Disclaimer: This web page is designed to aid farmers with their marketing and risk management decisions. The risk of loss in trading futures, options, forward contracts, and hedge-to-arrive can be substantial and no warranty is given or implied by the author or any other party. Each farmer must consider whether such marketing strategies are appropriate for his or her situation. This web page does not represent the views of Kansas State University. 
Disclosure:  Dr. Barnaby’s research was the basis for the privately developed Crop Revenue Coverage.

Disaster Aid Update and the Impact of Yield Variance on Disaster Aid Policy[1]

Define Wheat Price Cap Limit.  Disaster aid sign-up is scheduled to start on June 6, however there still remains a critical variable not defined.  As of the end of May, USDA had not decided which National Agricultural Statistics Service (NASS) wheat price to use for determining the disaster aid payment cap for wheat.  The NASS seasonal average prices are reported in their February report and include a wheat price for all wheat, winter wheat, spring wheat, and durum wheat.  The NASS all wheat price of $3.60 would be preferred by Kansas wheat growers for setting payment caps.  The alternative is to use the winter wheat price, which is $3.45.  It is possible the lower NASS price would reduce disaster payments for some wheat growers with high levels of crop insurance coverage. 

One would expect that a USDA decision will be forthcoming soon.  One would think a decision will be needed prior to the sign-up date.  One possible outcome is for USDA to use the all wheat NASS price for all classes of wheat except for durum.  In the past, USDA has often split durum wheat out separate from other classes of wheat.  All of the wheat disaster aid analyses published on this WEB site assume the higher $3.60 NASS wheat price.

The per acre payment cap was defined as 95 percent times average yield times the maximum of the APH price election or the NASS price.  The original payment cap was based on the APH price but later was changed to include the NASS price, if higher.  This change increased the expected budget cost for the disaster program.  As a result of this change, it is expected few if any Kansas producers will suffer a reduction in disaster aid payments on 2002 claims.  Growers also have the option to select the 2001 loss year if that is to their advantage. 

Non-Harvested Acres.  Disaster aid payments will also be reduced for acres that are not harvested.  The non-harvested disaster aid reduction was also applied to past disaster programs.  The elimination of harvest costs reduces grower paid expenses.  Therefore, the disaster aid payment reduction is “justified.”  Those Farm Service Agency (FSA) reductions in disaster aid payments are 12 percent for corn, wheat 12 percent, milo 14 percent, soybeans 15 percent, and cotton 19 percent (table 1).  For example, growers who accepted a corn loss appraised settlement from their insurance company, and then used their cows to salvage any remaining forage will be subject to the disaster aid reduction for non-harvesting.  FSA will calculate the corn disaster payment and then multiply that payment times 88 percent to generate the net disaster aid paid to growers who did not mechanically harvest.  However, growers who salvage a disaster corn crop by chopping the remaining forage for livestock feed would not be subject to an unharvested payment reduction.  It is unclear if the non-harvested payment reduction would apply to growers who sold their crop for salvage to a neighbor who then chopped the remaining corn stalks for livestock feed.  In this case, the remaining corn was mechanically harvested but not by the owner/operator.

The non-harvested payment reduction is not applied to crop insurance.  Under the crop insurance program, once the crop is appraised and the loss settled, growers are then free to salvage the crop by any method they desire.  They also have the option to simply till the remaining plant material. 

Many growers have complained that crop loss adjusters generally will not enter a zero yield even if there appears to be no production in the field.  Typically, these growers have argued that loss adjusters will nearly always report at least a one bushel yield remaining in the field.  Growers have argued that, once yield losses are that severe, the crop is not worth harvesting and the yield should be zeroed out generating a payment equal to the total coverage.  However, these same growers should feel fortunate that crop insurance does not apply the disaster aid rule and reduce the indemnity payment for non-mechanical harvesting.  Like so many public policies, this is another example of, “is the glass half full or half empty?”

Disaster Aid versus Crop Insurance.  Disaster aid versus crop insurance will again be debated among growers and policy makers.  Under the current crop insurance program, the subsidy system favors the high risk growing areas over the low risk growing areas.  Disaster aid is simply crop insurance with a 100 percent premium subsidy.  Therefore, disaster aid more strongly favors high risk growing areas over low risk growing areas.

For example, let’s assume four different locations ranging from a high level of yield risk to very low yield risk. The total premium before any subsidies are provided effectively equals the expected payment to growers assuming the insurance contracts are rated correctly.  In the long run, total premiums equal the expected total indemnity payments.  In the case example the premium equals the long run expected average payment of $14.50.  The annual indemnity payment would obviously vary from zero to a maximum payment equal to the liability.  Therefore, one would need to consider a period of time, for example 20 years.  One would sum the expected indemnity payments over the 20 year period then divide by 20 and the result would be $14.50 (line 1, table 2).  All 4 levels of risk exposure in this example have exactly the same expected long-run average annual payout of $14.50.  These four example farms also have exactly the same average yield of 150 bushels (line 2, table 2). 

However, these four hypothetical farms do not have the same amount of yield variability, as measured by the coefficient of variation (CV) (standard deviation divided by the mean).  Because the very high risk farm has a larger standard deviation, it only requires a 60 percent guarantee to generate an expected annual average payment of $14.50.  If the coverage level were raised to 65 percent, then the expected payment would be more than $14.50.  In the very low risk production region it requires an 80 percent guarantee to generate the same expected payout as a 60 percent guarantee in a high yield risk region.  The production guarantee is 90 bushels with a 60 percent guarantee and 120 bushels with an 80 percent guarantee (line 5, table 2).  However, the grower with a 120 bushel guarantee has less yield variability therefore, has the same odds of collecting as the grower with a 90 bushel guarantee in the high risk growing area.  The dollars of coverage per acre is the guaranteed bushels times the MPCI price election of $2.20 (line 6, table 2). 

Under the present subsidy system, the Risk Management Agency (RMA) pays 64 percent of the premium for the high risk example grower while paying only 48 percent of the premium for the low risk grower, yet both individuals have exactly the same expected payout (line 7, table 2).  Obviously because the premium subsidy percentage is higher for the high risk farm, the dollar per acre subsidy is also higher on line 8.  Line 9 shows the reduction in premium subsidy compared to the very high risk grower.  In this example, the low risk grower receives $2.32 per acre less premium subsidy then does the high risk grower, which translates to a 25 percent reduction in the dollar per acre premium subsidy (line 10, table 2).  The result is the low risk grower pays $5.22 in premium and expects to receive back $14.50 in indemnity payments over the long run.  In the very low risk growing region the grower pays $7.54 per acre premium with exactly the same expected payout of $14.50 over the long-run.  This represents a 44.4 percent increase in farmer paid premium for the low risk grower over the high risk grower (line 13, table 2).

Current 2003 MPIC-APH Corn Rates.  An example using current 2003 corn rates for Multiple Peril Crop Insurance based on Actual Production History (MPCI-APH) was developed in table 3.  The Texas and Colorado corn farms are irrigated, while the Iowa and Illinois corn farms are dryland.  The APH was assumed to be 150 bushels in all 4 locations.  Notice the long run average annual expected payment does vary because these are real rates but still are approximately $14.50.  In the actual farm example in table 3, the expected annual average payouts ranged from $14.42 to $14.87.  This analysis also assumes the long-run loss ratio for all 4 locations is 1.0.  The “long-run” 14 year loss ratio in this particular Illinois County is substantially below 1.0 but for the purposes of this analysis it was assumed the loss ratio will equal 1.0 once more years are added to the payment history.  With a loss ratio substantially below 1.0, Illinois growers effectively would not have collected $14.87 over the years from 1989 – 2002.  However, if 1983 and 1988 data were available and included in the loss ratio the expected payout of $14.87 might have been the result.

Assuming that these premium rates are set correctly it would require a 60 percent guarantee for the Texas grower to generate approximately $14.50 in long run average annual payments.  However, in Illinois it would require an 80 percent guarantee to achieve approximately the same expected payout.  As in the case example, the difference in farmer paid premiums is caused by the premium subsidy levels.  The Texas grower would receive a 64 percent premium subsidy, while in Illinois USDA would only pay 48 percent of the grower’s premium.  The result is the Texas grower receives a higher dollar per acre subsidy than does the Illinois grower with the same average yield and the same expected payout.  As the result of the higher subsidy, the Texas grower pays $5.19 per acre premium while the Illinois grower pays $7.73 premium per acre, yet both producers have approximately the same expected payout (line 11, table 3).  As a result, the Illinois grower pays 49% more in premium then the Texas grower but has approximately the same expected indemnity payment over the long-run. 

Because disaster aid provides the same percent yield guarantee (65%) and a 100 percent premium subsidy, a disaster aid policy clearly favors Texas or other high risk growing areas over low risk growing areas like Illinois .  Disaster aid policy assumes paying a percentage of growers’ average yield is “fair” for all growers.  However, disaster aid policy does not consider the CV around the mean yield.  Because the grower in Texas has a higher CV than the Illinois producer, the result is that the Texas producer has a much higher probability of collecting from a 65 percent guarantee under a disaster aid program than does an Illinois producer. 

One could argue growers who have exactly the same mean yield should have exactly the same expected payment under a disaster aid program.  However, in order to provide the same expected disaster aid payment it would require USDA to incorporate a measure of variance and not base payments solely on the mean yield.  One method for generating similar expected disaster aid payments would be to provide different disaster aid coverage percentages to reflect yield risk by state or crop reporting district. 

Crop insurance provides less discrepancy between the high risk and low risk growers because there is a premium paid by the producer.  However, if public policy wanted to make expected indemnity payments equivalent between low risk and high risk production areas one alternative is to provide the same percentage premium subsidy regardless of coverage or type of insurance contract selected.  If all 4 locations received exactly the same percentage premium subsidy then all 4 farms would pay exactly the same premium because all 4 locations have exactly the same expected payout (table 2). 

Disaster aid benefits are based off the assumption that the mean yield for setting benefits is “fair” across all regions, all producers.  Crop insurance provides less of a benefit than disaster aid to high risk producing areas because the grower must pay a share of the program cost.  Unlike disaster aid that only considers the mean yield, crop insurance also includes a variance measure that is incorporated into the premium rate.  However, the current crop insurance subsidy system still skews the benefits towards the higher risk growing areas. 

Other farm programs.  Other farm programs also base their benefits off of the mean yield, so the obvious question is does the standard deviation of the mean yield have any impact on the benefits to individual growers in high risk areas versus low risk growing areas?  The yield variance would have little, if any, impact on the direct payment because the mean yield used to set direct payments is based on early 1980’s yields.  In addition, these yields may not be the ones produced by the current owner/tenant because the direct payment program yield is attached to the land.  Also different CVs of yields would have no impact on the counter cyclical payment but the payment would be effected by the CV of price.

Based on simulation analysis, the marketing loan appears to be more favorable for low risk areas, or exactly the opposite of the crop insurance program that is more favorable to high risk growing areas under current policy.  Based on those simulations the analysis suggests the marketing loan favors farms with a low CV and a “large” negative price-yield correlation.  Illinois farms, at least north of I-70, probably are more likely to have a low CV and a negative price-yield correlation.  While the marketing loan was advantageous for growers in low risk farming regions combined with a negative price-yield correlation, the advantage was not large.  When considering disaster aid, crop insurance, or even other farm programs, policy makers need to understand that the mean yield is not the only consideration when evaluating the benefits from most farm programs.  Yield variability can clearly alter farm program benefits and how those benefits are distributed between different regions of the country.

CV Effect on Other Policy Alternatives for Disasters.  Other disaster related public policies will affect different regions of the Country based on yield variance, price-yield correlation and mean yield.  In a previous WEB page update, analysis was provided that showed the greatest financial loss occurred with a 35 to 40 percent yield loss.  Disaster aid payments require yield losses greater than 35 percent to trigger payments and net crop insurance payments are normally small with yield losses under 50 percent.  Growers with a large CV will benefit the most from a traditional disaster aid program.  Public policies that target multiple year losses also favor growing regions with high CV of yield.

A companion disaster aid program that triggers payments with an 8 percent yield loss and the full benefit is paid out with about a 43 percent yield loss would shift more payments to the lower risk growing regions when compared to a 35 percent deductible under current disaster aid policy.  Under this approach growers would receive no payments for yield losses over 43 percent but that level of loss could be covered with crop insurance.  Growers in low risk regions would be more likely to collect with an 8 percent deductible than a 35 percent deductible.  Also growers in low risk growing regions are less likely to suffer losses greater than 43 percent thus increasing the odds they would collect the full disaster aid benefit.  While this does shift more of disaster benefit to lower risk growing regions, the high risk growing regions would still have a higher frequency of disaster aid claims caused by a larger (CV) of the yield.

Putting a second payment trigger on the counter cyclical payment would trigger payments because of either low prices (current policy) or low yields (the second payment trigger).  If market prices increase there is no counter cyclical payment under present policy because it is assumed growers sell their crop for the higher price.  However, if they have a crop failure there is noting to sell at the higher price.  If the counter cyclical payment were also triggered by yields below 92 percent of the average yield then growers could also collect the counter cyclical payment in years of higher prices but no yield.  Those growers with yields would collect no counter cyclical payments when prices are above the counter cyclical strike price.  This policy would also limit payments to crops that have a counter cyclical payment program. 

One would think this zero/92 approach would also favor regions that have a high CV of yield.  However, the yield trigger would only apply when prices are above the counter cyclical strike price and that would favor regions with a negative price-yield correlation.  Regions and States like Illinois that have a high negative price-yield correlation are more likely to collect under the second yield trigger than current policy because prices tend to increase and eliminate the counter cyclical payment when the Corn Belt has yield losses.  Under current policy growers in regions that have a zero price-yield correlation are more likely to collect the counter cyclical payment in years when they have a yield loss than growers in regions with a negative price-yield correlation.

Some growers in the high risk growing regions have argued revenue insurance should be eliminated and all subsidies concentrated on yield coverage only.  If a grower has a zero yield then revenue insurance and yield insurance would pay exactly the same (assuming the price elections were the same). While true for revenue insurance, replacement revenue insurance could pay more than the yield only insurance but it would require an increase in market prices. 

In regions with high CV of yields relative to price CVs, the yield only converge (MPCI-APH) continues to be a popular crop insurance choice.  However, revenue insurance has been a popular crop insurance choice in the Corn Belt .  In this region, the CV of price may be (is) larger than the CV of yield.  Based on participation one would have to conclude growers have a very good understanding of the negative price-yield correlation and the CV for their yields and prices.

Summary.  Should public policy continue to provide disaster aid based on a single 35% yield trigger?  Should a crop insurance subsidy policy that does not recognize the differences in the CVs of yields continue?  Should public policy provide additional assistance for multiple year losses?

How one answers those questions will likely depend on the yield variability they experience.  There is no right or wrong answer; the final decision will likely be a political compromise.  However, thinking about the expected mean yield and the CV of yield for different regions may help public policy makers create more effective public policies for crop disasters.