March 15, 2005
A SOYBEAN RUST SCENARIO MODEL: 2005 CROP YEAR DECISION MAKING IN ILLINOIS
The 2005 crop will be particularly challenging
for Illinois soybean producers. Soybean rust, a fungal disease,
has moved up from South American and was found in the Southern US
in the fall of 2004. This is the first discovery of the disease
in the continental US. The disease is in the form of spores and
can spread through airborne pathways over wide geographic areas
(Isard et al, 20041). Weather patterns,
especially those from the South to North, will be the main factor
causing an outbreak in Illinois during the 2005 crop year.
For farm managers the situation in Illinois
is very different than for producers in South America, who have
been dealing with rust the last three crop seasons. Low latitude
regions (closer to the equator) maintain spore populations all year-round.
While infection rates may be higher, because the spores are always
present, decision making is simpler. A treatment regime of multiple
spray treatments is integrated directly into the farm management
plan. In high latitude regions where a killing frost is present,
annual infection is windborne/weather related. This makes infection
timing and infection extent much more uncertain; making farmer decision
making uncertain. If the farmer makes the decision that rust is
a threat and a response is necessary, a second source of uncertainty
arises as to how the farmer should respond?
The purpose of this article and the associated
model is to help farmers address these uncertainties.
Resources and Valuable Literature
Though beyond the scope of this article there
are numerous articles and websites devoted to rust. In terms of
general web sites dedicated to rust several of value are:
These provide not only information but contact
people. The University of Illinois Extension and the Illinois Department
of Agriculture, as part of the State's Rust Response Plan, are working
closely to provide Illinois farmers with high quality information
For analysis of economic impact, take a look
at a recent publication from the USDA, "Economic and Policy Implications
of Wind Borne Entry of Asian Soybean Rust into the United States"
Also Goldsmith constructed a rust cost calculator
and Schnitkey looked at the costs of adjusting the corn:soybean ratio (see
All three economic analyses bring to light the levels of uncertainty
involved in decision making about rust.
Tactical Decision Making
The current rust environment is a tactical challenge
for the upcoming season. Farmers have to respond in some way. Even
no decision is a decision for the upcoming cropping season. It is
important to note first off that while it is probable that rust
will be found in Illinois in 2005, it is not certain. And if it
is found, it is unclear how widespread the infection will be across
the state. So decision making may take the full range of actions;
from doing nothing, to multiple spray treatments, to avoiding soybeans
completely. For example, the organic soybean farmers in the state
and their customers are particularly challenged because there is
no known organic treatment for rust (Leopold Center, 20042).
In tactical decision making there are two dimensions
to consider, whether a management variable is significant or trivial
and if a variable is controllable or not controllable (Figure 1).
While a problem may involve many variables, and in a perfect world
one would like to address them all, many times this can not be done.
Hard choices need to be made to focus on a few key variables. Effective
managers focus on the variables that jointly have the highest returns
to profitability and on which they have the most control, the upper
right-hand quadrant of Figure 1.
The rust problem for Illinois farmers can be
cast in such a framework. The uncertainty for managers arises from
a variety of sources and can be either experiential or probabilistic.
Experientially, soybean producers are unfamiliar with fungicides.
Spraying for rust will add managerial, as well as logistical, demands
on the State's spraying infrastructure, protocols, and skills. More
narrowly, rust is a new disease in the U.S. 2005 will be the first
year continental U.S. farmers will experience managing the disease
on their farms.
There are also probabilistic uncertainties with
respect to infection and rust's impact on markets. Unknown is when
rust might arrive during the growing season, how it might be distributed
across a particular region or farm, how it might affect a particular
field, and what its final impact might be on crop yields and prices.
Figure 1. Tactical Decision Making Framework
There are 16 main variables that affect profitability
within a rust response model (Table 1). In terms of tactical decision
making, these variables can be cast in the two-dimensional framework
(Figure 2). Each variable in some form has an impact and may or
may not be controllable. Impact can be high, medium, or low/none.
The same can be true for control; highly controllable, moderately
controllable, or not controllable. For effective decision making
under such uncertainty the objective is to focus management on those
few areas where control and impact in combination would most favorably
affect profitability. As an exercise the variables have been set
within the tactical decision framework. The placement of the variables
within the framework is for illustrative purposes. Each manager
should decide on their own where variables should be placed and
if variables need to be added or removed.
Figure 2. Tactical Decision Making Framework
Table 1. Rust Decision Variables
An immediate conclusion drawn from Figure 2 is
that 1st tier managerial activities should focus on scouting, disease
management, and spray-related decision making. There is plenty of
time before infection and large quantities of high-quality educational
materials are available. In this way, rust, as an event, creates
an opportunity for generating returns to good management. Those
who are prepared will out perform those who are not. So preparation
is the key to successfully deal with rust. The prognosis is good
that rust's impacts may only be moderate at worst given that: 1)
there is time; 2) information is plentiful; and 3) disease management
is a high impact/controllable area.
Second tier activities should focus on cropping
decisions, spraying equipment, and materials. As the scenario analysis
below relates, while having impact the corn:soybean ratio has less
impact on profitability than the tier 1 activities associated with
good spray management. Spray timing is critical, so delays beyond
a few days could be costly. Make sure that spraying equipment and
fungicide product will be available when you want and where you
want. Currently orders for fungicide are being taken but not fulfilled.
This makes preparedness a little more difficult, especially given
that it is not certain Illinois will see rust in 2005. Begin now
to develop good communication with fungicide sales representatives
and spray contractors.
Third tier activities might involve risk management
(e.g., crop insurance) and special approaches to managing rust (e.g.,
use of spray tracks and adjusting row widths.) While employing these
activities is highly controllable, their impact is generally thought
to be more moderate with respect to rust.
For example, if crop insurance coverage levels
were maximized at 85% and allowed for 15% losses before payout,
this may be less than the expected losses from rust. A producer
needs to ask: what is the probability of rust on my farm and what
will be the damage levels when best management practices are employed3?
If a farmer believed that the probability was, say 25% that rust
would be on his/her farm in 2005 and with losses of 10% with good
management practices, the expected loss is [25% x 10%], or 2.5%,
much less than the 15% threshold for crop insurance. Alternatively,
because of one's location or because of potential difficulties securing
timely treatment for rust, expected probability of losses may be
high enough to warrant crop insurance. Such an analysis needs to
be conducted by the individual producer.
Finally, there are the high impact variables
that are uncontrollable, such as grain prices and rust infection
rate and location. These variables need to be factored in and analyzed
but they can not be directly managed. Use of risk management tools
such as insurance, contracts, and pricing options may prove helpful
to address the associated risks.
Rust Scenario Model
A Rust Scenario Model4
was constructed to better understand the impacts of these uncertainties.
For simplicity assume a 1,000 acre farm. This
farm produces only corn and soybeans. In 2004 it produced 500 acres
of corn and 500 acres of soybeans (Table 2). In 2005 two overall
decisions have to be made, the number of corn acres planted versus
soybean acres and how to treat the soybeans, if planted, for rust.
In 2006 assume there are no threat or impacts from rust5. A partial
budget focusing only on variable costs is developed using 2004 FBFM6
cost of production averages for central Illinois grain farms having
highly productive farmland. Most central Illinois farms plant about
50% of their acres in corn and 50% in soybeans, suggesting that
these cost estimates are appropriate. Following Schnitkey and Lattz
(2004) the model assumes a 10% yield drag and a $10/acre additional
nitrogen cost for corn following corn. Prices and yields are the
last five year average for central Illinois; 173 bu/ac for corn
and 49 bu/ac for soybeans. Corn prices are $2.25/bu and soybean
prices are $5.20.
Table 2. Model Variable Settings
With respect to costs associated with rust, several
costs and yield impacts are important. Following Goldsmith (2004),
cost per spray is assumed to be $20. As a starting point assume
all soybeans are sprayed. Planting fewer soybeans will result in
fewer acres that will require spraying. Losses are assumed to be
5% under a two-spray regime7.
As noted above, the model farm planted 500 acres
of corn and 500 acres of soybeans) in 2004, which is a ratio of
50% corn and 50% soybeans. In this crop year (2005) the farm can
plant 50:50 again, 75:25, or 100:0 (all corn and no soybeans). In
2006 it is assumed that the ratio will return to 50:50, 500 acres
of corn and 500 acres of soybeans. These three scenarios are compared
across the three-year period 2004-2006.
Rust Scenario Model Results
Three scenarios are compared, 50:50, 75:25 and
100:0 over the three year period 2004-2006.
This is the base year. There are no impacts from
rust and there are 500 acres of corn and 500 acres of soybeans planted.
Total revenue for the 1,000 acres is $322,025.
Table 3. Total Revenue for 2004 (Base Year)
In 2005, the farmer must decide how much corn
to plant and how many soybeans to plant.
Under the 50-50 scenario total revenue is $315,655, falling 2% ($6,370)
from the base year (Appendix1a). Revenue falls because of a 5% yield
loss on the 500 acres of soybeans. Yields are reduced from 49 bu/ac
to 46.55 bu/ac.
Under the 75:25 scenario total revenue increases
almost 9% over the 50:50 scenario, to $342,721 (Appendix1b). Rust
affects on yield are reduced because 50% (250 ac) fewer soybeans
are planted. Revenue gains from reduced soybean acres though are
mitigated in part by the yield drag occurring on part of corn acres.
Because 750 acres of corn are planted in 2005 in this scenario,
250 acres will need to be planted on 2004 corn ground. Coming out
of the 2004 crop year there was only 500 acres of soybean ground
available for rotation. So there will be a yield drag of 10 bu/ac
on 250 of the 750 corn acres planted in 2005. Effective yield on
those 250 acres falls to 155.7 bu/acre.
Under the final scenario where no soybeans are
planted, revenues increase 17% over the 50:50 plan to $369,788 (Appendix
1c). Now rust has no impact because no soybeans are planted. Yield
drag affects occur on half (500) the planted acres.
Variable Costs 2004
In the base year total variable costs per acre
are $192 and $111 respectively for corn and soybeans (Table 4) and
$151,500 in total for the entire farm (Table 5).
Table 4. Variable Costs/Acre for 2004 (Base Year)
Table 5. Total Variable Costs for 2004 (Base
Variable Costs 2005
Under the 50:50 scenario variable costs for corn
are $96,000 and $75,500 for soybeans, for a total of $171,500 (Appendix
2a). This would be an increase of 13% over the base year. The only
extraordinary cost for soybeans is an additional charge of $20,000
for spraying fungicide two times (@$20 per) on all 500 acres of
Under the 75:25 scenario, there the additional
fungicide cost, but it is only applied on 250 acres. There is also
an additional cost ($10/ac) of supplemental nitrogen for the 250
corn:corn acres. Total variable costs rise to $184,250 (19%) over
the base year (Appendix 2b)..
Under the 100:0 scenario, there is no additional
cost from the fungicide but supplemental nitrogen is now needed
on 500 acres of corn:corn ground. Variable costs per acre rise 27%
over the base year to $197,000 (Appendix 2c). This is due in part
to increased usage of nitrogen but also because variable costs per
acre are higher for corn than soybeans.
Net Revenue 2004
Net revenue, revenue minus variable costs, in
the base year is $170,525 or 171/ac (Table 6).
Table 6. Net Revenue for 2004 (Base Year)
Net Revenue 2005
Net Revenue under the 50:50 plan falls 15% to $144,155
reflecting both the revenue and cost impacts from rust (Appendix
3a). Under the 75:25 and 100:0 scenarios the differences with the
base year are reduced (Appendices 3b and 3c). Net revenue for 2005
for the two plans is $158,471 (-7%) and $172,788 (+1%), respectively;
both superior in 2005 to the 50:50 plan.
As described above, changing the corn:soybean
ratio is not simply a static problem but is dynamic, with carry-over
implications for the following crop year. Assume that the farmer
returns in 2006 to a 50:50 ratio and there is no threat of rust.
Under the 1st scenario, 50:50 in 2005, there is neither a yield
drag penalty, nor a soybean yield loss due to rust in 2006. Total
revenues would rise to $322,025, the same as in 2004 (Appendix 4a).
In Scenario 2 that employed a 75:25 plan in 2005, revenues would
fall almost $10,000 (3%) compared to the 50:50 plan, to $312,294
(Appendix 4b). In this scenario, 250 acres of corn would be two-year
continuous corn so would suffer a 10% yield drag.
Finally, the third scenario where in the previous
year (2005) no soybeans were planted, all 500 acres of corn would
suffer a yield drag. Half of the 500 acres of corn would be two
years of continuous corn and the other half would be three years
of continuous corn. Yield drag losses were assumed to the same for
both8. Revenue would decline almost $20,000 (6%), compared to the
50:50 plan, to $302,563 (Appendix 4c).
Variable Costs 2006
Under the 1st Scenario, 50:50 in 2005, costs
return to the 2004 levels and total $151,500 (Appendix 5a). Since
the 50:50 rotation remains intact no supplemental nitrogen is needed.
Also since rust is not an issue in 2006 (by assumption) there are
no additional costs for fungicides.
In Scenario 2 that employed a 75:25 plan in 2005, costs would rise
in 2006 $2,500 above the base year to $154,000 (Appendix 5b) This
reflects the additional costs of added nitrogen on 250 of corn:corn
In Scenario 3, all corn in 2005, would require significant supplemental
nitrogen. Variable costs would be $156,500, rising $5,000 (3%) over
the base year (Appendix 5c).
Net Revenue 2006
Net Revenue in 2006 for the three plans would
be $170,525, 158,294, and 146,063, respectively, for the 50:50,
75:25, and 100:0 plans (Appendices 6a-6c). The 50:50 plan's variable
costs for 2006 would be exactly the same as the base year (2004).
The 75:25 plan's costs would be 7% higher and the 100:0 plan's costs
would be 14% higher due to the effects on revenue from the yield
drag and the effects on costs from the higher nitrogen rates.
Net Revenue 2004-2006
When all three years are combined all three scenarios
are quite comparable, within $4,170 of each other (Table 7). The
highest net revenue, $489,375, occurs by using the 100:0 plan. It
is .86% or $1.39/ac/yr greater than the 50:50 plan. Net revenues
are $487,290 (+.43%) and $485,205, respectively for the 75:25 and
Analysis And Conclusions
Though it would appear then the three plans are
quite comparable. In fact they may not be when alternative analyses
are conducted using different assumptions. The above analysis was
just one application of a model than can be significantly altered
depending on the user's assumptions. The model will be posted on
www.farmdoc.uiuc.edu to allow for more in-depth analyses.
Several key assumptions might be quite different under a more complete
exercising of the model. For example:
The above results assume prices are static; that there is no supply
effect. Just a 10% price response in 2005 in soybeans from $5.20
to $5.72, makes the 50:50 plan preferred by $12,000 (1.6%), over
the 2004-2006 time period, when compared to the 100:0 plan. Correspondingly
what might be assumed about corn prices?
Using an alternative line of reasoning, risks from rust are eliminated
by going to all corn. So when increased risks from rust are incorporated
into the comparison, the 100:0 plan may in fact be preferred. For
example, the above results assumed rust losses to be 5%. Just by
assuming losses to 10% reduces the net revenue of the 50:50 plan
by $7,000 over the three year horizon, countering the impacts of
a positive soybean price response.
Alternatively, while a 100:0 crop mix removes certain uncertainties,
it adds new uncertainties as well. All benefits from crop diversification
are lost. Risk is increased simply by concentrating all activities
in corn rather than spreading risk across the two crops. There are
also uncertainties and associated with producing continuous corn.
The model assumed a 10% yield drag, but is that correct? What if
2005 was a drought year? Or the model does not capture the out years
of 2007 and beyond. For example, just increasing the yield drag
to 15% and keeping nitrogen costs the same would reduce the 3-year
net revenue of the 100:0 plan almost $19,000 or $6.49/acre/year.
Also the model assumed that the only additional costs from continuous
corn were the application of higher rates of nitrogen. Continuous
corn might even require other increases inputs, i.e., insecticides
or be more vulnerable to input price inflation.
Finally, an additional significant risk is being poorly positioned
in the market if rust doesn't impact the U.S. or if there is a supply
response positively affecting the soybean corn price ratio. For
example, some farms or regions may not be affected, allowing those
producers who choose the plant soybeans to take advantage of a positive
soybean price response. Similarly some producers may be more effective
managing the affects of rust. They too would be able to take advantage
of a positive price response.
In conclusion, many rust management experts are confident that
good managers will manage any outbreak of rust effectively. While
Illinois soybean producers certainly face a new threat, there is
a real opportunity for good management to mitigate many of the risks
posed by rust. In a commodity business, farm profitability is not
simply a case of managing well, but managing differentially well.
This has certainly been the case in Brazil where superior rust managers
have garnered superior returns.
The model shows that exchanging one known, rust, for another unknown,
continuous corn, may not be warranted. Each farmer will need to
study their own situation, use their own assumptions and conduct
their own analyses for their particular situation. The key is to
focus on the high impact controllable variables and prepare. What
that means with respect to rust is learn about the disease, how
to scout it, and how to treat it. Then put in place a rust management
plan, ready to deploy in event rust does come to Illinois in 2005.
1. Isard, S., C. Main, T. Keever, R. Magarey, S. Redlin, and J Russo. (2004) Weather-Based Assessment of Soybean Rust Threat to North America. Final Report APHIS. https://netfiles.uiuc.edu/ariatti/www/SBR/Cycle.html
2. Leopold Center. (2004). "Science Behind the
News: Looking at Asian Soybean Rust." The Leopold Letter.
Vol.16 (4): p. 5.
3. See http://www.rma.usda.gov/news/2004/07/715soybeanrust.html
for a discussion of crop insurance coverage.
4. A decision tool utilizing the model will soon be available to for use on www.farmdoc.uiuc.edu
5. Assuming no rust threat in 2006 is highly unrealistic. The analysis though is simpler and no insights are lost.
6. Farm Business Farm Management Association http://fbfm.ace.uiuc.edu/
7. For a 1-spray regime 20% losses are expected
if 100% of the soybeans are infected and 2% losses are expected
for losses under a 3-spray regime. All loss percentages are based
on personal communication with Drs. Monte Miles (University of Illinois)
and Glen Hartman (USDA).
8. No data exists on the yield drag of three-year
continuous corn versus two-year. So to be conservative the same
10% loss compared to rated corn was used.
Issued by: Gary Schnitkey, Department of Agricultural
and Consumer Economics