Development of Corn Stover Biofuel : Impacts on Corn and Soybean Markets and Crop Rotation

What would be the impacts of a viable market for corn stover? A partial equilibrium model and a linear programing model were used to determine to what extent the existence of a viable market for corn stover would affect the traditional corn-soybean crop rotation in the US. We find that with government support production of biofuel from corn stover could significantly increase. That boosts profitability of farming corn in combination with harvesting corn stover versus soybeans. We show that if corn stover is demanded for biofuel, then a major shift will be observed in crop rotations in the US.


Importance
First generation biofuels and their impacts on: 1) greenhouse gas emissions, 2) oil imports, and 3) markets for agricultural commodities and food prices have been examined from different angles in recent years.These studies showed that first generation biofuels, which are produced from food crops will not be able to replace a large portion of oil-based liquid fuels, because their rapid expansion could cause adverse impacts on food supply (Abbott, Hurt, & Tyner, 2008;Abbott, Hurt, & Tyner, 2011;Trostle, 2008;Zilberman et al., 2013) and/or induce major unintended land use changes which in turn will lead to increases in greenhouse gas emissions (Hertel, 2010;Tyner & Taheripour, 2012).Instead, second-generation biofuels produced from forest or crop residues such as wood chips, corn stover, or wheat straw offer an alternative to first generation biofuels.These residue-based biofuels could have minor impacts on both food prices and land use change.In addition, they could make a major contribution in greenhouse gas emission reduction targets if removed in a sustainable manner (English et al., 2013).Another advantage for second-generation biofuels produced from agricultural residues is that they provide a potential new source of income for farmers (Thompson & Tyner, 2013).
If second generation biofuels became economically viable and a massive volume of biofuels are produced from agricultural residues, it then can have major impacts on agricultural commodity markets.Corn stover is a major and abundant feedstock in the USA, which is expected to be used in biofuel production, if the technology becomes economically viable.Converting corn stover to biofuels, if a significant market develops, can affect profitability of corn production versus other crops produced in the USA, in particular.Thompson and Tyner suggest that if a viable corn stover market existed, it could have a large impact on farmers' crop rotations and land allocation decisions.However, that research was done at the farm level with a given set of crop prices and ignored interactions between farm and market level variables.The purpose of this paper is to determine to what extent producing biofuels from corn stover can affect demands for and supplies of corn and soybeans and their market prices.In addition, it examines to what extent producing biofuels from corn stover could affect switching from corn-soybean rotation to continuous corn.
To determine how a corn stover market would impact corn and soybean markets, a partial equilibrium model was developed which links corn, soybeans, corn stover, ethanol, dried distiller grains, and gasoline markets at the USA national level and determines market-clearing prices at different crude oil prices.The resulting prices were used in a linear programming model to determine how farmers would allocate their land.Converting corn stover to biofuel could change relative profitability of corn and soybeans in favor of corn, when corn stover price is higher than its production costs.Production costs of corn stover include fertilizer costs to maintain productivity of land after stover removal and collection, storage, and transportation costs (Thompson & Tyner, 2013).These authors have concluded that if a viable corn stover market existed, it could have a large influence on farmer's crop rotation in favor of corn-corn rotation.However, they ignored the fact that relative prices of corn and soybeans could be different in the presence of a viable market for corn stover.
Consider a case where converting corn stover to biofuel is profitable, either due to market forces or government supports.In this case, farmers who produce corn and soybeans will bring profitability of corn stover collection into account.When corn stover production is a profitable operation for a farmer, in each planting period he/she will compare gains from a joint production of corn and corn stover with profits from soybean production at given prices (Note 1).In this case, if the joint profits from corn production and corn stover is higher than the profit from soybean production, then corn will be produced.If a large group of farmers decides to follow this choice, then the corn-corn rotation will increase among farmers, which will lead to an increase in corn production and a reduction in soybean production, other factors being constant.This could affect relative prices of corn and soybeans at the market level.
In this paper we first develop a partial equilibrium model to examine impacts of converting corn stover to ethanol on markets for ethanol, gasoline, corn, corn stover, and soybeans at the USA aggregation level.The model developed in this paper is based on the model developed and used by (Tyner & Taheripour, 2008).Then we will feed the results of this model into the Purdue Crop/Livestock Linear Programing (PCLP) model (Doster et al., 2008) to examine farmers' land allocation behavior in the presence of a viable market for corn stover.We test the sensitivity of the land allocation process at the farm level with respect to changes in key economic factors.

Partial Equilibrium Model
The partial equilibrium model developed in this paper is an extended version of the model developed by and used in several other articles (Taheripour & Tyner, 2008;Tyner & Taheripour, 2007;Tyner, Taheripour, & Perkis 2010;Tyner, Taheripour, & Hurt August, 2012).The original model follows an integrated partial equilibrium modeling structure which displays linkages among crude oil, gasoline, ethanol, and corn markets.The model captures the demand and supply sides of the corn, corn ethanol, dried distiller grains, and gasoline markets.The demand side of the corn market consists of three major corn users: foreign users (q cxd ), domestic uses for food and feed (q cdd ), and the corn ethanol industry (q ced ).The foreign and domestic demands for food and feed are modeled using constant price elasticity functional forms.The corn demand for the ethanol industry is q ced = y.qse .Here, y shows the corn-ethanol conversion factor and q se is quantity of ethanol supply.Hence, total demand for corn is: q cd = q cxd + q cdd + q ced .On the supply side, a constant return to scale Cobb-Douglas production function is used to model the supply side of the corn market.In this production function, capital, labor, land, and a composite input (which represent fertilizer, pesticides, seeds, energy and other items) are used to produce corn.In this function all inputs, except the composite input, are constant in the short-run.This production function is used to determine the supply for corn, q cs .In this model, in the short-run the demand for liquid fuel, g gd , only responds to its own price using a constant price elasticity functional form.However, in the long-run, demand can grow with income and population.The supply side of the fuel market is comprised of gasoline producers and ethanol producers.Gasoline supply is produced from crude oil.The supply of gasoline, q gos , is a function of its price and the price of crude oil.It follows a constant elasticity functional form as well.Ethanol is produced from corn.The supply of ethanol, q es , is a function of its own price and the price of corn following a constant elasticity functional form as well.In this model it is assumed that every gallon of ethanol is presumed to contain 70% of the energy of a gallon of gasoline.Hence, total supply of gasoline equivalent is: q gs = q gos + 0.7*q es .Distiller's Dried Grains with Solubles (DDGS) is a co-product of the ethanol industry.DDGS can be used as a substitute for corn and to some extent soybean meal in the livestock industry and also alleviate impacts of ethanol production on the corn market.DDGS can increase profitability of ethanol industry as well.As a substitute for corn, DDGS covers a portion of corn demand: q DDGS = ϒ.q ced .Here q DDGS is the quantity of DDGS produced and ϒ is the corn-DDGS conversion factor.The model evaluates profitability of the ethanol industry and assumes that this industry will expand/contract until profits reach zero.Given this assumption and according to the predetermined supply and demand elasticities, the model determines equilibrium prices and their corresponding quantities for corn, ethanol, and gasoline and other endogenous variables for given exogenous variables defined in the model (such as crude oil price).
Several new components are introduced into this partial equilibrium model to handle new markets for soybeans, corn stover, and a biofuel produced from corn stover.The first component added to the model is a drop-in biofuel named bio-gasoline.This drop-in biofuel can be produced from corn stover with the following supply function: q gstov = A estov (p pg ) gstov (p stov ) -gstov .The energy content of bio-gasoline is assumed to be equal to the energy content of gasoline.In the presence of bio-gasoline, total supply of gasoline will be equal to: q gs = q gos + 0.7*q es + q gstov .Similar to ethanol industry, the bio-gasoline industry will expand until profits reach zero.Profits per gallon of bio-gasoline are estimated by: πs = p g -cap stov -var stov .Here, cap stov and var stov are capital and variable costs per gallon of bio-gasoline.At equilibrium πs=0.
In the new model it is assumed that the capital costs of producing corn ethanol is zero when markets operate below the existing production capacity.However, if an expansion in capacity is required the model takes into account the required capital costs.
The supply of and demand for corn stover are added to the model as well.The supply of stover is presented by: q stov = A stov (p stov /c stov ) estov .Here q stov and p stov represent supply and market price of stover, A stov is the constant term, c stov stands for production costs of stover per metric ton (including all costs items such as collection costs, costs to maintain land productivity, and transportation costs), and estov indicates price elasticity of supply.The production costs are divided into two segments of fixed and variable costs.The variable costs are assumed to be sensitive to changes in crude oil price to cover impacts of changes in crude oil price on collection and transportation costs of stover.The corn stover total production costs follows an increasing trend from $68 per ton at $60 per gallon of crude oil to about $100 per ton at $160 crude oil price.The demand for corn stover is determined by bio-gasoline production with the following equation: qd stov = q gstov (V stov ).Here V stov is the conversion rate of corn stover to bio-gasoline.For details about the cost structure of the bio-gasoline industry and corn stover activity see Fiegel (Fiegel, 2012).
A market for soybeans is also added to the model.In this market, the demand for soybeans is defined as: qd soy =A soy (1/(p soy ) esoy ).Here qd soy and p soy represent demand for soybeans and its price, A soy shows the constant term of the demand function, and esoy indicates the own price elasticity of demand for soybeans.The model determines the supply of soybeans (qs soy ) using the allocated land to this product (l soy ) from the following equation: qs soy =l soy .yieldsoy .Here, yield soy represents soybean yield.The model assumes that total supply of land (l tot ) for corn and soybean production is fixed in the short-run.Hence, it determines areas under soybean production using the following relationship: l soy =l tot -( q cs /yield corn ).
Finally, the model imposes a zero profit condition to allocate land between corn and soybeans.Indeed the model assumes that farmers maximize their profit when they allocate their land between corn and soybeans.It is assumed that at equilibrium: π soy =π corn +π stov .
The revised model is calibrated using data obtained for 2010 for the US economy.For details see Fiegel (Fiegel, 2012).Then the model is solved for several alternative crude oil prices (an exogenous key variable of the model) ranging from $60 per barrel to $160 per barrel.For each crude oil price two alternative policies are tested.The first policy assumes that the government will not support converting corn stover to biofuel.The second policy assumes that the government will pay a fixed subsidy of $1.01 per gallon of bio-gasoline.Note that in both cases mentioned here we assumed that the government pays no subsidy for corn ethanol.

Farm Level Model
To examine impacts of having a viable market for corn stover on crop rotation at a farm level the PCLP model which was originally developed at Purdue University is used and modified.PCLP is a linear programming model which helps determine profit-maximizing decisions for a given farm according to its possible crop activities, its existing resources, and according to current prices of commodities and input costs.The PCLP model takes specific data such as land, labor, capital, crop yields, crop prices, and detailed input costs and determines activities which maximize farmer's profits.To adapt this model a new activity called stover collection is added to this model.This activity covers all steps and their corresponding costs which are required to collect and sell corn stover to a bio-refinery at current prices.These steps and their corresponding costs are defined in detail in (Fiegel, 2012;Thompson & Tyner, 2013).Then the model is solved for a group of farmers who participated in the Top Crop Farmer Workshops in 2007-2010 under two alternative cases of with and without corn stover activity to find out their optimal choices under these two different cases.To tune the PCLP model with market conditions in the presence of corn stover activity, prices obtained from the partial equilibrium model are used.Then the sensitivity of the results with respect to changes in the assumptions and parameters behind the corn stover activity are examined.

Impacts at Market Level
The simulation results cover a wide range of cases for many variables under all alternative scenarios developed for this paper.In what follows, we only present a selection of key results which are important for our analyses.Two key variables are outputs of corn ethanol and stover bio-gasoline under alternative cases.The simulation results are presented in Table 1.  1 shows that the supply of bio-gasoline will be very limited at low levels of crude oil price in particular when the government does not support bio-gasoline production.However, with a bio-gasoline subsidy, the market will produce significant amounts of bio-gasoline especially at medium and higher crude prices.A similar impact would result from a mandate such as the Renewable Fuel Standard.For example, when the crude oil price is around $100 per barrel, bio-gasoline production will be about 0.56 billion gallons when no subsidy is paid, and it will be about 6.33 billion gallons with subsidy.The first case provides a status quo situation where there is no market for corn stover.The second case evaluates impacts at farm level in the presence of a market for corn stover.In this case it is assumed that: 1) corn-corn rotation with stover removal reduces tillage costs by $25 per acre; 2) corn-corn rotation does not affect corn and soybean yields; 3) corn stover farm price and corn stover delivery price are about $85.40 per ton, and $111.80 per ton, respectively; and 4) farmers will use the rake and bale system to remove corn stover.In cases iii to vii these basic assumptions are relaxed.In case iii it is assumed that corn-corn rotation with stover removal does not reduce tillage costs.Case iv takes into account the fact that rotation can affect yield.In case v the farm level and delivery prices used in the base case are reduced by 20%.Case vi investigates impacts of using a new technology to remove corn stover.This technology is known as the corn-rower.Compared to the rake and bale system, this technology eliminates the need to rake the stover after harvesting corn and in general reduces the costs of stover activity (Fiegel, 2012).Finally, the last case assumes that farmers will use the corn-rower technology to harvest corn stover, but this technology does not reduce tillage costs.
While the above experiments are tested for individual participant farmers, in what follows we report the results obtained from the pool of participant farmers.The key simulation results are shown in Table 2. Table 2 shows that in the base case when there is no stover removal activity, participant farmers allocate their land mainly to corn-soybean rotation or only prefer to produce soybeans.In this case farmers only allocate about 14.7% of their land to the corn-corn rotation.However, when corn stover removal is introduced as a new option under the base case scenario farmers allocate about 66% of their land to the corn-corn rotation and remove stover from their land.In this scenario about 15.8% of available land will remain in corn-soybean rotation.In this case corn stover will be removed from 78.2% of available land at a rate of 1.18 tons per acre.This shows that if corn Figure 1

Table 1 .
Key simulation results obtained from partial equilibrium model The simulation results indicate that the supply of corn ethanolTo examine impacts of a viable market for corn stover at a farm level we tuned the PCLP model with market clearing prices obtained from the PE model as described above.We then made the following experiments for each farm to assess its response with respect to changes in key economic variables:

Table 2 .
Land allocation patterns for alternative scenarios examined using PCLP model at farm level (figures are shares in total acreage in percent except otherwise noted)