic coverage, the government provides a free minimum level
of coverage to farmers for a small processing fee.\4 The government
pays the premium for this insurance. Farmers must sign up for this
program if they sign up for the annual USDA commodity programs;
obtain USDA farm ownership, operating, or emergency loans; or
contract to place land in the Conservation Reserve Program. They can
sign up through their local Consolidated Farm Service Agency
office--the USDA agency responsible for administering the program--or
obtain their policy from a participating private insurance agent.\5
The free catastrophic program protects farmers against extreme
losses. The program pays farmers only when they are able to harvest
less than 50 percent of their normal crop. The normal crop is
determined on the basis of a farmer's past production history as
reported to the USDA office or insurance agent. If a farmer does not
report past production, that farmer's normal crop is determined by
using a modified average production level for the county,\6
reduced by a discount, because of the uncertainty of the farmer's
expected production. For losses in production\7 below the 50-percent
level, farmers are paid 60 percent of USDA's estimated market price.
Farmers can purchase additional insurance from participating private
insurance companies. As authorized by the 1980 act redesigning and
expanding the program (P.L. 96-365, Sept. 26, 1980), the managers
of USDA's crop insurance program have entered into reinsurance
agreements authorizing the participating insurance companies to sell
the insurance and process the resulting claims. The government pays
part of the farmers' premium. Farmers who purchase this additional
insurance must choose both the coverage level (the proportion of the
crop to be insured) and the unit price (e.g., per bushel) at which
any loss is calculated. With respect to level of production, farmers
can choose to insure as much as 75 percent of normal production
(25-percent deductible) or as little as 50 percent of normal
production (50-percent deductible) at different price levels.\8 With
respect to the unit price, farmers choose whether to value their
insured production at USDA's full estimated market price or at a
percentage of the full price. USDA sets the premium rates and
assigns correspondingly higher premiums for higher production and
price levels.\9
The following example illustrates how a claims payment is determined.
A farmer whose normal crop production averages 100 bushels of corn
per acre and who chooses to buy insurance at the 75-percent coverage
level will be guaranteed 75 percent of 100 bushels, or 75 bushels per
acre. Assuming that the farmer had chosen the maximum price coverage
and that USDA had estimated the market price for corn at $2 per
bushel, the farmer would have total coverage of $150 per acre.
Should something like drought cut the farmer's actual harvest to 25
bushels, the farmer will be paid for the loss of 50 bushels per
acre--the difference between the insured production level of 75
bushels and the actual production of 25 bushels. The insurance would
pay the farmer's claim at $2 x 50 bushels, or $100.
In addition, the crop insurance program's "prevented planting"
provision pays farmers who have purchased insurance but never planted
crops because of adverse weather conditions. These farmers are
entitled to claims payments ranging from 35 to 50 percent of the
coverage they purchased, depending on the crop.
--------------------
\1 USDA refers to claims payments as indemnities.
\2 USDA refers to insurance coverage as liabilities.
\3 USDA offers insurance for 51 major crops, which include about 400
subgroups of the 1,265 for which USDA paid disaster assistance in
1988-93. See Disaster Assistance (GAO/RCED-94-208R, May 12, 1994).
For crops for which USDA does not offer insurance, a program similar
to catastrophic insurance is provided at no cost to the farmer.
\4 The fee is $50 per crop per county, capped at $200 per county and
$600 for all the farmer's crops in all counties. USDA can waive the
fee in hardship cases.
\5 The crop insurance program is funded through the Federal Crop
Insurance Corporation, a wholly owned government corporation created
in 1938 (7 U.S.C. 1503). Under USDA's fiscal year 1995
reorganization, the corporation's employees are within the
Consolidated Farm Service Agency--the agency that now administers the
crop insurance program.
\6 This modified average is generally calculated by multiplying the
production level USDA has assigned the farmer for calculating
deficiency payments under the income support program by the result of
dividing USDA's records of the average production in the county over
10 years by the average production USDA assigned all farmers in the
county under the income support program.
\7 USDA refers to the production level as the yield.
\8 As noted above, the free catastrophic insurance insures 50 percent
of production at 60 percent of USDA's estimated market price. The
additional insurance allows farmers to increase coverage above the
60-percent price level. For crop year 1996, USDA will offer
production coverage at the 50-percent level, increasing in 5-percent
increments to 85 percent. However, the 80- and 85-percent coverage
will be offered only on a pilot basis.
\9 The amount of USDA's subsidy varies by the level of coverage the
farmer chooses. The maximum subsidy is calculated on the basis of
the 65-percent coverage level and 100-percent of USDA's estimated
market price. This subsidy is about 42 percent of the total premium.
USDA provides the same dollar subsidy for the 75-percent coverage
level.
INSURANCE PREMIUM RATES ARE
BASED ON RISK, WHICH TYPICALLY
VARIES BY LOCATION, FARM, AND
FARMER
---------------------------------------------------------- Chapter 1:2
Critical to the success of the crop insurance program is aligning the
premium rates with the risk each farmer represents. The riskiness of
growing a particular crop varies from location to location, from farm
to farm, and from farmer to farmer. If the rates are too high for
the risk represented, farmers are less likely to purchase insurance,
lowering the program's income from premiums. Conversely, if the
rates are too low, farmers are more likely to purchase crop
insurance, but because the rates are too low, the income from
premiums will be insufficient to cover the claims.
To align crop insurance premium rates with the risk represented, USDA
establishes rates that vary by crop, location (county), farm, and
farmer. Because of all the combinations involved, literally hundreds
of thousands of premium rates are in place. For this review, we
examined crop insurance rates at the state level for six major crops:
barley, corn, cotton, grain sorghum, soybeans, and wheat. For these
crops, the average premium rates for crop insurance purchased at the
65-percent coverage level in 1994 varied widely among the states. As
shown in figure 1.1, the average rates\10 ranged from a low of $1.95
per $100 of insurance coverage for wheat in one state to a high of
$32.94 per $100 of insurance coverage for soybeans in another state.
Figure 1.1: Highest, Lowest,
and Average Premium Rates for
Six Crops, 1994
(See figure in printed
edition.)
Note: Rates are at the 65-percent coverage level. The averages are
calculated by dividing the total premiums (including the government's
subsidy) by the total insurance coverage (the production level
multiplied by the coverage and price levels). The averages are
calculated for each state and for all states combined.
Source: GAO's analysis of USDA's data.
To adjust the hundreds of thousands of rates it publishes each year,
USDA goes through a multistep process involving considerable computer
analysis and judgment. USDA's objective is to set the rates that
each farmer pays according to the risk associated with the farmer's
location, crop, past production, and past losses. For the six crops
we reviewed, USDA begins its rate-setting process each year by
looking at the crop insurance experience over the past 20 years for
each county and state. On the basis of a county's and state's
historical experience, USDA sets a basic rate for each crop in each
county at the 65-percent coverage level for average production.
Using this basic rate, USDA makes adjustments to establish rates for
other coverage levels and for farmers whose production levels are
higher or lower than the county's average. This latter adjustment is
based on USDA's research showing that farmers with
higher-than-average production levels are less likely to experience
losses.
USDA aligns rates with risk in several other ways as well. For
example, it imposes an additional premium on those farmers who insure
individual fields\11 rather than all fields combined, purchase hail
insurance, and are high risk as evidenced by frequent and high
experience with claims. Moreover, for those farmers who have
production records for fewer years than required to establish the
amount of production that can be insured, USDA uses the modified
average production level for their county, adjusting the production
down according to the number of years for which the farmers have
provided records. USDA's rate-setting methodology is described in
more detail in appendix I.
--------------------
\10 USDA refers to this figure as the earned premium rate.
\11 USDA refers to this approach as unit coverage.
PROGRAM HAS HISTORY OF
FINANCIAL LOSSES
---------------------------------------------------------- Chapter 1:3
Since 1980, when the Congress redesigned and expanded the crop
insurance program to be the primary form of agricultural disaster
assistance, the program has not been financially sound. USDA has
regularly paid out more in claims than it received in premiums paid
by farmers and the government. Two key requirements of the 1980
legislation were to (1) operate the program on a financially sound
basis and (2) eliminate the need for government-funded disaster
assistance by having most farmers buy crop insurance.\12 The program
has never met either requirement.
First, to be financially sound, the program needed to realize more
income from premiums, including the government's subsidy, than it
paid to settle farmers' claims so that it could build up a cash
reserve to pay farmers' claims in years of catastrophic loss. As
shown in figure 1.2, the claims paid per $1 of premium (including the
government's subsidy) for crop years 1981 through 1994 varied greatly
from year to year, averaging $1.41. During this period, claims
exceeded premiums by a total of $3.3 billion. The highest claims
payments in relation to premiums were in 3 catastrophic
years--resulting from severe droughts in 1983 and 1988 and excessive
moisture and severe flooding in 1993. Excluding the 3 catastrophic
years, the average claim per dollar in premiums was $1.22. Thus,
even in years without catastrophic losses, the program consistently
operated at a loss; catastrophic years just made the situation worse.
Figure 1.2: Claims Paid per
Premium Dollar Collected, Crop
Years 1981-94
(See figure in printed
edition.)
Source: GAO's analysis of USDA's data.
Moreover, the Congress's goal of having most farmers buy crop
insurance to eliminate the need for direct government disaster
payments was not reached. Farmers never insured more than 40 percent
of their eligible acres, and the pressure for direct disaster
assistance continued. In fact, the Congress passed emergency
disaster legislation to cover several crop years in the 1980s and
each crop year from 1988 through 1993. Over the period 1981-93, USDA
paid farmers about $11 billion in disaster assistance payments.
Adding this to the government's $8 billion share of the cost of crop
insurance, the government's spending to assist farmers who lost crops
exceeded $19 billion over the 13-year period. Figure 1.3 depicts the
outlays by year.
Figure 1.3: USDA's Crop
Insurance and Disaster
Assistance Outlays, 1981-93
(See figure in printed
edition.)
Notes: Data on disaster assistance payments for 1988 through 1993
are by crop year. All other data are by fiscal year.
Source: GAO's analysis of USDA's data.
The crop insurance program's financial condition is influenced by
several key management activities that, taken together, determine
whether the program will produce sufficient income to cover claims.
These key activities are
setting appropriate premium rates,
setting and enforcing the rules for calculating a farmer's normal
production level,
establishing the periods when insurance can be sold, and
setting and enforcing the rules for adjusting claims.
Historically, these activities, individually and collectively, have
prevented the crop insurance program from reducing its losses to an
acceptable level. As we reported throughout the 1980s, USDA's crop
insurance program unsuccessfully attempted to achieve financial
soundness at the same time it was rapidly expanding to include more
crops and locations.\13 In 1993, the crop insurance program's acting
manager acknowledged to a congressional committee that during the
1980s, the agency had focused "solely" on improving participation in
the program and "sacrificed" actuarial soundness.
Moreover, we and USDA's Inspector General reported problems with the
private insurance companies' claims adjustments. In 1993, the
Inspector General estimated an overpayment rate for claims of about 9
percent--an improvement over the 16-percent overpayment rate in 1987
payments that we had previously reported.\14
Furthermore, we had previously identified inherent problems with crop
insurance and problems in the design of the crop insurance program
that made it exceedingly difficult for the program to be financially
sound.\15 Crop insurance is an inherently difficult proposition
because many weather-related hazards can reduce crop production over
large areas of the nation, thereby increasing the chance that a
substantial number of policies will require payments during the same
year. This widespread impact reduces the probability that financial
stability can be achieved because risk pooling--the concept that
limited premiums are paid by many to fund claims paid to relatively
few--is less likely to be successful if most of the insured farmers
simultaneously face severe losses.
For example, in the severe drought of 1988, 92 percent of the 34,773
crop insurance policies purchased by wheat farmers in North Dakota
and Montana resulted in payments for claims, as did 58 percent of the
65,159 policies purchased by corn farmers in Iowa, Minnesota, and
Illinois. Similarly, in 1993--a year with extensive moisture and
flooding--72 percent of the 71,131 crop insurance policies purchased
by Iowa and Minnesota corn farmers resulted in payments for claims,
as did 56 percent of the 54,909 policies purchased by soybean farmers
in the same two states.\16
Statutes and regulations designed to encourage participation in the
program have further limited USDA's ability to make the program
financially sound because they encourage participation at the expense
of appropriate rates. These provisions include (1) allowing all
farmers to participate regardless of risk (entitlement); (2) allowing
farmers to insure for production levels higher than would be expected
on the basis of their production history, thereby increasing the
likelihood that claims will be paid; (3) restricting USDA's ability
to increase premiums; and (4) allowing farmers more time to assess
current growing conditions before purchasing insurance, which enables
them to better determine the likelihood of loss and to purchase
insurance when that likelihood is high.
--------------------
\12 Beginning in the mid-1970s, disaster assistance has often been
provided on an ad hoc basis when widespread weather-related damage
occurs.
\13 Crop Insurance: FCIC Should Strengthen Actual Production History
Program Controls(GAO/RCED-89-19, Dec. 15, 1988); Crop Insurance:
Federal Crop Insurance Corporation Needs to Improve Decision-Making
(GAO/RCED-87-77, July 23, 1987); More Attention Needed in Key Areas
of the Expanded Crop Insurance Program (GAO/RCED-84-65, Mar. 14,
1984); and Concerns About the Actuarial Soundness of the Federal Crop
Insurance Program (letter dated Aug. 10, 1982).
\14 Federal Crop Insurance Corporation--Crop Year 1991 Claims, USDA
Office of Inspector General, Audit Report No. 05600-4-Te (Sept. 30,
1993); and Crop Insurance: Private Company Loss Adjustment
Improving, but Overpayments Still High (GAO/RCED-90-32, Nov. 7,
1989). USDA's estimate was based on a statistical sample of all
payments, whereas our estimate was based on a statistical sample of
payments over $20,000.
\15 Crop Insurance: Federal Program Faces Insurability and Design
Problems (GAO/RCED-93-98, May 24, 1993).
\16 Such widespread claims far exceeded the premiums paid. For crop
year 1993, farmers in Iowa and Minnesota received payments for
damages to their corn and soybean crops totaling $542 million. This
amount was about 6 times more than the $97 million in premiums
collected (including the government's subsidy). It would take
another 5 years of the same amount of premiums, without any claims,
to collect enough premiums to equal the claims paid in 1993.
THE CONGRESS ENACTED VARIOUS
MEASURES TO IMPROVE CROP
INSURANCE PROGRAM'S FINANCIAL
SOUNDNESS
---------------------------------------------------------- Chapter 1:4
As a result of persistent problems and high costs in the delivery of
crop insurance to farmers, potential reform of the crop insurance
program was a major focus in developing the 1990 farm bill. However,
congressional and administration officials were unable to reach
agreement on a design for the crop insurance program that fostered
high participation, eliminated the need for expensive ad hoc disaster
assistance legislation, and stayed within budget guidelines.
Consequently, in the 1990 legislation the Congress reemphasized the
need for the crop insurance program to achieve financial soundness by
mandating that USDA raise the premium rates, where necessary.
However, the Congress limited the increase for any farmer to no more
than 20 percent per year.
Continuing to be concerned about the losses in the crop insurance
program, the Congress, in the Omnibus Budget Reconciliation Act of
1993, directed USDA to improve the crop insurance program's financial
condition. The act required USDA, by October 1, 1995, to lower the
program's projected losses (loss ratio) from an average of over $1.40
paid in claims for every $1.00 of premium taken in down to $1.10. In
response to the legislation, USDA developed a blueprint explaining
how it expected to improve the program's financial condition by
reducing losses to the level specified in the legislation.
In October 1994, the Congress made additional changes. Under the
Federal Crop Insurance Reform Act of 1994 (P.L. 103-354, Oct. 13,
1994, title I), the Congress combined the existing crop insurance
program and the new catastrophic insurance program for which USDA
pays the farmers' premiums. By adding the catastrophic coverage, the
Congress planned to eliminate the need for ad hoc, emergency disaster
assistance for crop losses. This change should resolve the inherent
conflict in the program between expanding participation and achieving
financial soundness. The legislation also repeated the requirement
that USDA lower the projected loss ratio to $1.10 in claims paid for
every $1 in premiums on and after October 1, 1995. This requirement
remains in effect through September 30, 1998; thereafter, the amount
paid in claims must be reduced to $1.075 for every $1 in premiums.
The act also specifically provided that USDA establish insurance
rates that will fulfill the requirement for 1998.
The estimated cost of the integrated program, according to USDA's
budget request for fiscal year 1996, is $2.1 billion, which will be
partially offset by about $600 million in premiums paid by farmers.
Thus, the net cost to the government is estimated at $1.5 billion.
The estimated outlays consist of about
$1.6 billion in payments of claims to farmers and
$500 million for USDA's and the insurance companies' operating and
delivery costs.
USDA'S BLUEPRINT DESCRIBES PLAN
FOR ACHIEVING IMPROVED
FINANCIAL CONDITION
---------------------------------------------------------- Chapter 1:5
In response to the 1993 legislation, USDA released its Blueprint for
Financial Soundness on March 2, 1994. USDA described 18 initiatives
intended to improve the financial stability of the crop insurance
program. USDA had started most of these initiatives before the
legislation was enacted. The initiatives most critical to promoting
the success of the crop insurance program are setting appropriate
rates, charging higher rates to high-risk producers. establishing
accurate production levels, and setting appropriate deadlines for
purchasing insurance.
In September 1994, USDA contracted with the actuarial firm of
Milliman and Robertson to perform an overall evaluation of its
rate-setting process. This review is expected to be completed by
September 1996. The last comprehensive review of USDA's rate-setting
methodology was completed in 1983 by the same firm.
OBJECTIVES, SCOPE, AND
METHODOLOGY
---------------------------------------------------------- Chapter 1:6
Concerned about the financial condition of the crop insurance
program, the Ranking Minority Member of the Senate Committee on
Agriculture, Nutrition, and Forestry asked us to examine whether USDA
(1) set insurance rates to achieve the legislative requirement of
collecting premiums sufficient to cover 91 percent of the claims
paid--termed "91-percent adequacy" in this report; (2) reduced the
losses caused by high-risk farmers; (3) based claims payments on
farmers' normal production levels; and (4) set deadlines for farmers
to purchase crop insurance before planting begins. These activities,
taken together, substantially determine the program's financial
soundness.
As part of our review, we examined an initial draft of USDA's
blueprint. On the basis of this analysis, we briefed crop insurance
program officials on actions that we believed could be taken to
reduce the program's losses. In response, USDA added more specific
time frames for accomplishing tasks.
To determine the extent to which USDA's premium rates for crop
insurance were adequate under the legislative requirement, we met
with crop insurance program officials at USDA's headquarters in
Washington, D.C., the Department's main crop insurance field office
in Kansas City, Missouri, and selected regional service offices. We
reviewed USDA records and past studies to understand the Department's
actions to set premium rates. We also obtained USDA's computer files
for crop insurance to evaluate the adequacy of the rates.
In addition, we interviewed insurance representatives from the
private sector and reviewed insurance literature. We also reviewed
previous reports by GAO and USDA's Inspector General.
For our review, we evaluated the adequacy of the premium rates for
1991-95 for six of the seven major crops insured by USDA. We
selected these six crops for review because they were the largest
programs for which USDA used the same methodology to set the rates.
For 1994, the income from premiums for these six crops totaled about
$721 million.\17 For these six crops, the losses experienced were at
about the same level--$1.37 compared with $1.41 in claims payments
for each $1 in income--as in the overall program for the period 1981
through 1994. As shown in table 1.1, the six crops account for 74
percent of the claims paid and 76 percent of the premiums collected
under the program.
Table 1.1
Premiums and Claims Payments for Crops
Reviewed Compared With All Crops, Crop
Years 1981-94
(Dollars in millions)
Percenta Percenta
ge of Claims ge of
Crop Premiums total paid total
------------------------------ -------- -------- -------- --------
Wheat $1,508 19 $2,333 21
Cotton 772 10 1,148 10
Corn 2,002 25 2,175 19
Sorghum 219 3 336 3
Soybeans 1,383 17 2,006 18
Barley 196 2 326 3
Total for 6 crops reviewed $6,079 76 $8,324 74
Other 45 crops $1,890 24 $2,919 26
======================================================================
Total for 51 crops $7,969 100 $11,242 100
----------------------------------------------------------------------
Note: Columns may not add to total because of rounding.
Source: GAO's analysis of USDA's data.
In the absence of a USDA annual or periodic evaluation showing how
the rates it establishes each year compare with the rates that its
historical data indicate are needed to pay future claims, we
developed benchmark rates to measure the adequacy of USDA's basic
premium rates that it sets at the 65-percent coverage level and
average production level. We developed the benchmark rates by
generally following USDA's methodology for setting premium rates.
USDA uses the past 20 years' claims experience to set its rates each
year. USDA believes that the past 20 years' claims experience
provides the basis for setting rates each year that are needed to
produce sufficient income from premiums to pay future claims.\18
For example, if claims payments averaged $100 over the past 20 years
and the insurance sold averaged $1,000 in coverage, the benchmark
rates would be 10 percent of the amount of the insurance coverage
sold, or $10 per $100 of coverage. Although the future claims paid
would vary from year to year, they would be expected to average about
$100 per year. Thus, to achieve a rate that is 91 percent adequate,
USDA would need to set the rate at $9.10.
Following USDA's methodology, we used 20 years of historical data for
the insurance claims paid and insurance coverage sold to calculate a
benchmark premium rate for each crop in each county and state.
Because USDA sets its basic rates at the county level on the basis of
the historical experience in the county and state, we calculated
benchmark rates for each crop overall, weighting the county
experience to the state crop, national crop, and national level (six
crops combined). We then compared these benchmark rates with USDA's
basic premium rates for the year reviewed to assess the adequacy of
USDA's rates. Appendix II provides more detail on our methodology.
Appendix III lists the results of our analysis by crop, state, and
year.
USDA applies mathematical factors to its basic rates to set rates for
coverage and production levels above and below those used to set the
basic rates. To determine the accuracy of these other rates, we
compared the relative losses at the various levels over the period
1990 through 1994. Appendix II provides more detail on our
methodology.
To evaluate the effectiveness of USDA's program to target high-risk
farmers for individual rate increases, we identified the
policyholders that USDA targeted for 1993, the most recent
information available at the time of our analysis. We used USDA's
historical results from 1992 to estimate the reductions in claims and
increases in premiums that would result from targeting high-risk
farmers. Appendix IV provides more information on our methodology.
To determine the effectiveness of USDA's revised rules for estimating
a farmer's expected production level, we analyzed USDA's experience
for crop year 1994. We calculated the difference between the
production level each farmer qualified for in 1994 and the production
the farmer would have qualified for if the 1993 rules had continued.
Appendix V provides more information on our methodology.
To determine whether USDA's deadlines for purchasing crop insurance
were appropriate, we determined the extent to which USDA permitted
farmers to purchase crop insurance after the planting period had
begun. We compared the deadlines for purchasing insurance with the
initial date USDA establishes for planting. We briefed USDA
officials on our initial comparison, showing them that many deadlines
needed to be set earlier. They included in their blueprint a plan
for changing these deadlines. We compared these revised deadlines
with the initial dates set for planting.
We conducted our review from August 1993 through August 1995 in
accordance with generally accepted government auditing standards.
Although we did not assess the accuracy and reliability of USDA's
computerized databases, we used the same files that USDA uses to set
its rates.
--------------------
\17 In 1994, USDA, for its 51 crops, collected premiums totaling $949
million for $13.6 billion of insurance coverage on about 800,000
policies.
\18 The governing legislation does not permit USDA to set rates to
recoup previous losses.
CHANGES IN PREMIUM RATES HAVE
IMPROVED PROGRAM'S FINANCIAL
CONDITION, BUT MANY RATES REMAIN
TOO LOW
============================================================ Chapter 2
For the six crops we reviewed, the basic premium rates are, on
average, approaching the level necessary to achieve the legislative
requirement of 91-percent adequacy. The basic rate is set, by
county, for the 65-percent coverage level and the average production
level for each crop. However, for certain crops in certain states,
these basic rates remain too low. USDA has generally not raised
rates sufficiently because it was concerned that higher rates would
reduce sales of crop insurance.
While the basic rates are approaching the 91-percent adequacy
requirement, the rates for coverage higher or lower than the basic
rates have not been set to ensure that premiums are aligned with
risk. Most farmers purchase crop insurance coverage at these other
rates. USDA has not adjusted the mathematical factors applied to the
basic rates to calculate these other rates because of the time and
resources required. However, USDA is currently reviewing these
factors.
Finally, while the program has been moving in the direction of
adequate income to cover 91 percent of the claims paid, USDA recently
made a decision that further calls into question the program's
ability to meet that requirement. USDA increased the benefits
provided under the program's "prevented planting" provision for crop
year 1995 without first adjusting the premium rates. USDA
acknowledges that this change will result in payments of up to $135
million in claims.
BASIC RATES, ON AVERAGE, ARE
NEARLY ADEQUATE TO ACHIEVE
LEGISLATIVE REQUIREMENT
---------------------------------------------------------- Chapter 2:1
According to our analysis of the basic premium rates USDA established
for the six crops reviewed, the rates overall are nearly adequate to
meet the Congress's legislative requirement of charging premiums that
are projected to cover at least 91 percent of claims--resulting in $1
in income from premiums for every $1.10 paid in claims.
As figure 2.1 shows, USDA's basic rates for the six crops reviewed
were about 84 percent adequate overall in 1991, and this percentage
increased slightly in the following years. The rates in 1994 and
1995 were just below the requirement of 91-percent adequacy. In
1995, the rates were 89 percent adequate, meaning that USDA should
receive about $0.98 in income for every $1.10 in claims paid.
Figure 2.1: Adequacy of USDA's
Premium Rates for Six Crops
Combined Compared With
Legislative Requirement,
1991-95
(See figure in printed
edition.)
Note: The adequacy of the premium rates is measured at the
65-percent coverage level and average production level.
Source: GAO's analysis of USDA's data.
BASIC RATES FOR SOME CROPS AND
STATES DO NOT MEET LEGISLATIVE
REQUIREMENT
---------------------------------------------------------- Chapter 2:2
While the overall basic rate is approaching the requirement of
91-percent adequacy for the six crops combined, the ultimate
achievement of this requirement is being hampered because the basic
rates for some crops are not adequate. Furthermore, the basic rates
in many states are not adequate. USDA did not raise the rates for
these programs as much as it could have because of concern that
higher rates would discourage farmers from buying crop insurance.
RATES FOR SOME CROPS DO NOT
MEET REQUIREMENT
-------------------------------------------------------- Chapter 2:2.1
As table 2.1 shows, USDA's basic premium rates for some crops in 1995
are still well below the 91-percent requirement. For the six crops
reviewed, the rates for cotton and soybeans exceed the requirement of
91 percent, while the others fall short. In fact, the corn
rates--accounting for 37 percent of the crop insurance business for
the six crops--were the farthest from the requirement at 81 percent.
This shortfall occurred because 1993 (when claims payments were very
high) was added to the rolling 20-year database used for setting
rates and 1973 (when claims payments were lower) was deleted, without
a corresponding increase in the premium rates.
Table 2.1
Adequacy of 1995 Premium Rates for Six
Major Crops
1995
Number Rate rate Adequa
of needed charge cy of
Total state per d per 1995
premiums crop $100 $100 rate
(million progra covera covera (perce
Crop s)\a ms\b ge ge nt)
---------------------------- -------- ------ ------ ------ ------
Cotton $130 17 $16.53 $16.87 102
\c
Soybeans 128 31 5.24 $4.92 94
Wheat 158 38 9.63 8.36 87
Grain sorghum 22 25 10.80 9.39\c 87
Barley 14 29 11.19 9.62 86
Corn 268 43 6.64 5.37 81
Total/average $721 183 8.25 7.34 89
----------------------------------------------------------------------
Notes: Columns do not add to total because of rounding. Rates are
at the 65-percent coverage level and average production level.
\a Premiums are for crop year 1994 because 1995 was not complete at
the time of our review.
\b As discussed in chapter 1, we based our analysis of the adequacy
of the rates for each crop in each state, which we identify as one
state crop program.
\c From 1991 to 1995, the basic rate increased by 30 percent for
cotton and by 35 percent for grain sorghum. These were the largest
increases for the six crops we reviewed.
Source: GAO's analysis of USDA's data.
Our analysis of the adequacy of the basic rates is consistent with
USDA's blueprint, which stated that only about 30 percent of the
crops the Department analyzed met the required level of adequacy.
The results of both our and USDA's analysis depend heavily on the
number of years included and the weight assigned to each year. For
example, in 1983 USDA's consultant suggested changing from the
current methodology of giving equal weight to each year of the 20
years' experience to giving greater weight to more recent years'
experience. Specifically, the consultant suggested assigning a
50-percent weight to the experience for the most recent 10 years and
a 50-percent weight to the experience for all available years.\1 We
found that the consultant's approach had a significant impact on the
premium rates for three crops. For soybeans, barley, and wheat, the
adequacy was reduced from 94 to 87, 86 to 77, and 87 to 76 percent,
respectively. The impact is greatest on these three crops because of
changes in the level of losses that have occurred in the most recent
10 years. In response to our evaluation, USDA's senior actuary for
crop insurance told us that the Department will have its actuarial
consulting firm evaluate whether the trend in losses in recent years
requires a change in USDA's methodology. He said this evaluation
would be completed in late September 1995.
--------------------
\1 According to an expert on actuarial science, Charles L.
McClenahan, the period of experience selected involves judgment, but
whatever period is used must be representative. He noted that there
"is a natural preference for using the most recent incurred
experience available since it is generally most representative of the
current situation." Foundations of Casualty Actuarial Science (New
York: Casualty Actuarial Society, 1990), pp. 40-41.
RATES SET FOR MANY STATES
ARE INADEQUATE
-------------------------------------------------------- Chapter 2:2.2
For the 183 state crop programs\2 we examined, only 54 had basic
rates that were at least 91 percent adequate for 1995. These 54
programs were generally those that had the greatest volume of
insurance. For the remaining 129 programs, 40 were approaching
91-percent adequacy--ranging from 80 to just under 91 percent. The
other 89 programs, representing about 24 percent of the crop
insurance premiums for the six crops in 1994, had basic rates that
were less than 80 percent adequate. As table 2.2 shows, many of
these 89 programs had not charged adequate rates for the entire
1991-95 period. As the table also shows, the size of these state
programs varied significantly, from as low as $100 in premium income
annually to as much as $61 million.
Table 2.2
State Crop Programs With Premium Levels
That Were Less Than 80 Percent Adequate,
Crop Years 1991-95
Number of
state crop Total
Year programs Low High premiums
---------------------- ---------- ---------- ---------- ----------
1995\a 89 $100 $61,372,06 $173,828,0
1 60
1994 76 100 6,559,380 53,261,819
1993 79 240 4,476,928 43,581,227
1992 75 181 5,088,520 49,947,988
1991 80 861 6,028,470 48,817,991
----------------------------------------------------------------------
\a Because the amount of premiums for 1995 is not yet complete, we
used the amount of 1994 premiums for each state to provide
perspective on the size of the state crop programs.
Source: GAO's analysis of USDA's data.
The increase in the number of programs that are less than 80 percent
adequate for 1995 resulted in part from the addition of large corn
programs in four states, totaling about $119 million in premiums.
These four programs had been 80 percent or more adequate--often more
than 90 percent adequate--in 1991-94 but this percentage dropped
dramatically in 1995. This drop occurred because (1) the severe
losses in 1993 were added to the historical database for establishing
the 1995 rates and a year from the 1970s when losses were lower was
deleted and (2) USDA did not increase the rates as much as it could
have for corn in these four states.
Our analysis of the adequacy of the basic rates is consistent with
USDA's blueprint, which stated that some areas of the country met the
legislative requirement while others did not.
--------------------
\2 We examined the adequacy of the rates for each of the six crops in
each state where insurance is offered--such as the average rate for
corn in Iowa--and refer to each as a "state crop program."
USDA DID NOT RAISE RATES AS
MUCH AS IT COULD HAVE
WITHOUT EXCEEDING
LEGISLATIVE LIMIT
-------------------------------------------------------- Chapter 2:2.3
For the state crop programs that were less than 80 percent adequate,
USDA often did not sufficiently increase the basic rates where
necessary. Rates that are less than 80 percent adequate in any year
would require at least a 14-percent increase (of 80 percent) to reach
the 91-percent adequacy requirement. USDA did not always raise the
rates sufficiently even though most of the increases imposed were
less than the 20-percent statutory maximum.
USDA increased the rates most in 1992 and least in 1993, as shown in
figure 2.2. In 1992, USDA increased 71 percent of the rates for
state crop programs by 10 percent or more. In contrast, in 1993 USDA
increased the rates for only 10 percent of the state crop programs by
10 percent or more, while increasing the rates for 68 percent of the
state crop programs by less than 5 percent. For 1995, USDA again
moved towards greater increases by raising the rates for 59 percent
of the state crop programs by 10 percent or more.
Figure 2.2: Rate Increases for
State Crop Programs That Were
Less Than 80 Percent Adequate
in Crop Years 1995 and 1991-94
(See figure in printed
edition.)
Note: Rate increases and rate adequacy are measured at the 65-
percent coverage level and average production level.
Source: GAO's analysis of USDA's data.
USDA has not sufficiently raised rates out of concern that higher
rates will discourage farmers from buying crop insurance. For
example, in 1994 the crop insurance program manager testified that
USDA did not want to cause "sticker shock" and drive away the farmers
who are buying crop insurance.\3 He said that USDA was "trying to
raise rates in a relatively gentle way--10 percent instead of 20
percent a year--to phase them in." Similarly, USDA's blueprint stated
that increasing the rates to the levels suggested by experience in
the most recent 20 years may not be good public policy and "extremely
high premium rates will preclude realization of the social benefits
and public policy goals of the program because participation will be
discouraged."
We recognize that rate increases could cause some farmers to limit
their insurance coverage to the free catastrophic insurance program
because they conclude that the additional insurance program is not to
their financial advantage. However, as long as USDA sets rates that
are less than 91 percent adequate, it will not have the premium
income necessary to ensure that it meets the legislative requirement
of $1 in premiums for each $1.10 in claims paid. Furthermore, USDA
does not routinely evaluate and report on the adequacy of its rates.
As a result, USDA does not calculate the expected shortfall between
the income from premiums and the claims paid.
--------------------
\3 Testimony before the Subcommittee on Agriculture, Rural
Development, Food and Drug Administration, and Related Agencies,
House Committee on Appropriations, Mar. 17, 1994.
RATES FOR SOME LEVELS OF
COVERAGE AND PRODUCTION ARE NOT
ADEQUATE TO MEET LEGISLATIVE
REQUIREMENT
---------------------------------------------------------- Chapter 2:3
While establishing appropriate basic rates is critical to the
financial condition of the crop insurance program, the majority of
all insurance is purchased at rates for coverage and production
levels that are above or below those covered under the basic rates.
For this insurance, our analysis showed that in relationship to the
basic rates, the rates are
too high for coverage at the 75-percent level and too low at the
50-percent level,
too low at the higher levels of production, and
either too high or too low for the lower levels of production,
depending on the crop.
As a result, the rates for both coverage and production levels are
not aligned with risk. This occurs because USDA does not
periodically review and update the calculations it uses to adjust
rates above and below the basic rate.
RATES CHARGED FOR DIFFERENT
LEVELS OF COVERAGE ARE NOT
ALIGNED WITH RISK
-------------------------------------------------------- Chapter 2:3.1
To set the rates for the 75-percent and 50-percent coverage levels,
USDA applies preestablished mathematical factors to the basic rate.\4
However, these factors have not resulted in rates that are aligned
with risk. According to our analysis, the rates were too high at the
75-percent coverage level and too low at the 50-percent coverage
level in relationship to the basic rates.\5
For crops insured at the 75-percent coverage level, USDA set premium
rates ranging from 19 to 27 percent more than required. (See table
2.3.) As a result, the 1994 income from premiums was about $30
million more than required for this coverage. Although grain sorghum
had the greatest percentage of rates in excess of those required,
corn had the greatest amount of additional premium income because the
program was much larger.
Table 2.3
Amount That Premiums Were Greater Than
Needed for Crop Year 1994 for 75-
Percent Coverage Level
(Dollars in millions)
Percentage
that premium
rates were Total Amount that
greater than premiums for premiums
required 75-percent were greater
relative to coverage than
Crop basic rates level required
---------------------------- ------------ ------------ ------------
Grain sorghum 27 $1.8 $0.5
Cotton 27 3.3 0.9
Barley 26 3.2 0.9
Wheat 25 29.1 7.2
Soybeans 21 33.3 6.9
Corn 19 75.5 14.1
======================================================================
Total 21 $146.4 $30.3
----------------------------------------------------------------------
Note: Totals do not add because of rounding.
Source: GAO's analysis of USDA's data.
For crops insured at the 50-percent coverage level, the rates were
about 11 percent too low, resulting in a shortfall in premium income
of about $3 million for crop year 1994. The impact was much less
than at the 75-percent coverage level because only about $30 million
in insurance was sold at the 50-percent coverage level. However, the
potential impact of setting rates too low for the 50-percent coverage
level is much greater for future years. Beginning in crop year 1995,
USDA provided free catastrophic insurance to farmers at the
50-percent coverage level. In its fiscal year 1996 budget request,
USDA estimated that it will need $350 million to cover its costs to
pay these premiums for all crops. Assuming the six crops we reviewed
represent about 75 percent of the free insurance provided--the
proportion of the program they have historically represented--then
about $263 million of USDA's estimate is for these crops. Since the
50-percent coverage rate is 11 percent too low, USDA's budget request
could be understated by about $29 million.
--------------------
\4 USDA multiplies the basic rate at the 65-percent coverage level by
154 percent to arrive at the rate for 75-percent coverage and by 72
percent to arrive at the rate for 50-percent coverage.
\5 If the factors USDA applies to the basic rate to calculate the
rates at the 75- and 50-percent coverage levels are correct, the loss
ratios for each coverage level should be about the same over a number
of years.
RATES CHARGED FOR DIFFERENT
LEVELS OF PRODUCTION ARE NOT
ALIGNED WITH RISK
-------------------------------------------------------- Chapter 2:3.2
USDA also adjusts the basic rates for production, set at the county
average, for farmers whose historical production level is above or
below the county's average. As with the varying rates for coverage,
however, these adjustments do not result in rates that accurately
reflect the risk involved at each production level. Specifically,
according to our analysis the rates are too low for all crops at the
higher production levels and too high for some crops at the lower
production levels. The net effect is that premium income is too low.
The greatest dollar shortfall resulting from these problems occurred
in the cotton and corn programs.
USDA's basic rate applies to the farmer whose average production is
about equal to the average for all producers in the county. However,
many farmers' average production is above or below the county's
average, and USDA's research shows that the higher a farmer's
production level, the lower the chance of a loss. Therefore, USDA
establishes rates for different production levels using a
mathematical model that sets rates according to preestablished
relationships between production levels. The rates per $100 of
insurance coverage decrease as a farmer's average production
increases.
The mathematical model USDA applies to the basic rate to calculate
rates for production levels higher and lower than the coverage under
the basic rate does not result in correct rates. For above-average
production, USDA's rates should have been from 13 to 33 percent
higher than currently set. As shown in table 2.4, USDA needed an
additional $55 million in premium income in 1994 for the six crops.
Although barley would have required the greatest percentage increase
in premiums, cotton required the greatest amount of additional
premiums because the cotton program is much larger.
Table 2.4
Additional Premiums Required for Crop
Year 1994 for Production Levels Above
Basic Rate
(Dollars in millions)
Percentage
that premium Total
rates were premiums for
too low in production
relation to levels above Additional
the basic the basic premiums
Crop rates rate needed
---------------------------- ------------ ------------ ------------
Cotton 25 $78.9 $19.7
Corn 13 99.9 12.5
Wheat 17 64.2 11.0
Soybeans 21 43.0 9.1
Barley 33 5.6 1.9
Grain sorghum 18 6.3 1.1
Total 19 $298.0 $55.4
----------------------------------------------------------------------
Note: Totals do not add because of rounding.
Source: GAO's analysis of USDA's data.
At below-average production, premiums were about evenly split for
1994 between crops with rates higher or lower than needed. Overall,
as shown in table 2.5, the premiums were only slightly too low for
this group.
Table 2.5
Additional Premiums Required for Crop
Year 1994 for Production Levels Below
Basic Rate
(Dollars in millions)
Percentage
that rates
were too low Total
or too high premiums for
(-) relative production Additional
to basic levels below premiums
Crop rates basic rate needed
---------------------------- ------------ ------------ ------------
Corn 10 $91.6 $9.5
Soybeans 5 41.5 2.2
Grain sorghum 10 11.0 1.1
Barley -14 5.0 -0.7
Cotton -10 36.6 -3.5
Wheat -11 53.7 -5.7
Total 1.2 $239.3 $3.0
----------------------------------------------------------------------
Note: Totals do not add because of rounding.
Source: GAO's analysis of USDA's data.
As our analysis shows, the inaccurate rates had the greatest impact
on income from premiums for cotton (a net shortfall of about $16
million) and for corn (a net shortfall of about $22 million). For
cotton, this shortfall occurred because farmers were allowed to
insure their crop at production levels higher than their historical
production levels, according to USDA officials.\6 As a result, a
greater volume of insurance was sold at higher production levels than
was warranted on the basis of the farmers' experience. Beginning in
1994, USDA changed the requirements for calculating farmers'
production levels so that the amount of production insured would be
more closely aligned with the farmers' actual production history.
According to USDA officials, this change should result in cotton
farmers' purchasing insurance at reduced production levels. However,
as with other types of crops, USDA does not require cotton farmers to
decrease the amount of production coverage by more than 10 percent
per year until their coverage coincides with their actual production
experience.
For corn, the shortfall occurs because the rates for production above
and below the basic rates were both too low, according to our
analysis. This situation indicates that the mathematical model is
not appropriate for corn.
--------------------
\6 Beginning in 1989, USDA allowed cotton farmers to insure at the
highest production level resulting from three different calculation
methodologies, including historic production. However, USDA
officials told us that most farmers purchased insurance at production
levels calculated by the other methodologies.
USDA DOES NOT PERIODICALLY
REVIEW AND UPDATE FACTORS
FOR CALCULATING DIFFERENT
COVERAGE AND PRODUCTION
RATES
-------------------------------------------------------- Chapter 2:3.3
The misalignment of rates with risk occurs because USDA has not
revised the factors it applies to the basic rate to arrive at
different coverage and production levels. USDA officials told us
that they had not had the time and resources to revise the factors
since they were established in the 1980s. Moreover, USDA's senior
actuary told us that they have not developed a plan for how often the
factors ought to be evaluated and updated.
Nonetheless, these officials said they were working to improve their
capability to set rates. USDA is changing its computer database to
enable it to more easily evaluate the crop insurance program's past
performance and set new rates. This effort is expected to be
completed in time for setting the 1997 crop rates. In September
1994, the Department contracted with an actuarial consulting firm to
evaluate its factors for adjusting basic rates to other coverage
levels. In addition, in response to our analysis of rates for
production levels, USDA's senior actuary said the Department will
have the consulting firm evaluate the accuracy of its mathematical
model and recommend any specific changes needed.
USDA'S EXPANSION OF BENEFITS
UNDER PREVENTED PLANTING
PROVISION FURTHER JEOPARDIZES
MEETING LEGISLATIVE REQUIREMENT
---------------------------------------------------------- Chapter 2:4
While USDA is taking a number of actions to improve the crop
insurance program's rate structure, it recently made a decision that
will weaken the program's financial condition. For 1995, USDA
increased the benefits provided under the prevented planting
provision of the crop insurance program. This decision will increase
the claims paid by at least $135 million for 1995, according to
USDA's estimates.\7
Under the prevented planting provision, included in crop insurance
policies beginning in crop year 1994, farmers who could not plant
crops because of adverse weather conditions could receive insurance
payments at 50 percent of the insurance coverage level they
purchased.\8 In June 1995, USDA expanded the coverage to 75 percent,
for crop year 1995 only.\9 In addition, for 1995 farmers who could
not plant the crop they insured but were able to plant a different
crop will receive 25 percent of their coverage level in insurance
payments, whereas in the past, they would not have received any
insurance payments.\10
USDA increased the coverage level even though the 1995 crop insurance
rates were not set to include this additional coverage. USDA
officials recognized that this decision would hurt the program's
financial condition. Specifically, USDA's decision memorandum states
that the decision
". . . could arguably be seen as stretching the statute's
requirement that Federal Crop Insurance Corporation . . .
cannot make changes which adversely impact actuarial soundness
and must achieve a loss ratio of 1.10 by October 1, 1995."
In advising on this decision, USDA's Office of General Counsel said
that in determining rates and coverages, the manager of the program
"should make the specific determination that the action will not
adversely affect the `actuarial soundness' of the program."
Despite this advice, the decision memorandum recommended the change,
while recognizing its increased cost to the program. USDA's Acting
Deputy Administrator for Risk Management said that USDA's decision
was based on broader policy concerns that had to be considered along
with actuarial concerns. As crop year 1995 progressed, many farmers
were prevented from planting the crop they had insured and were
uncertain about the benefits. USDA believes that farmers were
confused about the program's requirements and restrictions because of
the rapid expansion of the crop insurance program in crop year 1995.
Moreover, in offering prevented planting coverage for the first time
in crop year 1994, USDA recognized that changes would be required in
future years as it gained experience with this provision. Also, USDA
concluded that it needed to correct an inconsistency in its coverage
that resulted in three different levels of claims payments for
farmers similarly affected by excessive moisture.\11 USDA was
concerned that if changes were postponed, farmers might not accept
the new crop insurance program and might call upon the Congress to
revise it. Therefore, USDA concluded that changes to the prevented
planting program were needed immediately.
According to the decision memorandum, the increased claims payments
were to be recovered beginning in crop year 1996. However, according
to USDA's Office of General Counsel, the governing legislation does
not permit USDA to set premium rates to recover past losses.
Instead, USDA can set rates only to cover anticipated future claims
payments. Therefore, USDA intends to include the $135 million in
claims payments in the historical database that it uses to calculate
future premium rates to cover estimated future claims payments.
--------------------
\7 This decision is described in a June 16, 1995, decision memorandum
for the Secretary of Agriculture. An attachment to the decision
memorandum estimates that this decision will result in additional
claims payments of $157.5 million. A more recent USDA estimate for
the two crops most affected (corn and wheat) estimates additional
claims of $135 million after deducting claims payments that would be
avoided.
\8 For hybrid seed corn, cotton, and rice, the percentage is 40, 35,
and 35, respectively.
\9 For hybrid seed corn, cotton, and rice, the percentage is 60, 52.5
and 52.5, respectively.
\10 For hybrid seed corn, cotton, and rice, the percentage is 20,
17.5, and 17.5, respectively.
\11 Specifically, farmers who planted an insured crop that failed
were eligible for claims payments coverage; farmers who were
prevented from planting an insured crop and did not plant a
subsequent crop were eligible for claims payments; but farmers who
were prevented from planting an insured crop and planted a substitute
crop were not eligible for any claims payments. USDA believed that
this third group should receive some amount of claims payments.
CONCLUSIONS
---------------------------------------------------------- Chapter 2:5
USDA has taken steps to improve the overall financial condition of
the crop insurance program for the six crops we reviewed by raising
the program's basic premium rates. On average, the basic rates are
approaching the 91-percent adequacy requirement the Congress set for
the program. However, this overall improvement masks some serious
problems in the basic rates set for some crops and in some states.
USDA recognizes the need to raise the basic rates, and it plans to
review its weighting methodology to ensure that the basic rates are
accurate. At the same time, because of concerns that farmers would
stop purchasing crop insurance, USDA has failed to raise the basic
rates promptly to ensure achievement of 91-percent adequacy. Keeping
farmers in the program is a legitimate goal. However, without
sufficient increases in the basic rates, the legislative requirement
cannot be met. Currently, USDA's management and the Congress cannot
project the program's losses because USDA does not annually evaluate
and report on the adequacy of the basic rates. Until that is done,
USDA and the Congress will be unable to routinely know whether the
program is meeting its legislative requirement, and, if not, what
adjustments need to be made to the basic rates.
In addition to the problems with the basic rates, USDA has not
adjusted the factors applied to the basic rates to arrive at accurate
rates for coverage and production levels different from those covered
by the basic rates. Most purchases of crop insurance occur at these
other levels. This lack of accurate rates benefits some farmers and
penalizes others: Farmers pay too much for coverage at the higher
coverage level and too little at the lower coverage level.
Similarly, farmers pay too little for production levels above average
and too much or too little for production levels below average,
depending on the crop. Ultimately, the crop insurance program loses
money. USDA has recognized that these rates will continue to be
incorrect because the mathematical factors it uses to set them are
incorrect. USDA officials stated that they have not had the time and
resources to periodically evaluate these factors. In response to our
analysis, USDA officials are evaluating the mathematical factors to
determine what changes are needed. However, these officials are not
developing a plan to periodically reevaluate whether these factors
continue to result in correct rates.
Finally, the difficulty in achieving the legislative requirement has
been compounded by USDA's recent program policy decision to increase
the coverage for prevented planting, even though USDA's Office of
General Counsel advised against it. This decision added an estimated
$135 million in claims payments that were not and cannot be recovered
through premium rates because the governing legislation prohibits it.
The prohibition raises further doubts about whether USDA's decision
to increase prevented planting levels was appropriate.
MATTER FOR CONSIDERATION BY THE
CONGRESS
---------------------------------------------------------- Chapter 2:6
If the Congress wants to ensure the financial viability of the crop
insurance program, it may wish to prevent USDA from making program
policy decisions that are not funded under the crop insurance
program's rate structure. To do so, the Congress would need to amend
the Federal Crop Insurance Reform Act of 1994 to specifically
prohibit the Secretary of Agriculture from making policy decisions
that increase benefits without first increasing the rates to cover
the anticipated claims.
RECOMMENDATIONS TO THE
SECRETARY OF AGRICULTURE
---------------------------------------------------------- Chapter 2:7
To meet the 1994 legislative requirement that USDA reduce losses and
set premiums to cover 91 percent of the claims paid, we recommend
that the Secretary of Agriculture direct the Deputy Administrator for
Risk Management to take the following actions:
Annually raise premium rates up to the 20 percent authorized by the
Congress, if needed, to cover future claims under the
legislative requirement of 91-percent adequacy. As part of this
rate-setting process, the Deputy Administrator should report the
expected adequacy of premium rates each year, by crop and by
state, so that USDA's management and the Congress can be kept
informed of the program's financial condition. If the rates are
not raised as required, USDA should include in its annual report
the estimated cost of subsidizing farmers' purchase of crop
insurance in areas where the rates are inadequate.
Develop and implement a plan for periodically evaluating the
mathematical factors used to set coverage and production levels
above and below the basic rates to ensure that these factors
continue to result in correct rates.
AGENCY COMMENTS AND OUR
EVALUATION
---------------------------------------------------------- Chapter 2:8
USDA made a number of comments on our findings and conclusions.
Overall, USDA agrees with our conclusion that the basic premium rates
for the 1995 crop year are 89 percent adequate. However, USDA
believes the program's financial soundness has been improved even
more than these rates suggest when the other changes, such as
increasing the premiums of high-risk farmers and improving the
calculation of farmers' insured production levels, are taken into
account. In addition, USDA noted that our analysis does not reflect
the likely influence of the rates for crop year 1996 on rate
adequacy. USDA believes its policy of gradual increases, coupled
with the slightly lower rates indicated for 1996 by the 20 years of
experience used to set them, should bring the 1996 rates closer to
the level required.\12 USDA believes that its actions, in
combination, should bring the 1996 crop insurance rates closer to
91-percent adequacy.
We recognize that some of the changes to the crop insurance program
discussed in this report are improving the program's financial
condition. We also recognize that the estimated savings from these
changes, as well as the excess premiums for the 75-percent coverage
level, may come close to offsetting the shortfalls in premiums that
we have identified.\13 However, when the $135 million shortfall
resulting from the prevented planting decision is included, the net
shortfall for the program as a whole is substantial.
In addition, we cannot determine the extent to which the 1996 premium
rates will further improve the program's financial soundness because
they are still being developed. However, in response to USDA's point
that the required 1996 premium rates will be more adequate because of
the change in the rolling 20-year database on which the rates are
based, we estimate that this change could raise the adequacy of the
rates. This assumes that the 1996 rates would, on average, be at
least as high as the 1995 rates. However, we estimate that the rates
could still be less than adequate for some crops unless the rates are
increased. For example, we estimate that the rate for corn would be
87 percent adequate without a rate increase, while the rate for wheat
would be 85 percent adequate.
USDA recognized that its decision to increase prevented planting
coverage for crop year 1995 added to the program's overall exposure
without a matching adjustment to 1995 premium rates, as our report
states. However, USDA said the report should recognize that its
decision was based on broad policy concerns that farmers were
suffering. We recognize USDA's position, but we still believe that
decisions with this magnitude of impact on the program's financial
soundness should be made with congressional consultation.
USDA disagreed with our recommendation that it raise rates by up to
the 20 percent authorized by legislation when needed but agreed with
our recommendation that an annual report showing the expected
adequacy of premium rates each year by crop and state was feasible.
It did not, however, clearly state whether it would prepare such a
report. With respect to our recommendation that rates be raised by
20 percent, USDA repeated its position, which we noted earlier, that
raising rates up to the maximum authorized should not be a standard
practice because abrupt increases may discourage farmers from
purchasing crop insurance. While we also recognize this possibility,
as we previously stated, unless rates are raised as much as allowed
when needed, the premium rates for many crop programs will continue
to fall short of the legislative requirement of 91-percent adequacy.
USDA also had several comments on a proposed recommendation in a
draft of our report that it report to the Congress as a part of its
budget request on the additional funds the program would need to
subsidize farmers' purchase of crop insurance when the rates are
inadequate. USDA questions whether this requirement should be a part
of the budget process because of the overlap in the preparation of
crop-year rates and fiscal year budget requests. Instead, USDA
believes that such information could appropriately be included in an
annual report to the Congress. We believe USDA's view has merit and
have revised our recommendation accordingly.
--------------------
\12 This difference occurs because the experience for crop year
1994--a year with very low claims payments--enters the 20-year
database used to set rates and the experience for crop year 1974--a
year with higher relative claims payments--is dropped. This was true
for each of the six crops we reviewed except wheat.
\13 In this analysis, we did not consider the savings from USDA's
program for high-risk farmers because the claims payment experience
of these farmers has been removed from the database for setting rates
for the crop insurance program.
USDA IS TAKING ACTION TO IDENTIFY
HIGH-RISK FARMERS
============================================================ Chapter 3
Beyond establishing a sound overall structure for premium rates,
aligning these rates with risk requires USDA to charge higher rates
to the individual farmers who present the highest insurance risk. To
accomplish this, USDA has instituted a program to identify those
farmers with frequent and substantial claims so that it can increase
their premiums and/or reduce the production levels they can insure.\1
Without this program, the overall rates would have to be raised more,
thereby penalizing lower-risk farmers. This in turn would make
lower-risk farmers less likely to purchase crop insurance and
contribute to reducing the program's financial stability. USDA's
program for targeting high-risk farmers for rate increases is
generally sound and will reduce the government's outlays for crop
insurance, although not by as much as USDA estimated.
--------------------
\1 USDA refers to the high-risk program as the Nonstandard
Classification System.
USDA'S ACTIONS TO IDENTIFY
HIGH-RISK FARMERS
---------------------------------------------------------- Chapter 3:1
The Department implemented the high-risk program in 1991 to reduce
the high losses associated with some farmers in the crop insurance
program. Over the period 1981 through 1989, USDA had found that
about 6 percent of the policies accounted for about 28 percent of the
total claims paid. The high-risk program improves the crop insurance
program's financial soundness by (1) reducing the production levels
at which high-risk farmers are insured and/or (2) charging high-risk
farmers increased rates that are more in line with their claims
history.
To be placed in this high-risk program, a farmer must
have received claims payments in at least 3 years, or if
information on more than 5 years' experience is available, in 60
percent of the years;
have had a cumulative adjusted loss ratio of about 4.0 or more
(i.e., $4 or more in claims paid for each $1 in premiums);\2 and
require a rate increase of at least 10 percent from the previous
year.
USDA has expanded the high-risk program from one crop in
1991--soybeans--to 37 crops in 1995. By 1993, the program included
11 crops that accounted for about 90 percent of the crop insurance
purchased.
In addition, in response to its 1994 appropriations legislation, USDA
developed a modified high-risk program for counties where losses were
high. These were counties that had paid out more than $1.10 in
claims for each $1 in premiums in 70 percent of the years (1980-92)
in which the crop program was offered.\3 To be placed in this
program, a farmer in these counties must
have received claims payments in at least 3 years, or if
information on more than 5 years' experience is available, in 60
percent of the years;
have had a cumulative adjusted loss ratio of about 2.25 or more
(i.e., $2.25 or more in claims payments for each $1 in
premiums); and
require a rate at least 10 percent higher than would have otherwise
been charged.
--------------------
\2 The actual loss ratio varies depending on the premium rate paid.
USDA calculates an adjusted loss ratio (identified as a "z" score) by
using an actuarial formula that considers both the loss ratio and the
premium rate charged. The higher the loss ratio and premium rate,
the higher the adjusted loss ratio and the more likely that a farmer
will be included in this high-risk program. For crop year 1995, USDA
reduced the adjusted loss ratio for some crops in order to maintain a
specified percentage of farmers in the high-risk program for each
crop.
\3 For this analysis, USDA adjusted the historical premiums to the
1993 premium rates, as the legislation required.
HIGH-RISK PROGRAM WILL PRODUCE
SAVINGS, ALTHOUGH NOT AS MUCH
AS ANTICIPATED
---------------------------------------------------------- Chapter 3:2
USDA's plan for targeting high-risk farmers will reduce the
government's outlays for the crop insurance program, although not by
as much as the Department had originally estimated. According to
USDA's blueprint, the high-risk program will reduce crop insurance
claims from an average of $1.40 in claims for every $1 in premiums to
an average of between $1.30 and $1.35. USDA estimated that the
program would result in savings of about $70 million for crop year
1993. However, we estimated savings from the program of about $33
million for 1993.
Our estimate is lower because we based it on the actual program USDA
implemented in 1992 and 1993, which did not include as many farmers
as the Department's estimate assumed. This estimate was based on
USDA's original plan to select 2 percent of the policyholders. In
practice, however, after changing its targeting criteria in 1992 and
1993, USDA selected only 1.5 percent. Furthermore, over one-third of
those identified had already ceased buying crop insurance before
being selected for the program. Therefore, about 1 percent of all
policyholders were included in the program. (App. IV contains a
more detailed discussion of our calculations and methodology.)
Additional savings may not be significant after the first year that
farmers are included in the high-risk program. Most of the savings
are realized in the first year, when many high-risk farmers choose to
stop purchasing crop insurance rather than pay the higher rates. For
those who remain, most of the rate increases occur in the first year;
the rate increases in succeeding years are similar to those for other
farmers. For example, for farmers who remained in the program after
being targeted in 1992, the premiums paid averaged 67 percent more in
1992 than in 1991 and 7 percent more in 1993 than in 1992.
CHANGES TO ESTABLISH ACCURATE
PRODUCTION LEVELS ARE UNDERMINED
BY LACK OF VERIFICATION OF
FARMERS' PRODUCTION HISTORY
============================================================ Chapter 4
To ensure that the crop insurance program realizes the congressional
requirement of receiving a projected $1 in premiums for each $1.10 in
claims paid, USDA needs reasonable estimates of farmers' normal
production. This information will help ensure that farmers do not
purchase insurance for production levels higher than they are likely
to produce and, as a result, make claims for production losses that
are not real. To achieve this objective, USDA has recently changed
the way it establishes farmers' production levels to more closely
align them with actual production history. USDA's action should
reduce the government's outlays, although not as much as USDA had
anticipated. However, this change has a critical weakness. USDA
does not require that loss adjusters verify the accuracy of the
production history supplied by the farmers and therefore lacks
assurance that it is insuring production at the appropriate level.
PLANNED ACTIONS TO DETERMINE
FARMERS' NORMAL PRODUCTION
LEVELS
---------------------------------------------------------- Chapter 4:1
According to USDA's blueprint, the level of production insured may be
the single most important factor in determining the success or
failure of the crop insurance program. The insured production level
is key because it forms the basis for calculating insurance premiums
and payments on claims. Consequently, a production level that is too
high compared with the productive potential of the farmer and the
land will increase the frequency and amount of a farmer's claim.
Conversely, a production level that is too low will not effectively
protect farmers from loss and, because the production level is
regarded as insufficient, will discourage farmers from buying
insurance.
Before crop year 1994, farmers could base the level of production for
which they purchased crop insurance on 10 years' actual production,
or, for those years for which farmers did not report actual
production, on a modified average production level for the county.
USDA concluded that the option of basing a production level on a
modified county average was adversely affecting the crop insurance
program's financial condition. This option benefited farmers whose
production was below the modified county average. It enabled them to
get a higher level of production coverage than their historic
production levels would have warranted. Therefore, some farmers may
have paid lower premiums than they should have and received claims
that exceeded what would have been warranted by their historic
production levels.
To address this problem, in crop year 1994 USDA began penalizing
farmers who did not have at least 3 years of production history. The
revised rules should discourage farmers from using the modified
average production level for the county and encourage them to provide
their actual production history. Under the revised rules, USDA uses
65 percent of the modified average production level for the county
for 4 years if the farmer reports no actual production,
80 percent of the modified average production level for the county
for 3 years if the farmer reports actual production for 1 year,
90 percent of the modified average production level for the county
for 2 years if the farmer reports actual production for 2 years,
and
100 percent of the modified average production level for the county
for 1 year if the farmer reports actual production for 3 years.
After the first 4 years, the level of production that can be insured
is the simple average of the actual production reported for up to 10
years.
NEW METHOD FOR SETTING
PRODUCTION LEVELS WILL PRODUCE
SAVINGS, ALTHOUGH NOT AS MUCH
AS ANTICIPATED
---------------------------------------------------------- Chapter 4:2
The actions USDA has taken to revise production levels will reduce
the government's outlays, but not by as much as it estimated. USDA
estimated in its blueprint that its actions would reduce crop
insurance claims over time from $1.40 for every $1 in premiums to
between $1.25 and $1.30. This estimate equates to a savings of
between $75 million and $113 million annually.
However, we estimated that these savings would be about $44 million
for crop year 1994. Our estimate differs from USDA's primarily
because USDA limited any reduction in a farmer's insured production
level to no more than 10 percent annually. In addition, with the
change in the calculation of production levels, farmers with 4 to 8
years of production history had increases in their production levels.
(App. V contains a detailed discussion of our methodology for
calculating the savings.)
USDA IS NOT REQUIRING LOSS
ADJUSTERS TO VERIFY PRODUCTION
HISTORY SUPPLIED BY FARMERS
---------------------------------------------------------- Chapter 4:3
Although USDA recognizes the importance of accurate production levels
to the program's integrity, it does not require that loss adjusters
verify the production history provided by farmers. Therefore, USDA
cannot be assured that it is paying claims accurately.
USDA allows farmers to certify the production level that they insure.
It also requires farmers to retain records supporting their certified
production level for 3 years. However, USDA does not require
insurance adjusters to verify the accuracy of the production levels
supplied by farmers.
Over the years, we and USDA's Office of Inspector General have
consistently found that USDA's process for verifying production
histories has been inadequate. In 1988, we reported that USDA did
not have adequate procedures to ensure that farmers' reported
production levels were accurate.\1 According to our analysis of
USDA's data, 37 percent of the production levels examined were
inaccurate, largely because of inaccurate certifications by farmers.
Therefore, we recommended that for each claim, USDA require loss
adjusters to verify the production data supporting the production
level insured. We noted that such verifications could be minimized
by spot-checking the supporting data for a farmer's production level
for some, rather than all, years. Likewise, in 1989 USDA's Inspector
General found inaccurately reported production levels in about half
of the cases reviewed and recommended that USDA require review of the
production levels for each claim until an acceptable error rate is
achieved.\2 Despite these recommendations, USDA has not established
an acceptable error rate and is not requiring verification.
--------------------
\1 Crop Insurance: FCIC Should Strengthen Actual Production History
Program Controls (GAO/RCED-89-19, Dec. 15, 1988).
\2 Federal Crop Insurance Corporation--Crop Year 1988 Insurance
Contracts With Claims, USDA Office of Inspector General, Audit Report
No. 05600-1-Te (Sept. 1989).
CONCLUSIONS
---------------------------------------------------------- Chapter 4:4
USDA's changes in the way it calculates farmers' production levels
should improve the program's financial condition because the revised
methodology will result in more accurate estimates of a farmer's
expected production. However, USDA will not get the full short-term
benefit it anticipated from the change because it limited the
reduction in a farmer's insured production level to no more than 10
percent annually. Therefore, farmers can continue for some time to
insure at production levels higher than their experience justifies.
Moreover, a long-standing problem that could erode the positive
benefit of more accurate production levels is the fact that USDA does
not require verification of production when claims are adjusted. We
believe this problem needs to be addressed.
RECOMMENDATIONS
---------------------------------------------------------- Chapter 4:5
We recommend that the Secretary of Agriculture direct the Deputy
Administrator for Risk Management to take the following actions:
Remove the 10-percent annual limit on reduction in farmers' insured
production levels so that the level of production insured is
aligned with the farmers' actual production history. If not,
USDA should include in an annual report to the Congress the
estimated cost of subsidizing the additional losses that will be
incurred.
Require that the production history provided by farmers be verified
when claims are adjusted.
AGENCY COMMENTS AND OUR
EVALUATION
---------------------------------------------------------- Chapter 4:6
In commenting on our recommendation concerning the 10-percent limit
in farmers' insured production levels, USDA recognizes that there is
a cost associated with its policy of limiting reductions in insured
yields. However, USDA believes that this policy provides a more
"gentle landing" for farmers than would occur in instances in which
farmers have recently suffered severe losses. USDA also agreed that
reporting on the impact of this policy on the estimated cost of
subsidizing additional losses in an annual report to the Congress is
workable.
Concerning our recommendation aimed at improving the verification of
production history, USDA agrees that it needs to look at ways to
better ensure that it is obtaining an adequate number of
verifications. However, USDA believes that it needs time to identify
the most appropriate point in the process for such verification.
USDA plans to consult with the companies with which it has insurance
contracts and arrive at a workable verification plan by May 31, 1996.
USDA HAS REDUCED THE RISK IN THE
TIMING OF INSURANCE SALES, BUT
SOME ADDITIONAL CHANGES ARE NEEDED
============================================================ Chapter 5
Until recently, USDA allowed farmers to purchase crop insurance after
they knew whether early growing conditions, such as the amount of
moisture in the subsoil, might result in poor production. Such late
deadlines for purchasing insurance increased the likelihood that
those who bought insurance during the planting period would file
claims. In recognition of the importance of purchasing deadlines to
the crop insurance program's financial soundness, 1994 crop insurance
legislation required USDA to set deadlines that were 30 days earlier
than in 1994.\1 This was to prevent farmers from buying crop
insurance close to or in the planting period, when they can better
evaluate the probability of a loss. While the revised deadlines will
reduce the extent of this problem, the underlying problem of USDA's
approach to setting these deadlines remains.
The legislation builds on the proposal USDA included in its blueprint
for setting the last date for purchasing insurance for the 1995 crop
year 15 to 30 days earlier than it had in 1994. USDA's proposal was
in response to our analysis of three crops in 111 crop-producing
areas.\2 For 33 percent of these areas, USDA allowed insurance sales
to continue well into the planting period (8 to 60 days past the
initial planting date). In another 32 percent of these areas, the
deadlines for purchasing insurance were near the start of the
planting period (up to 7 days before or after USDA's initial planting
date).
Although USDA moved the deadline for new purchases of crop insurance
30 days earlier in the year, it generally did not move the related
deadline for cancelling insurance. The cancellation deadline is the
last date that current insurance purchasers may cancel their coverage
before it continues in force for another year. Before crop year
1995, the purchasing and cancellation deadlines were on the same
date, as would be expected. By not moving the cancellation deadline,
USDA allows many current purchasers to make the decision to renew or
cancel their crop insurance coverage well into the planting period.
USDA officials explained that they could not change the cancellation
date without first publishing the proposed change for comment in the
Federal Register. USDA officials said that they were in the process
of making this change and expect to have revised all cancellation
dates by crop year 1997.
While USDA had set its purchasing deadlines 30 days earlier in the
year as the legislation required, it has not addressed the underlying
problem--these deadlines are not designed to address actual
production situations. Instead, the dates have historically been
set, and continue to be set, to ease the administration of the crop
insurance program. The dates are set for a several-state area rather
than for local growing conditions. Specifically, USDA has two
principal purchasing deadlines for spring-planted crops and two for
fall-planted crops. These deadlines have historically fallen into
the planting period for many crops in many areas. However, USDA does
not have written procedures or criteria for its field offices to
follow in reviewing and updating the purchasing deadlines on the
basis of the planting dates in each crop-growing region.
Consequently, the revised national deadlines correct the situation we
identified in most cases, but not all. With this change, 12 percent
of the areas we reviewed--compared with 66 percent formerly--had
purchasing deadlines that continued into the planting period.
Without establishing a procedure for routinely reviewing and updating
these deadlines on the basis of planting practices in each region,
USDA will continue to have some deadlines that extend into the
planting period.
Moreover, USDA does not record crop insurance sales dates in its
database. Therefore, it cannot evaluate the relationship between the
claims paid and the number of days before the planting period that
the insurance was purchased.
--------------------
\1 The Federal Crop Insurance Reform Act of 1994 (P.L. 103-354, Oct.
13, 1994, title I).
\2 For this analysis, we compared USDA's deadlines for purchasing
crop insurance and initial planting dates for corn, grain sorghum,
and soybeans in 1992. For example, in Nebraska USDA had one deadline
for insuring corn (April 15) but two planting areas with different
initial planting dates (April 6 for one part of the state and April
25 for the other).
CONCLUSIONS
---------------------------------------------------------- Chapter 5:1
USDA has improved the financial condition of the crop insurance
program by moving purchasing deadlines 30 days earlier in the year.
However, by not routinely setting these deadlines by crop-growing
regions, USDA enables some farmers to better evaluate growing
conditions and increases the likelihood that they will purchase crop
insurance when growing conditions are poor. As result, USDA
increases the probability of a shortfall--that claims paid will
exceed $1.10 for each $1 in premiums. Furthermore, by not recording
purchase dates in its database, USDA cannot adequately evaluate the
relationship between the claims paid and the number of days before
the planting period that insurance was purchased.
RECOMMENDATIONS
---------------------------------------------------------- Chapter 5:2
We recommend that the Secretary of Agriculture direct the Deputy
Administrator for Risk Management to
set purchasing deadlines before the initial planting date in all
areas of the country and establish criteria and procedures for
routinely reviewing these deadlines to ensure that they continue
to occur before initial planting dates and
record the date that insurance is purchased in order to better
evaluate the relationship between purchasing deadlines and
claims payments.
AGENCY COMMENTS AND OUR
EVALUATION
---------------------------------------------------------- Chapter 5:3
USDA agrees with our recommendation that USDA set purchasing
deadlines before the initial planting date in all areas of the
country and establish criteria and procedures for routinely reviewing
these deadlines to ensure that they continue to occur before the
initial planting dates. USDA noted that the legislative requirement
to move all purchase deadlines 30 days earlier for crop year 1995
resulted in some purchase deadlines being too early and inconsistent.
USDA'S RATE-SETTING METHODOLOGY
=========================================================== Appendix I
Annually, analysts in the U.S. Department of Agriculture's (USDA)
crop insurance field office in Kansas City, Missouri, follow a
multistep process, which is partially automated, to establish the
premium rates for each of the 51 crops in the federal crop insurance
program.
For each crop, USDA analysts begin by extracting data on counties'
crop experience for all years available (up to 20) from the mainframe
computer. The data elements for each crop, crop year, and county
include (1) the dollar amount of the insurance coverage sold (also
referred to as insurance in force),\1 (2) the dollar amount of the
claims paid,\2 and (3) the average coverage level.
The analysts remove from the resulting database information on
historical insurance in force and claims payments to farmers who
incur frequent and severe losses relative to other farmers. The
premium rates for these individuals are established separately under
the high-risk program. The analysts remove these data to avoid
setting rates that are higher than necessary for the risk represented
by the farmers who are not considered high-risk.
The analysts then adjust the historical data to the 65-percent
coverage level, the level at which USDA sets its basic premium rate.
Using the adjusted data, the analysts compute a loss-cost ratio for
each crop in each county. They calculate this ratio by dividing the
total claims payments by the total insurance in force; the result is
stated as a percentage. For example, if the claims paid in one year
totaled $7.36 and the insurance in force was $100, the loss-cost
ratio is 7.36 percent. The percentage is the rate that USDA would
need to charge per $100 of insurance coverage sold if the total
premiums are to equal the total claims payments over a number of
years. For example, 7.36 percent would indicate that a rate of $7.36
is required per $100 insurance coverage sold. In calculating
loss-cost ratios, USDA uses the latest data available, which are for
the period ending 2 years before the year for which the rates are
being established. For example, USDA established crop year 1995
rates in 1994; the most recent 20-year period was for crop years 1974
through 1993.
The analysts determine a preliminary rate--called the county unloaded
rate--by first calculating a loss-cost ratio for each of the 20
years. The analysts then divide the data into two segments--the 4
years with the highest loss-cost ratios and the 16 years with the
lowest loss-cost ratios. For the 4 years with the highest loss-cost
ratios, they cap the ratios at the loss-cost ratio for the
highest-loss year in the 16-year segment. Using the capped loss-cost
ratios for each of the 4 years, the analysts calculate the average
for all 20 years to establish the county unloaded rate.
The analysts adjust the county unloaded rates to minimize the
difference between the rate for each county and that of its
neighboring counties. To make this adjustment and "smooth" the rates
from county to county, the analysts use a weighting process, called
the concentric circle method.
USDA then develops a surcharge for catastrophic coverage for each
crop in each state. This surcharge is added to the adjusted unloaded
rate for each county in the state. The analysts pool, at the state
level, the amount of insurance in force and the claims payments for
the 4 years with the highest loss-cost ratios in each county that was
not factored into the county unloaded rates. Using these data, they
compute a statewide surcharge for catastrophic coverage (pooled
claims payments divided by pooled insurance in force). USDA limits
the surcharge to a maximum of 5 percentage points, or $5 per $100 in
insurance coverage. Beginning in 1995, any excess above the 5
percent is factored back into the individual county's unloaded rate
on the basis of the amount that the county contributed to the excess.
The analysts then add the state catastrophic surcharge to each
county's unloaded rate. This results in a basic rate for the county
for the 65-percent coverage level and is also considered to be for
the average production level in that county.
At this point, the analysts calculate a rate at the 65-percent
coverage level for each farming practice, such as whether the insured
acreage is irrigated or dryland, and for each crop type, such as
winter wheat or spring wheat. Using historical data, the analysts
periodically establish a factor for adjusting the basic rate for each
crop in each county to align the rate with the risk associated with
each farming practice and crop type. The analysts apply these
factors to the basic rate to calculate a rate for each farming
practice and crop type.
To facilitate review by regional service office underwriters, the
analysts convert the rate from the 65-percent coverage level to the
75-percent coverage level. This conversion is made because the
underwriters are more familiar with the 75-percent coverage level.
The analysts make the conversion by multiplying the rate for each
county-crop practice and type by a previously established factor of
154 percent.\3 The field underwriters review these rates for
reasonableness on the basis of their knowledge and continuing
research on farmers' experiences with the particular crop in the
county and the surrounding area; they recommend changes when they
believe adjustments are warranted. On the basis of the
recommendations, analysts in the Kansas City office make any
adjustments.
Following this review and any resulting adjustments, the analysts
adjust the rates upward for the risk represented when farmers choose
to subdivide their farming operation for a given crop into multiple
units for crop insurance purposes. They do this because USDA's
historical data show that farming operations insured on a multiple
unit basis are more likely to make claims than those insured as one
unit. The analysts divide the basic rate by a factor of 0.9 (in
effect, imposing a 10-percent surcharge) to arrive at the rate
charged farmers who subdivide their land into units. If a farmer
chooses to insure the entire farming operation for a crop as one
unit, USDA then allows a 10-percent discount from the rate.
The rates for the 50-, 65- and 75-percent coverage levels that the
analysts have calculated are for farmers whose historic production
level, or yield, is about equal to the average for all producers in
the county. However, many farmers' average production level is above
or below the county's average. According to USDA, farmers' chances
of having a loss decrease as production increases. Therefore, USDA
has established a series of rates (9 for barley, corn, grain sorghum,
soybeans, and wheat, and 13 for cotton) that decrease as production
levels increase. USDA refers to these as rate spans; for most crops,
it numbers them R01 through R09. To adjust the rate for the average
production level--referred to as the midpoint rate (generally
R05)--to the other production levels, the analysts apply a
mathematical model to the R05 rate. The R06 through R09 spans are
for producers whose production levels are higher than average. The
model reduces the rates as production levels increase from the R06 to
the R09 level. Conversely, the R01 through R04 spans are for farmers
whose production is lower than average. The model increases the
rates as the production level decreases, and the R01 span is charged
the highest rate.
Figure I.1 shows the rates for corn in one county to illustrate the
nonlinear relationship of the rates by production level. At the
lowest production level for the county--less than 45 bushels per
acre--the premium rate is $17.40 per $100 coverage; at the midpoint
level--83 to 95 bushels per acre--the rate is $7.10; and at the
highest level--more than 133 bushels per acre--the rate is $4.60.
Figure I.1: Crop Insurance
Rates for Production Level for
Corn in One County, Crop Year
1995
(See figure in printed
edition.)
Note: Rates are at the 65-percent coverage level.
Source: GAO's illustration is based on USDA's 1995 rates for dryland
corn in Cooper County, Missouri.
To account for the government's subsidy, the analysts further adjust
the rate table so that the rates exclude this subsidy. The analysts
then examine the completed rates to ensure that none exceed the
maximum increase of 20 percent per year allowed by law and make
adjustments downward, where necessary.
As a final step, discounts are developed for farmers who buy hail and
fire protection from private insurance companies. The analysts
establish a maximum discount for each crop in each county on the
basis of the prevailing costs of private insurance in the area.
Farmers who buy hail and fire protection from another insurer are
granted a discount from the USDA rate that is equal to the amount
they paid for the protection up to a preestablished maximum set by
USDA.
--------------------
\1 USDA refers to the insurance coverage sold as liabilities.
\2 USDA refers to the claims paid as indemnities.
\3 Later, the analysts adjust the rates back to the 65-percent
coverage level. To calculate the rate for the 50-percent level, the
analysts multiply the 65-percent coverage level by 72 percent.
METHODOLOGY FOR EVALUATING EXTENT
TO WHICH USDA SET CROP INSURANCE
PREMIUM RATES AT REQUIRED LEVELS
========================================================== Appendix II
To estimate the extent to which USDA's premium rates for crop
insurance were adequate to achieve an income of at least $1 in
premiums for every $1.10 in claims paid to farmers, we examined the
extent to which USDA's basic rates for crop years 1991 through 1995
achieved this requirement (referred to in this report as "adequacy"),
measured the extent to which USDA had raised the premium rates each
year since 1990, and examined the accuracy of the rates for different
coverage and production levels. We based our analysis on six major
crops that represent about 75 percent of USDA's crop insurance
business: corn, wheat, soybeans, cotton, grain sorghum, and barley.
METHODOLOGY FOR CALCULATING
EXTENT TO WHICH USDA'S RATES
WERE ADEQUATE
-------------------------------------------------------- Appendix II:1
To determine the adequacy of USDA's premium rates, we compared USDA's
basic rates (set at the 65-percent coverage level and average, or
midpoint, production level) for each crop and each year to a
benchmark premium rate that we calculated. Generally, we followed
USDA's methodology for setting the basic rates in developing our
benchmark rates. Just as USDA does, we used the insurance experience
for the most recent 20-year period available--at the time the rates
were established--to calculate the benchmark rates for each year.
Although USDA sets premium rates for each crop at the county level,
it uses the historical experience of both the county and state to do
so. Therefore, to develop a benchmark rate that also considered
county and state experience, we weighted experience and rates at the
county crop level up to the state crop, national crop, and national
level (all crops combined) for the six crops reviewed. We also
calculated a benchmark rate, as suggested by a USDA consultant, that
gives more weight to recent years' experience.
Once we calculated a benchmark rate at the state crop, national crop,
and national level for each year, we calculated the extent to which
USDA's rates were adequate (as a percentage), by dividing USDA's
weighted average basic rate by our benchmark rate.
BENCHMARK RATES
------------------------------------------------------ Appendix II:1.1
To calculate benchmark rates for each crop for 1991-95, we obtained
the computer files that USDA used to set insurance rates each year.
These data--for crop years 1969-93 for each of the six crops we
reviewed--included the dollar amount of premium income, claims paid,
and dollar amount of insurance coverage sold (insurance in force).
The data also included the average coverage level for each crop in
each county (county crop program). We then made two adjustments to
the data that USDA also makes. We (1) removed from the database the
experiences of the high-risk producers, which USDA handles under a
separate program, and (2) used USDA's mathematical formula to adjust
the claims paid and insurance in force to a uniform 65-percent
coverage level. In addition, we removed records that had zero or
missing values in all fields or a negative value in any field.
Following these adjustments to the database, we calculated benchmark
rates for each crop in each county by taking a 20-year\1
simple average of the loss-cost ratios. The 20-year period used was
the period for which USDA had complete data available when it set its
premium rates each year. For example, the 1995 benchmark rates were
based on the 1974-93 period. To obtain total benchmark premiums for
a county, we multiplied the benchmark rate for that county and crop
by the adjusted amount of insurance in force for that county and crop
for the last year in the 20-year period.
We then summed the benchmark premiums and adjusted the amount of
insurance in force by county crop program for each state and crop.
To calculate an estimated benchmark premium rate by state crop
program and crop year, we divided the total adjusted amount of
insurance in force into the total benchmark premiums. We also summed
the data to the national level and made an identical calculation to
determine a benchmark premium rate for the crop for the nation and a
national benchmark premium rate for the six crops combined.
--------------------
\1 For counties with less than 20 years of experience, we used the
number of years for which experience was documented.
WEIGHTED AVERAGE BASIC RATES
------------------------------------------------------ Appendix II:1.2
Recognizing that varying amounts of insurance coverage are sold in
the many counties where insurance is offered and that the premium
rates vary by county, we calculated a weighted average basic rate for
each crop to allow each basic rate to have a weight corresponding to
the amount of insurance sold at that rate. To make this calculation
for each crop for each year, we obtained the amount of insurance in
force and the published premium rates from USDA's database by county,
farming practice, and crop type. Because USDA's published rates
include a 10-percent charge for optional unit coverage, we adjusted
the rates downward to eliminate the charge included for this
coverage. We made this adjustment because USDA deducts the amount
from the rate if an insured farmer does not elect this option.
To calculate USDA's weighted average basic rate at the state crop and
national crop level, we made a series of calculations. First, for
each county crop program and year, we multiplied the basic rates (at
65-percent coverage and midpoint of the production levels) for each
farming practice and crop type by the amount of insurance in force
for the same year and for each farming practice and crop type.\2
Second, we summed the premiums and the amount of insurance in force
for each county crop program and each year and divided the total
premiums by the insurance coverage to obtain a weighted average basic
rate for the county and crop.\3 Third, we multiplied the weighted
average rate for that county and crop for each year by the adjusted
amount of insurance in force that we used to calculate the benchmark
premium rates--the same information that USDA had available when it
set the premium rates--to obtain the total premiums in the county
that USDA could expect when it published its rate schedule. Fourth,
we summed the total premiums in the county and total amount of
insurance in force for each county crop program and year to the state
crop and national crop level. Fifth, we divided the total premiums
by the total insurance coverage in force for each crop and year to
obtain a weighted average state crop rate, national crop rate, and
national premium rate.
--------------------
\2 Although USDA sets rates annually with data on experience ending 2
years before the year for which the rates are set, we applied the
rates to the same year's insurance in force because USDA changes its
coding for some information from year to year. This makes it very
difficult to match the rates for one year with the insurance in force
in any previous years. For the six crops in 1995, USDA published
about 25,300 rates at the 65-percent coverage and midpoint of the
production levels.
\3 We multiplied 1995 rates by the 1994 insurance coverage in force
because 1995 data was not complete at the time of our review. USDA
representatives worked with us to identify the several coding changes
necessary to match 1994 and 1995 data for this calculation.
USDA CONSULTANT'S SUGGESTED
WEIGHTING OF EXPERIENCE
------------------------------------------------------ Appendix II:1.3
We calculated an additional benchmark rate, following the same
methodology used for the original benchmark rate, except we gave
greater weight to experience in recent years. This approach was
suggested by USDA's actuarial consulting firm, Milliman and
Robertson, in a 1983 study. Following the consulting firm's
suggestion, we assigned 50-percent weight to the data for the last 10
years and 50-percent weight to the data for the entire 20-year
period.\4 In addition to the benchmark rate we had already computed,
we computed a second rate averaging the loss-cost ratios for the most
recent 10 years. We then averaged the two rates--that is, for 20
years and 10 years--to obtain a benchmark rate suggested by USDA's
consultant.
--------------------
\4 Although the consultant suggested that all years of historical
data be used, following USDA's approach we used 20 years of
historical data and did not determine the impact that using all
available years of history would have had on USDA's premium rates.
CALCULATING EXTENT TO WHICH
USDA INCREASED ITS PREMIUM
RATES
-------------------------------------------------------- Appendix II:2
To determine the extent to which USDA raised the premium rates each
year from 1991 to 1995, we calculated the average premium rates by
state crop program for each year and then calculated the extent to
which USDA raised or lowered these rates from one year to the next
for the six crops reviewed. We calculated a weighted average premium
rate to give each rate a weight corresponding to the amount of
insurance sold at that rate. To do so, we extracted from USDA's
database information on the total amount of insurance coverage in
force by county crop program, farming practice, and crop type for
crop years 1990 through 1994. We multiplied this amount by the
premium rate for each farming practice and crop type to obtain the
total premiums.\5 We summed these premiums and insurance in force
amounts to the state crop program level and divided the total
premiums by the total insurance in force. This provided a weighted
average premium rate by state crop program and crop year.
--------------------
\5 We multiplied the 1995 premium rates by the 1994 insurance in
force.
CALCULATING EXTENT TO WHICH
RATES FOR EACH COVERAGE AND
PRODUCTION LEVEL WERE CORRECT
-------------------------------------------------------- Appendix II:3
To determine the adequacy of USDA's premium rates for farmers
purchasing insurance at coverage and production levels different from
the basic rate (65-percent coverage and the midpoint of the
production levels), we compared the relative loss experience of the
different levels available over the period 1990 through 1994 to the
loss experience at the basic rate. USDA sets the basic rate for each
crop in each county and multiplies the basic rate by predetermined
mathematical factors to arrive at the rate for 75- and 50-percent
coverage levels and to arrive at rates for the several different
production levels above and below the midpoint rate span. If the
factors are correct, then the average loss ratios for each coverage
and production level over a number of years should be about the same
as that experienced at the basic rates, and the premiums will be
accurate. Conversely, if the factors are incorrect, then the average
loss ratios for different coverage and production levels will vary,
and the premium rates will be incorrect.
COVERAGE LEVELS
------------------------------------------------------ Appendix II:3.1
To measure whether rates for coverage levels above and below the
65-percent coverage level were correct for each crop, we recalculated
the premiums and claims paid at the 75-percent coverage level as if
the insurance had been sold at the 65-percent level and recalculated
the premiums and claims paid at the 65-percent coverage level as if
the insurance had been sold at the 50-percent level. In addition,
for each crop, we (1) calculated the loss ratios for each year for
both the original coverage level and the recalculated coverage level,
(2) averaged the loss ratios for the 5 years for both the original
coverage level and the recalculated coverage level, (3) calculated
the percentage that the loss ratios at 75-percent coverage were more
or less than at the recalculated 65-coverage, and (4) calculated the
percentage that the loss ratios for the recalculated 50-percent
coverage were more or less than at 65-percent coverage.
To calculate the amount by which the premiums were more or less than
required, we multiplied the 1994 premiums for each crop and coverage
level by the percentage that the average loss ratio (over the 5-year
period) was higher or lower than the loss ratio for the 65-percent
coverage level.
PRODUCTION LEVELS
------------------------------------------------------ Appendix II:3.2
To measure whether the premium rates for production levels above and
below the basic rates were correct for each crop, we (1) calculated
the loss ratio for the basic rates, which are set at the county's
average production, and the rates for production higher and lower
than the basic rates for 1990 through 1994; (2) averaged the loss
ratios for each of the three groups for the 5-year period; and (3)
calculated the percentage by which the loss ratio for the production
levels higher and lower than the basic rates differed from the loss
ratios for the basic rates.
To calculate the amount by which premiums were more or less than
required for these production levels, we multiplied the 1994 premiums
for each crop by the percent that the average loss ratio (over the
5-year period) was higher or lower than the loss ratio for the basic
rates.
RESULTS OF ANALYSIS OF CROP
INSURANCE PREMIUM RATES
========================================================= Appendix III
Tables III.1 through III.6 present the results of our analysis of
USDA's basic rates (i.e., at the 65-percent coverage level and
average production level) for the six crops we reviewed. For each
crop, state, and year (1991 through 1995), the table shows the
weighted average premium rate (arrived at by weighting the county
premium rates by the insurance coverage sold);
extent to which the premium rate was adequate, expressed as a
percentage; and
amount of insurance coverage sold (insurance in force).\1
Table III.1
Weighted Average Premium Rate, Adequacy
of Premium Rate, and Insurance in Force
by State for Barley, Crop Years 1991-95
Adequacy
of
Premium Insurance Premium Insurance Premium Adequacy Insurance Premium Adequacy Insurance Premium premium
rate per Adequacy of in force rate per Adequacy of in force ($ rate per of premium in force rate per of premium in force rate per rate
$100 premium rate ($ in $100 premium rate in $100 rate ($ in $100 rate ($ in $100 (percent
State coverage (percent) millions) coverage (percent) millions) coverage (percent) millions) coverage (percent) millions) coverage )
------------ ---------- ------------ ---------- ---------- ------------ ----------- ---------- ---------- ---------- --------- ---------- ---------- -------- --------
Ariz. $3.51 58 $0.02 \a \a $0 \a \a $0 \a \a $0 $4.14 115
Calif. 12.89 39 1.90 $16.08 50 1.07 $16.93 57 0.76 $16.28 66 0.51 17.93 80
Colo. 7.30 90 0.94 8.15 91 1.05 7.54 109 0.78 7.64 104 0.89 7.81 131
Del. 2.97 339 0.03 3.42 380 0.02 3.24 616 0.01 3.06 367 0.05 3.24 248
Idaho 4.73 48 6.37 5.14 50 5.44 4.77 46 4.87 5.01 48 5.15 5.60 52
Ill. 5.31 31 0.01 5.56 73 0.01 5.99 24 0.01 7.20 33 0\b 8.82 51
Ind. 3.91 57 0.02 4.41 88 0.02 4.87 36 0.02 5.67 36 0.01 6.93 54
Iowa 6.75 952 0.01 7.64 1,594 0.01 7.39 206 0.01 7.49 152 0.02 8.12 64
Kans. 10.20 37 0.32 11.78 42 0.26 11.05 41 0.12 11.57 48 0.19 13.21 54
Ky. 4.68 21 0.06 5.49 27 0.03 5.45 33 0.02 6.57 50 0.02 8.01 73
Md. 3.51 \a 0\b \a \a 0 \a \a 0 \a \a 0 \a \a
Mich. 5.87 37 0.15 7.09 61 0.12 6.82 55 0.15 7.77 65 0.15 8.60 74
Minn. 8.16 89 25.76 9.27 110 16.90 8.81 102 16.41 9.65 113 16.03 10.04 106
Mo. 7.39 152 0.03 8.60 157 0.03 9.51 67 0.02 11.38 82 0.02 12.78 189
Mont. 8.83 63 54.78 9.53 68 40.93 9.32 69 33.51 10.45 73 31.19 11.32 79
Nebr. 10.99 47 0.15 13.05 57 0.10 11.64 61 0.07 15.00 82 0.09 16.53 75
Nev. \a \a 0 9.27 10 0.01 7.42 20 0.14 \a \a 0 \a \a
N. Mex. 15.86 38 0.03 19.55 46 0.02 19.90 50 0.02 20.73 65 0.01 22.79 78
N.Y. 3.87 \a 0\b 4.41 48 0\b \a \a 0 5.22 16 0\b 5.49 16
N.C. 4.60 25 0.02 5.24 75 0\b 5.51 51 0.01 6.47 69 0.01 7.49 125
N. Dak. 8.26 84 122.39 9.16 95 85.31 8.29 87 83.09 8.30 90 73.12 8.91 88
Ohio \a \a 0 4.14 \a 0\b \a \a 0 \a \a 0 \a \a
Okla. 10.00 27 0.05 10.82 91 0.06 11.00 71 0.08 12.76 44 0.09 14.49 52
Oreg. 3.51 93 2.73 3.76 98 1.82 3.98 86 1.17 4.01 100 0.87 4.16 98
Pa. 3.43 284 0.01 3.93 320 0.01 3.92 318 0.01 4.04 95 0.03 4.22 155
S.C. 5.58 31 0.02 6.50 49 0.01 7.11 26 0.01 7.85 29 0.02 9.65 50
S. Dak. 11.93 71 6.96 13.12 75 5.83 12.82 76 4.01 12.71 77 4.13 13.40 81
Tenn. 6.21 49 0.05 7.20 53 0.01 7.84 142 0.01 9.35 61 0.01 11.12 84
Tex. 12.14 34 0.22 13.87 54 0.13 14.12 42 0.09 15.89 58 0.06 17.22 65
Utah 5.41 32 0.61 5.57 27 0.58 6.87 30 0.35 7.83 34 0.15 8.70 39
Vt. \a \a 0 \a \a 0 \a \a 0 5.40 12 0\b 6.48 16
Va. 4.43 47 0.49 5.18 66 0.33 5.08 59 0.32 5.55 55 0.28 6.37 68
Wash. 5.00 112 8.68 4.93 112 5.29 5.04 95 4.03 4.98 88 4.40 5.00 92
Wis. 8.97 62 0.26 10.41 72 0.18 10.28 65 0.26 10.56 77 0.66 11.50 72
Wyo. 3.66 67 3.17 4.44 90 4.00 4.37 105 3.37 4.36 104 3.10 4.36 115
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Notes: For 1995, no figures are shown for insurance in force because
the year was not complete at the time of our review.
Premium rates are based on the weighted average of rates at the
65-percent coverage and average production level.
\a Data were insufficient to perform analysis.
\b Although some insurance was in force, the amount was too small to
show in this presentation.
Source: GAO's analysis of USDA's data.
Table III.2
Weighted Average Premium Rate, Adequacy
of Premium Rate, and Insurance in Force
by State for Corn, Crop Years 1991-95
Adequacy
of
Premium Insurance Premium Insurance Premium Adequacy Insurance Premium Adequacy Insurance Premium premium
rate per Adequacy of in force rate per Adequacy of in force ($ rate per of premium in force rate per of premium in force rate per rate
$100 premium rate ($ in $100 premium rate in $100 rate ($ in $100 rate ($ in $100 (percent
State coverage (percent) millions) coverage (percent) millions) coverage (percent) millions) coverage (percent) millions) coverage )
------------ ---------- ------------ ---------- ---------- ------------ ----------- ---------- ---------- ---------- --------- ---------- ---------- -------- --------
Ala. $12.01 81 $3.29 $12.36 84 $4.53 $12.38 85 $3.26 $13.95 $92 3.93 $14.45 88
Ark. 7.71 29 0.48 8.76 30 1.25 9.65 26 0.97 11.68 40 1.05 13.95 39
Calif. 4.28 434 0.13 4.27 139 0.49 4.35 148 0.57 4.08 350 0.76 3.92 428
Colo. 5.04 149 41.08 5.19 150 42.20 5.06 146 42.61 5.00 138 59.56 5.01 120
Conn. 6.21 160 0.04 7.20 289 0.01 7.29 134 0.03 7.47 151 0.03 7.47 133
Del. 5.81 80 2.79 5.89 83 2.95 5.13 84 2.52 5.42 86 3.35 5.75 74
Fla. 8.47 42 1.02 10.53 53 1.15 10.62 60 0.91 12.63 77 1.02 13.62 76
Ga. 9.57 59 7.33 11.16 63 10.85 12.04 73 7.76 13.14 84 8.50 14.06 84
Idaho 3.55 33 0.16 4.20 26 0.07 3.99 32 0.25 3.81 34 0.12 4.42 26
Ill. 3.08 67 541.69 3.27 78 677.37 3.31 72 635.11 3.79 81 736.01 3.86 82
Ind. 3.53 81 176.91 3.64 87 240.25 3.67 75 223.11 4.04 84 260.86 4.15 87
Iowa 3.54 82 1,032.45 3.50 86 1,085.29 3.52 83 999.58 3.79 90 1,295.15 4.00 78
Kans. 4.54 76 80.91 4.60 77 83.55 4.52 74 89.09 5.25 78 109.43 5.63 81
Ky. 7.33 55 23.32 8.49 70 23.85 8.49 73 20.89 9.33 85 23.57 9.34 90
La. 5.16 41 10.34 6.10 56 11.63 6.51 43 7.53 7.53 49 13.27 8.99 58
Maine 6.21 \a 0\b 7.20 310 0\b 7.29 470 0\b 7.02 604 0.02 7.02 755
Md. 6.06 85 9.86 7.16 106 11.96 7.18 94 10.13 7.58 95 12.99 7.93 89
Mass. 5.49 97 0.03 6.57 139 0.02 7.20 127 0.08 7.63 143 0.05 7.60 128
Mich. 7.92 84 21.61 8.22 90 22.30 8.64 91 44.54 9.14 81 34.18 9.55 71
Minn. 5.70 102 376.62 5.45 102 381.71 5.49 101 371.08 5.54 99 630.46 6.14 76
Miss. 7.56 32 2.14 8.90 35 3.13 9.21 34 1.31 10.99 37 3.27 13.05 50
Mo. 8.57 73 66.92 9.15 84 72.82 9.00 80 61.29 9.70 86 96.32 9.97 83
Mont. 6.83 47 1.38 6.80 54 1.15 6.70 51 1.09 7.17 58 1.19 8.00 63
Nebr. 4.30 104 480.26 4.18 103 487.24 4.15 105 484.13 4.29 99 614.23 4.38 94
N.J. 6.97 62 0.21 7.99 84 0.24 8.15 87 0.20 8.58 87 0.24 8.77 75
N. Mex. 9.81 158 1.18 10.73 197 1.78 10.52 215 1.36 10.17 124 1.84 9.90 125
N.Y. 7.16 58 2.09 8.42 82 1.29 8.08 69 7.66 8.76 57 4.66 9.25 69
N.C. 8.63 68 22.84 9.57 78 25.01 9.35 87 20.96 10.31 95 26.28 10.97 89
N. Dak. 10.87 64 27.12 12.19 78 25.94 11.02 77 18.15 11.44 70 25.01 12.51 67
Ohio 4.01 89 82.02 4.12 95 104.58 4.08 84 94.51 4.37 92 106.24 4.42 94
Okla. 6.23 74 3.32 6.20 65 3.41 6.72 77 3.65 7.07 90 5.49 7.51 86
Oreg. 3.27 40 0.49 3.76 51 0.18 3.31 32 0.39 3.33 25 0.21 3.94 33
Pa. 7.02 68 12.67 7.26 68 17.97 7.47 54 18.93 8.36 61 20.76 9.05 62
R.I. \a \a 0 \a \a 0 \a \a 0 \a \a 0 \a \a
S.C. 11.32 62 5.06 13.32 67 6.75 13.59 74 6.00 15.17 88 7.96 16.44 83
S. Dak. 7.34 86 118.91 7.38 94 122.44 7.55 92 111.62 7.69 90 209.34 8.23 82
Tenn. 9.28 75 3.53 10.32 75 5.44 10.14 80 4.34 11.18 90 6.35 11.46 97
Tex. 6.89 63 55.89 7.69 54 48.96 8.36 51 56.06 9.02 82 69.60 8.76 84
Utah 4.53 34 0.27 5.28 26 0.27 5.64 10 0.09 5.03 23 0.18 5.86 43
Vt. 6.35 216 0.13 7.49 366 0.13 7.07 340 0.13 7.02 154 0.11 6.97 124
Va. 6.94 42 18.29 8.12 53 16.95 8.52 56 15.27 9.56 65 18.88 10.44 65
Wash. \a \a 0 \a \a 0 \a \a 0 \a \a 0 \a \a
W. Va. 7.34 37 1.54 8.62 53 1.92 9.21 44 1.68 10.67 55 1.86 11.90 58
Wis. 7.38 93 57.10 7.34 96 63.48 7.87 95 111.98 8.62 88 160.96 9.14 79
Wyo. 7.37 71 1.39 7.13 92 1.89 7.04 112 1.86 7.09 96 3.17 7.44 72
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Notes: For 1995, no figures are shown for insurance in force because
the year was not complete at the time of our review.
Premium rates are based on the weighted average of rates at the
65-percent coverage and average production level.
\a Data were insufficient to perform analysis.
\b Although some insurance was in force, the amount was too small to
show in this presentation.
Source: GAO's analysis of USDA's data.
Table III.3
Weighted Average Premium Rate, Adequacy
of Premium Rate, and Insurance in Force
by State for Cotton, Crop Years 1991-95
Adequacy
of
Premium Insurance Premium Insurance Premium Adequacy Insurance Premium Adequacy Insurance Premium premium
rate per Adequacy of in force rate per Adequacy of in force ($ rate per of premium in force rate per of premium in force rate per rate
$100 premium rate ($ in $100 premium rate in $100 rate ($ in $100 rate ($ in $100 (percent
State coverage (percent) millions) coverage (percent) millions) coverage (percent) millions) coverage (percent) millions) coverage )
------------ ---------- ------------ ---------- ---------- ------------ ----------- ---------- ---------- ---------- --------- ---------- ---------- -------- --------
Ala. $8.50 104 $40.26 $8.96 103 $42.76 $8.93 96 $44.10 $9.27 94 $56.31 $9.62 94
Ariz. 3.70 251 11.01 3.77 268 13.34 3.54 284 55.95 3.79 168 38.42 3.87 143
Ark. 9.45 70 16.47 10.57 71 12.95 10.65 74 9.26 11.49 57 10.13 12.93 61
Calif. 3.59 107 8.96 3.58 113 5.35 3.76 132 14.72 3.89 126 9.75 3.81 112
Fla. 11.33 70 3.61 13.25 113 4.77 12.99 88 4.33 13.63 87 5.99 13.72 88
Ga. 13.69 93 27.55 16.13 105 30.28 15.97 110 35.47 13.55 96 47.91 14.16 92
Kans. 14.57 19 0.06 14.62 33 0.10 15.93 35 0.02 19.39 41 0.01 19.46 33
La. 11.49 67 49.21 12.29 70 38.38 12.95 69 32.07 14.25 71 32.38 15.20 74
Miss. 6.77 69 33.52 7.75 71 29.64 8.00 70 28.17 8.43 78 37.98 9.10 72
Mo. 13.22 121 0.33 12.15 127 0.55 12.18 126 0.51 11.60 114 1.27 11.65 113
N. Mex. 11.32 71 3.43 12.89 80 3.06 14.17 86 3.23 15.57 87 3.35 17.20 98
N.C. 11.69 152 7.91 11.87 104 8.07 11.35 90 10.10 11.68 100 14.68 11.50 109
Okla. 9.84 79 16.12 9.39 82 17.84 10.60 80 20.97 12.75 86 21.42 14.06 92
S.C. 15.93 93 6.30 16.97 102 5.35 16.40 105 5.88 15.59 101 6.37 15.26 98
Tenn. 9.10 80 5.98 10.65 89 6.63 9.94 76 7.05 10.34 74 12.08 11.45 81
Tex. 13.80 84 569.60 15.13 92 515.97 15.70 96 583.23 18.26 100 606.88 19.50 106
Va. 14.72 133 0.27 13.93 221 0.22 12.32 710 0.56 12.24 232 0.98 11.61 98
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Notes: For 1995, no figures are shown for insurance in force because
the year was not complete at the time of our review.
Premium rates are based on the weighted average of rates at the
65-percent coverage and average production level.
Source: GAO's analysis of USDA's data.
Table III.4
Weighted Average Premium Rate, Adequacy
of Premium Rate, and Insurance in Force
by State for Grain Sorghum, Crop Years
1991-95
Adequacy
of
Premium Insurance Premium Insurance Premium Adequacy Insurance Premium Adequacy Insurance Premium premium
rate per Adequacy of in force rate per Adequacy of in force ($ rate per of premium in force rate per of premium in force rate per rate
$100 premium rate ($ in $100 premium rate in $100 rate ($ in $100 rate ($ in $100 (percent
State coverage (percent) millions) coverage (percent) millions) coverage (percent) millions) coverage (percent) millions) coverage )
------------ ---------- ------------ ---------- ---------- ------------ ----------- ---------- ---------- ---------- --------- ---------- ---------- -------- --------
Ala. $5.96 20 $0.10 $6.82 26 $0.06 $7.60 15 $0.03 $9.66 35 $0.03 $11.51 38
Ark. 5.83 32 3.32 6.71 36 6.09 7.33 31 2.22 9.14 40 3.00 10.72 40
Colo. 11.80 44 1.55 12.95 56 1.15 13.36 57 1.24 16.47 75 1.32 16.95 75
Del. 4.95 15 0.06 5.85 23 0.03 6.21 31 0.04 5.98 35 0.05 6.49 42
Fla. 8.75 40 0.02 10.62 29 0.02 11.10 34 0.01 13.83 45 0.01 16.07 50
Ga. 7.90 22 0.12 9.13 31 0.11 9.51 24 0.07 10.63 31 0.12 12.76 32
Ill. 6.19 42 2.32 6.98 46 3.11 7.43 52 2.32 8.70 62 2.17 9.47 68
Ind. 5.24 44 0.07 5.91 83 0.14 5.81 76 0.12 6.60 88 0.12 7.38 132
Iowa 6.40 61 0.07 7.57 42 0.12 7.50 68 0.06 8.60 97 0.06 10.26 58
Kans. 7.36 79 75.34 8.34 98 80.53 8.66 93 69.61 8.79 97 79.90 9.01 95
Ky. 4.91 59 0.07 5.67 145 0.07 5.83 35 0.03 6.43 92 0.16 7.15 44
La. 4.80 24 4.17 5.85 33 3.81 6.18 30 2.62 6.97 32 3.02 8.30 43
Md. 4.14 62 0.02 4.77 71 0.03 4.67 58 0.04 5.64 109 0.04 6.77 41
Minn. \a \a 0 \a \a 0 9.45 229 0.01 9.63 311 0\b 9.63 79
Miss. 4.95 40 0.43 5.92 34 0.84 6.12 32 0.40 7.21 34 0.34 8.60 42
Mo. 8.20 53 5.13 9.64 63 5.42 10.28 67 3.87 10.88 74 5.14 11.94 81
Nebr. 4.95 107 48.71 5.61 124 55.40 5.54 112 40.89 5.49 117 44.80 5.58 105
N. Mex. 9.89 41 3.41 11.67 40 3.55 13.12 46 3.27 16.25 57 3.91 18.39 62
N.C. 7.94 41 0.03 9.75 143 0.01 10.09 36 0.02 10.06 186 0.01 12.00 141
N. Dak. 17.91 23 0.03 20.30 24 0.01 18.52 27 0.01 \a \a 0 \a \a
Ohio 3.87 7 0\b 4.41 13 0.01 4.41 20 0\b \a \a 0 \a \a
Okla. 8.48 62 2.82 8.98 67 2.58 10.62 74 2.06 11.84 73 2.56 12.66 74
Pa. 5.85 49 0.03 6.93 26 0.04 7.06 44 0.05 8.55 26 0.02 10.26 33
S.C. 6.62 11 0.02 8.28 42 0\b 7.71 20 0.01 \a \a 0 11.42 24
S. Dak. 9.73 52 3.67 11.54 79 5.72 11.45 88 3.33 11.94 71 2.75 13.33 75
Tenn. 5.01 19 0.10 5.61 29 0.16 6.33 20 0.07 7.52 43 0.06 9.00 51
Tex. 7.71 69 81.50 8.45 75 97.40 9.31 79 66.02 9.83 77 73.33 10.60 89
Va. 6.32 26 0.10 7.20 32 0.05 7.32 37 0.09 9.08 52 0.12 10.96 44
Wis. 7.47 20 0.02 7.56 24 0\b 7.68 29 0\b 8.37 32 0\b 10.08 34
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Notes: For 1995, no figures are shown for insurance in force because
the year was not complete at the time of our review.
Premium rates are based on the weighted average of rates at the
65-percent coverage and average production level.
\a Data were insufficient to perform analysis.
\b Although some insurance was in force, the amount was too small to
show in this presentation.
Source: GAO's analysis of USDA's data.
Table III.5
Weighted Average Premium Rate, Adequacy
of Premium Rate, and Insurance in Force
by State for Soybeans, Crop Years 1991-
95
Adequacy
of
Premium Insurance Premium Insurance Premium Adequacy Insurance Premium Adequacy Insurance Premium premium
rate per Adequacy of in force rate per Adequacy of in force ($ rate per of premium in force rate per of premium in force rate per rate
$100 premium rate ($ in $100 premium rate in $100 rate ($ in $100 rate ($ in $100 (percent
State coverage (percent) millions) coverage (percent) millions) coverage (percent) millions) coverage (percent) millions) coverage )
------------ ---------- ------------ ---------- ---------- ------------ ----------- ---------- ---------- ---------- --------- ---------- ---------- -------- --------
Ala. $11.16 56 $3.32 $12.30 59 $2.13 $12.30 60 $1.52 $12.77 67 $1.71 $13.99 76
Ark. 12.46 67 21.43 13.83 70 17.07 14.04 69 17.11 15.41 78 25.41 16.80 80
Del. 7.60 122 2.13 6.79 116 1.98 6.52 113 1.64 6.74 121 2.45 6.76 102
Fla. 9.78 36 0.29 11.28 43 0.41 11.61 47 0.29 13.06 58 0.38 14.61 63
Ga. 14.90 63 3.28 16.34 63 3.16 16.84 70 2.65 18.97 81 4.90 20.11 82
Ill. 3.22 103 248.41 3.04 102 306.59 2.95 95 309.48 3.31 104 346.16 3.31 105
Ind. 3.49 95 88.63 3.43 93 107.37 3.38 90 107.90 3.60 99 120.31 3.66 101
Iowa 3.09 149 505.28 2.98 147 465.18 2.89 141 479.86 3.01 154 597.01 3.05 123
Kans. 6.25 48 49.87 7.26 65 46.83 7.59 63 52.66 8.42 73 70.73 8.71 74
Ky. 9.27 66 6.32 9.69 70 5.33 9.54 70 4.56 9.88 75 7.95 10.54 82
La. 14.76 68 19.60 17.06 75 17.83 17.18 68 18.17 17.94 71 19.79 18.38 70
Md. 8.05 94 3.49 7.33 93 3.57 7.31 91 3.73 7.89 88 6.33 7.92 81
Mich. 6.85 77 14.32 7.23 85 13.44 7.32 92 16.72 7.56 90 18.30 7.66 91
Minn. 4.75 129 349.02 4.61 127 319.27 4.52 131 323.53 4.65 126 455.52 4.98 101
Miss. 10.25 54 22.20 11.61 58 19.74 11.63 59 19.94 12.15 67 24.60 13.04 70
Mo. 6.61 76 82.59 7.34 86 74.54 7.22 81 70.82 7.70 87 115.65 8.12 86
Nebr. 3.73 80 112.12 3.77 87 108.76 3.77 86 112.65 4.02 93 138.75 4.05 91
N.J. 6.93 38 0.58 7.10 44 0.45 6.96 45 0.53 7.94 54 0.73 8.08 58
N.Y. 7.12 41 0.08 7.57 38 0.04 8.12 61 0.39 9.36 56 0.25 9.40 84
N.C. 10.53 68 5.37 12.45 78 4.02 12.96 84 4.70 13.76 84 7.40 14.02 83
N. Dak. 6.05 81 25.60 6.23 92 26.52 6.08 98 24.21 6.22 105 32.31 6.57 94
Ohio 4.69 110 64.84 4.59 108 70.84 4.42 100 74.94 4.59 108 77.66 4.61 112
Okla. 12.67 47 2.45 13.78 52 0.91 15.38 54 1.77 16.79 57 2.82 18.31 71
Pa. 6.32 50 0.91 6.00 67 1.32 6.04 51 1.11 6.79 59 1.47 6.90 64
S.C. 18.74 90 0.82 21.91 96 0.46 21.03 103 0.83 21.80 100 1.37 22.54 95
S. Dak. 5.16 92 94.37 5.14 98 97.71 5.25 93 79.39 5.40 94 162.75 5.65 84
Tenn. 9.74 58 3.54 10.66 62 2.69 10.28 64 2.29 11.03 82 4.26 11.99 79
Tex. 16.90 57 2.03 19.16 68 2.83 21.63 74 2.67 22.37 83 3.58 24.07 89
Va. 5.07 42 4.64 5.98 54 3.75 6.36 52 4.36 7.12 64 11.22 8.00 64
W. Va. 4.90 60 0.06 5.66 52 0.08 6.24 49 0.13 6.54 63 0.14 7.46 105
Wis. 5.39 80 5.28 5.26 85 7.11 5.51 83 8.88 5.94 78 14.91 6.23 71
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Notes: For 1995, no figures are shown for insurance in force because
the year was not complete at the time of our review.
Premium rates are based on the weighted average of rates at the
65-percent coverage and average production level.
Source: GAO's analysis of USDA's data.
Table III.6
Weighted Average Premium Rate, Adequacy
of Premium Rate, and Insurance in Force
by State for Wheat, Crop Years 1991-95
Adequacy
of
Premium Insurance Premium Insurance Premium Adequacy Insurance Premium Adequacy Insurance Premium premium
rate per Adequacy of in force rate per Adequacy of in force ($ rate per of premium in force rate per of premium in force rate per rate
$100 premium rate ($ in $100 premium rate in $100 rate ($ in $100 rate ($ in $100 (percent
State coverage (percent) millions) coverage (percent) millions) coverage (percent) millions) coverage (percent) millions) coverage )
------------ ---------- ------------ ---------- ---------- ------------ ----------- ---------- ---------- ---------- --------- ---------- ---------- -------- --------
Ala. $6.22 46 $1.70 $7.40 45 $1.26 $7.93 41 $0.82 $9.00 47 $0.80 $10.79 52
Ariz. \a \a 0 \a \a 0 \a \a 0 \a \a 0 3.89 296
Ark. 6.34 36 10.13 7.32 41 14.80 7.67 36 13.40 9.16 44 9.84 11.35 56
Calif. 9.57 50 2.88 10.05 51 2.54 10.60 53 2.10 9.39 55 1.59 10.16 57
Colo. 12.82 136 40.99 12.49 125 41.76 12.60 133 49.22 12.63 127 55.03 12.49 123
Del. 2.34 1,383 0.14 2.25 900 0.07 2.07 337 0.10 1.89 270 0.13 2.16 359
Fla. 8.03 39 0.32 9.33 49 0.23 9.00 41 0.20 10.40 57 0.17 11.75 71
Ga. 7.40 69 5.82 8.49 81 3.96 8.65 79 3.99 8.90 85 4.48 9.57 97
Idaho 4.05 53 23.74 4.27 58 25.71 4.23 59 26.65 4.71 61 23.99 7.24 67
Ill. 4.68 50 7.77 5.10 53 9.69 4.76 46 18.91 5.73 41 11.73 6.66 44
Ind. 3.67 83 2.39 3.74 79 3.03 3.33 71 11.08 3.94 55 6.63 4.56 61
Iowa 9.36 59 0.32 9.80 71 0.39 9.24 62 0.38 10.66 56 0.34 11.99 56
Kans. 6.68 105 263.33 6.61 99 301.16 6.76 97 303.38 6.74 90 310.42 7.13 90
Ky. 6.05 85 1.57 6.63 100 1.82 6.38 63 1.84 6.81 62 1.81 7.39 77
La. 8.06 27 5.57 9.31 31 3.91 8.99 26 1.85 10.75 29 1.97 13.24 34
Maine \a \a 0 \a \a 0 \a \a 0 \a \a 0 \a \a
Md. 2.51 306 0.20 2.59 161 0.15 2.55 172 0.12 2.44 217 0.15 2.58 505
Mich. 4.25 63 3.50 4.73 70 3.88 4.64 70 4.01 4.88 74 5.90 5.53 73
Minn. 7.46 96 112.07 8.12 111 121.58 7.98 98 119.41 8.69 110 135.47 9.20 90
Miss. 7.50 38 3.78 8.66 38 3.11 8.72 33 3.16 10.13 42 2.21 12.50 65
Mo. 7.95 71 11.40 8.89 80 9.96 8.60 70 11.57 9.82 78 12.79 10.68 78
Mont. 7.14 69 208.06 7.55 70 229.44 7.38 68 229.72 8.56 74 232.72 9.20 80
Nebr. 7.25 118 73.15 7.11 112 76.03 7.19 107 82.70 7.27 94 82.00 7.67 94
Nev. \a \a 0 \a \a 0 \a \a 0 6.39 32 0.18 7.65 45
N.J. 2.61 \a 0.01 2.64 2,008 0.01 \a \a 0 \a \a 0 \a \a
N. Mex. 14.48 47 5.59 13.91 51 4.14 15.13 58 3.54 17.13 66 3.69 17.78 70
N.Y. 4.34 99 0.09 5.06 118 0.12 5.26 84 0.02 5.43 52 0.08 \a \a
N.C. 6.14 62 2.09 7.00 72 1.81 6.92 74 1.91 7.17 80 2.68 7.85 90
N. Dak. 6.75 85 435.68 7.49 94 466.22 6.96 86 452.21 7.16 89 512.90 7.80 90
Ohio 3.15 99 3.45 3.24 95 4.50 3.06 94 7.63 3.40 76 9.09 3.91 80
Okla. 7.00 93 75.23 7.16 90 85.71 6.95 84 104.11 6.93 82 116.59 7.47 82
Oreg. 3.06 151 34.32 3.14 146 40.19 3.02 122 39.82 2.99 105 38.61 2.98 118
Pa. 2.40 136 0.06 2.59 80 0.09 2.66 91 0.05 2.74 119 0.09 2.96 96
S.C. 7.48 62 1.38 8.71 75 1.05 8.92 78 1.28 9.18 85 1.88 10.24 104
S. Dak. 11.13 83 66.98 11.78 92 80.57 12.04 90 77.62 12.08 89 82.00 10.82 95
Tenn. 7.06 59 1.20 8.16 56 1.57 8.50 43 0.65 9.09 44 0.55 10.88 58
Tex. 11.31 71 84.05 11.76 80 74.64 11.48 73 87.55 11.95 78 80.94 12.70 84
Utah 7.18 57 3.45 8.28 63 2.65 8.91 57 3.08 9.93 59 1.64 10.66 68
Va. 5.36 63 1.00 5.78 81 0.97 6.01 71 0.69 6.18 69 1.13 6.90 82
Wash. 3.20 112 90.67 3.35 110 115.96 3.42 97 103.71 3.50 95 97.40 3.34 109
W. Va. 4.39 17 0.04 4.15 40 0.03 4.33 35 0\b 5.40 20 0\b 6.66 17
Wis. 6.01 46 0.32 6.91 56 0.34 6.92 55 3.01 7.84 38 2.25 9.06 58
Wyo. 7.37 76 5.64 7.47 84 5.75 7.34 83 6.06 7.34 85 6.32 7.46 85
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Notes: For 1995, no figures are shown for insurance in force because
the year was not complete at the time of our review.
Premium rates are based on the weighted average of rates at the
65-percent coverage and average production level.
\a Data were insufficient to perform analysis.
\b Although some insurance was in force, the amount was too small to
show in this presentation.
Source: GAO's analysis of USDA's data.
--------------------
\1 USDA refers to the insurance coverage sold as liabilities.
ESTIMATED SAVINGS FROM USDA'S
PROGRAM TO TARGET HIGH-RISK
FARMERS
========================================================== Appendix IV
To estimate the annual savings resulting from USDA's program to
target high-risk farmers for individual rate increases and/or
decreases in their insured production levels, we obtained summary
information from USDA for crop years 1983 through 1991 on the
premiums paid by and the claims paid to 22,551 policyholders included
in this program for crop year 1993.\1 We also obtained data on the
insurance experience of those targeted for the program in crop years
1992 and 1993 who continued to purchase insurance. We reviewed the
estimate of savings made by a contractor and identified the reasons
for differences between the contractor's estimate and our estimate.
We determined that savings from the high-risk program consisted of
three elements: (1) the claims exceeding premiums that were avoided
because farmers dropped out of the crop insurance program once they
were targeted for higher rates; (2) the additional premiums paid by
farmers who, once targeted, remained in the program and paid higher
rates; and (3) the reduced claims paid to farmers who, once targeted,
remained in the crop insurance program.
--------------------
\1 This was the most recent year for which complete information was
available at the time we prepared our estimate.
GAO'S ESTIMATE
-------------------------------------------------------- Appendix IV:1
To make our estimate, we first determined to what extent farmers
continued to purchase crop insurance once they were targeted for the
high-risk program. Using USDA's analysis of farmers' decisions for
crop year 1992,\2 we found that (1) 36.7 percent had stopped
purchasing crop insurance before being targeted for the high-risk
program, (2) 41.2 percent stopped purchasing crop insurance once
targeted, (3) 18 percent continued purchasing crop insurance at
higher rates, and (4) 4.1 percent continued purchasing crop insurance
at the same rate. As table IV.1 shows, we estimate that the
resulting savings total $32.9 million.
Table IV.1
Estimated Savings Resulting From
Targeting High-Risk Farmers for
Increased Premiums and/or Decreased
Production Levels, Crop Year 1993
(Dollars in millions)
Source of savings Amount
------------------------------------------------------------ --------
Claims exceeding premiums that were avoided because of $25.3
farmers who, once targeted, dropped out of the program
Additional premiums paid by farmers who, once targeted, 2.0
remained in the program and paid higher premiums
Claims avoided for farmers who, once targeted, remained in 5.6
the program but had reduced claims
======================================================================
Total estimated savings $32.9
----------------------------------------------------------------------
To arrive at our estimate of $25.3 million in savings resulting from
farmers canceling their insurance, we subtracted the total premiums
paid by the 1993 target group for 1983 through 1991 ($152.5 million)
from the total claims paid over the same period ($704.4 million) to
determine the total amount that the claims paid exceeded the premiums
($551.9 million). We calculated an annualized estimate by which the
claims paid exceeded the premiums ($61.3 million) by dividing the
result by 9 (the number of years in the period 1983-91). We
multiplied this amount by 41.2 percent, the percentage of farmers who
we estimate quit purchasing crop insurance once targeted for the
high-risk program.
To arrive at our estimated $2 million in additional premiums for crop
year 1993, we divided the premiums the target group paid for 1983
through 1991 by 9 to arrive at the average annualized premiums paid
($16.9 million). We multiplied the result by 67 percent, the average
increase in premiums that farmers paid the first year they were in
the high-risk program,\3 to arrive at the increased premiums the
group would have paid if all continued to purchase insurance ($11.4
million). We multiplied this amount by 18 percent, the percentage of
farmers that we estimate continued to purchase crop insurance.
To arrive at our estimated $5.6 million in decreased claims for crop
year 1993, we multiplied our estimate of the annualized amount that
the claims exceeded the premiums ($61.3 million) by 51 percent, the
average decrease in the amount of claims farmers received the first
year they were in the high-risk program,\4 to arrive at the decrease
in the claims the group would have received if all continued to
purchase insurance ($31.3 million). We multiplied this amount by 18
percent, the percentage of farmers that we estimate continued to
purchase crop insurance.
--------------------
\2 This was the most recent period covered by the USDA analysis
available at the time we prepared our estimate.
\3 The 67 percent is based on our analysis of farmers included in the
program for the first time in crop year 1993.
\4 The 51 percent is based on our comparison of the total claims paid
in 1992 and 1993 to farmers who were in the high risk program in 1993
for the first time.
DIFFERENCES BETWEEN GAO'S AND
CONTRACTOR'S ESTIMATES
-------------------------------------------------------- Appendix IV:2
The primary reason for the differences between the USDA contractor's
estimate of $70 million in annual savings from the high-risk program
and our estimate of $32.9 million is the difference in the number of
farmers affected by the program. The USDA contractor's estimate was
based on the assumption that USDA would select 2 percent of all
policyholders for the program. In practice, however, USDA selected
1.5 percent. Furthermore, over one-third of those identified had
already stopped buying crop insurance before being selected for the
program. Therefore, the percentage of policyholders who were
actually affected by the program dropped to about 1 percent.
ESTIMATED SAVINGS FROM USDA'S
ACTIONS TO IMPROVE THE ACCURACY OF
FARMERS' PRODUCTION LEVELS
=========================================================== Appendix V
To estimate USDA's savings from the changes in the rules for
calculating a farmer's production level for crop year 1994,\1 we
obtained information on the 2,209,177 insurance policy units for the
six crops we reviewed from USDA's database for crop year 1994.
According to USDA officials, the primary savings would result from
claims payments that were reduced because of lower insured production
levels. Therefore, we estimated the reduction in production levels
that USDA approved for 1994. Our calculations were based on
1,174,037 policy units for which complete information was available.
According to our analysis, USDA decreased the approved production
levels for policy units for those farmers who had 2 years or less
actual production experience and increased the approved levels for
policy units for those farmers who had 3 years or more of experience.
(See table V.1.)
Table V.1
Estimated Changes in Farmers' Approved
Production Levels From Crop Year 1993 to
Crop Year 1994
Percentage change
in production
Number of years Number of level from 1993 to
of actual production policy units 1994
------------------------------ ------------------ ------------------
0 298,816 -35.0
1 295,738 -16.2
2 215,051 -3.6
3 191,715 +2.4
4 42,587 5.2
5 32,252 +3.4
6 21,903 +3.0
7 21,422 +3.2
8 20,029 +1.8
9 19,131 +1.0
10 15,393 .0
All combined 1,174,037 -12.8
----------------------------------------------------------------------
Source: GAO's analysis of USDA's data.
To estimate the impact of the changes, we divided the policy units
into the two groups and calculated a weighted average change for each
group, as follows:
Policy units based on actual production experience of 2 years or
less whose approved production levels were reduced from 1993 to
1994. The weighted average reduction was 20.15 percent.
Policy units based on actual production of 3 or more years that
were not subject to any reductions in approved production
levels. These policy units had a weighted average increase of
2.85 percent in approved production levels.
In developing our estimate, we assumed that (1) the experience for
all crops would be about the same as the experience for the six crops
we reviewed; (2) recent premiums of about $750 million annually for
all crops were representative; and (3) the claims paid would be $1.10
for each $1 in premiums, as the law requires. In addition, we used
USDA's assumption that claims payments would change by 1.5 percent
for each 1-percent change in approved production levels. We also
weighted the percentage change in production levels from 1993 to
1994--for the policy units for which we had complete information--to
all the policy units.
We initially estimated that claims would drop by $102 million for
farmers with an actual production history for 2 years or less and
increase by $21 million for farmers with an actual production history
for 3 years or more. The estimated net savings is $81 million. We
arrived at our estimate as follows:
For the farmers with 2 years or less of production history, we
multiplied the 1994 premiums of $306 million by the expected
loss ratio of 1.1 for an estimated $337 million in total claims
payments. We then multiplied the expected decrease of 20.15
percent in production levels by 1.5 percent to determine the
expected change in claims payments, for a total decrease of
30.23 percent, or $102 million.
For the farmers with 3 years or more of production history, we
multiplied the 1994 premiums of $444 million by the expected
loss ratio of 1.1 for an estimated $488 million in total claims
payments. We then multiplied the expected increase of 2.85
percent in production levels by 1.5 percent to determine the
expected change in claims payments, for a total increase of
4.275 percent, or $21 million.
Our initial estimate of $81 million is within the range of savings
that USDA estimated in its blueprint--from about $75 to $112
million.\55 However, USDA did not include in its estimate the
increase in approved production levels for farmers with 3 or more
years of production history. If USDA's estimate would have
considered the $21 million increase, our estimate would have been
near to the midpoint of USDA's estimate.
In implementing its changes in the rules, USDA imposed the following
limitations:
decreases were limited to no more than 10 percent annually and
increases to no more than 15 percent and
beginning farmers were excluded.
To estimate the impact of these limitations, we compared what the
1994 approved production levels would have been without any
limitations with the production levels that USDA approved. Following
the same methodology as in our previous calculation, we calculated
that these limitations reduced the estimated $81 million in savings
by about $37 million. Thus, we estimated the savings from USDA's
changes in the rules for calculating approved actual production
levels to be about $44 million for crop year 1994.
--------------------
\1 USDA refers to production levels as yield.
\55 Although USDA's blueprint does not specify the dollar amount of
savings expected, the Department estimated reductions in claims
payments from $1.40 for every $1.00 in premiums to between $1.25 and
$1.30, or between 7.15 and 10.72 percent. These percentages were
applied to $1.05 billion in claims payments--the average $750 million
in premiums in recent years multiplied by 140 percent-- resulting in
a range of about $75 million to $112 million.
MAJOR CONTRIBUTORS TO THIS REPORT
========================================================== Appendix VI
Carl Lee Aubrey, Project Leader
Stephen M. Brown
Thomas M. Cook
Ruth A. Decker
Donald L. Ficklin
Shirley A. Klaudt
Fredrick C. Light
Carol Herrnstadt Shulman
Robert R. Seely, Jr.
David R. Solenberger
*** End of document. ***