The Impression of Corn: Ethanol Policy’s Effect on America’s Largest Commodity 1978: First ethanol tax credit to blenders ($0.40/gal) 1983: Tax credit raised to $0.50/gal Luke Byers 1984: Tax credit raised to $0.60/gal University of Wyoming Honors College 1990: Tax credit lowered to $0.54/gal 2001: Tax credit lowered to $0.53/gal 2003: Tax credit lowered to $0.52/gal Abstract 2005: Renewable Fuel Standard established, requiring an average ethanol content of 10% in gasoline Corn is the largest and most widely produced agricultural 2011: Tax credit repealed commodity in the United States. What many people do not realize, though, is that there is an entrenched connection Each ethanol tax credit is an award given to ethanol blenders between the everyday practice of filling up one's automobile –wholesalers to fuel retailers— from the federal government. with gasoline and the price trend of this economic giant. That Prior to the institution of the first ethanol tax credit in 1978, connection is through ethanol. This study is observing several ethanol was seldom, if ever, included in consumer-grade policy changes on ethanol production incentives instituted by vehicle fuel blends. In fact, prior to 1981, data on corn ethanol the United States between 1978 and 2011. These observations production is so scant in comparison to the production of other are then applied to economic and econometric models in fuels and fuel substitutes that it is reasonable to suggest that which the projected changes deduced per economic theory ethanol’s production was limited to small consumer items and are tested against the statistical changes that occurred in the laboratory chemicals. The growth pattern of ethanol market for corn between 1970 and 2018. The two forms of production following the initial tax credit and its effect on corn policy of special focus are an ethanol tax credit, effectuated markets is the economic antecedent upon which this study is between 1978 and 2011, and an ethanol proportion mandate, predicated. enacted in 2005 and still in effect. Economic theory suggests that the tax credit will create multiple effects through an Research Expectations and Theoretical Models inward consumer demand shock in the retail fuel market and an outward producer derived demand shock, while a mandate Suppose in a suggested economy, where the behavior of the will prompt a uniform corn price increase. A regression model market is assumed to be the aggregation of individual market was employed to explain the time series data, and the results behavior from producers and consumers, the individual validated the economic theory: the tax credit was correlated consumer requires to travel a fixed number of miles. No matter with opposing price-moving forces between supply and what, the consumer must purchase enough fuel required to demand, and the mandate was correlated an overall price travel the fixed number of miles. For the sake of argument, increase. These focused results were found to be statistically suppose that the consumer must travel 10,000 miles in one significant. marketing year. In order to travel those 10,000 miles, the consumer must purchase a certain amount of fuel. The Introduction number of miles that a single unit of fuel will allow the consumer to travel is dependent on the efficiency of the fuel. Over the course of the last five decades, a plethora of United Suppose, for example, that one gallon of fuel that is comprised States industries from numerous economic sectors have of 100% gasoline allows the consumer to travel 25 miles. In this become increasingly dependent on corn-based products. The case, for the consumer to travel the 10,000 requisite miles in economic stimuli could be postulated as reducible to increased the marketing year, they would need to purchase 400 gallons technical efficiency in corn production coupled with of fuel. protectionist political measures directed to corn producers, Similarly, suppose that one gallon of fuel that is comprised but this is not the phenomenon in question in this study. What of 90% gasoline and 10% ethanol allows the consumer to travel is considered here, however, is that the market for consumer- 20 miles. The consumer then would need to purchase 500 grade vehicle fuel is a good that has not been excluded from gallons of this fuel in order to travel 10,000 miles. Let us call the permeation of corn-based products into its market the 100% gasoline fuel “E0” and the 90% gasoline, 10% ethanol constitution. In fact, the contemporary consumer would he fuel “E10”. In a graphical supply-demand representation, the hard pressed to find fuel at their local gasoline pump that did equilibrium price and quantity of E0 would be at some price not contain at least some level of corn-base ethanol in it, P*1 and quantity Q*1 = 400. (The graphical model below whether it be 10%, 15%, 85%, or some other percentage. illustrates the general principles of this postulation, from The market for consumer-grade fuel, however, was not which the tenets of this study’s theory and research always inclusive of corn-based ethanol. The United States’ expectations will unfold. Subsequent graphical models will be public policy on the incorporation of ethanol in fuel has employed throughout this section to support the general changed many times over the past fifty years. As a quick theory and inform the statistical model that this study will use enumeration of the policy adjustments that the market has to evaluate the research question.) experienced, here is a list of selected events: E0 Supply and Demand called Pe*1 and Qe*1. In a doubly derived market for corn, equilibrium price and quantity can be called Pc*1 and Qc*1. It P D is important to note that it is note that it is not the consumers E10 who are demanding either gasoline or ethanol, only fuel. It is fuel blenders who demand gasoline and ethanol and who then sell the fuel to fuel retailers who in turn sell to customers E0 (barring vertically integrated producers). Since consumers of fuel would prefer E0 to E10 because it minimizes costs, the quantity supplied of E10 is very low and the quantity supplied of E0 is very high. This is assuming that the cost of blending E0 is the same as the cost of blending E10. P*1 This would manifest itself in identical supply curves of consumer fuel, and since the demand curve for E0 is shifted outward as compared to E10 (since the consumer prefers E0), 0 Q*1 Q it is therefore advantageous to supply E0 than E10. This would create an inward shift in the supply curve of E10, further raising the equilibrium price of E10 and further encouraging If all fuel in the economy were to instantly change from E0 to consumer substitution to E0. This encourages a high demand E10, then in order to meet the 10,000 mile requirement, the for gasoline and a low demand for ethanol. The equilibrium demand curve would have to shift outwards until the new price of gasoline is therefore high, and the equilibrium price of equilibrium quantity, call it Q*2, is equal to 500. Since supply ethanol is low at their respective starting values. Since ethanol is held constant and the supply curve is upward-sloping, this is produced primarily from corn, the low demand for ethanol shift of the demand curve outward would create a new creates a low demand for corn (all else held constant) and a intersection of supply and demand at a higher point on the low price for corn at its starting values. price axis, call it P*2, so that P*2 > P*1. In addition to this, since The problem is perpetuated by the fact that the 500 > 400, Q*2 > Q*1. Therefore, (P*2 x Q*2 > P*1 x P*1), assumption of equal production costs is invalid. The relative meaning that the cost of using E10 is always higher to the inefficiency of ethanol with respect to gasoline is realized in consumer than the cost of using E0. If we assume that in production costs, making it, in terms of purchasing power, reaching the 10,000-mile requirement the consumer would more costly for blenders to produce E10 than E0. Assuming prefer to minimize costs, then it is preferable that the perfect competition, the supply curves of E10 and E0 are consumer purchase E0 as opposed to E10. assumed to be identical to the respective supply curves of fuel blenders, and these would be the supply curves faced by consumers. Therefore, the slope of the supply curve for E10 is E0 and E10 Supply and Demand noticeably steeper than the slope of the supply curve for E0, so that for every point of intersection of the fuel demand curve P D D(E10) with each supply curve, E10 will always be costlier to the E10 consumer than E0. Insert a tax credit for blenders that reimburses them a flat rate for every gallon of ethanol they incorporate into a retail P*2 E0 fuel blend. While the supply curve for ethanol producers does not change, this creates an outward shift in demand by fuel blenders for ethanol. While this drives up the equilibrium price for ethanol, it does not drive up the price realized by blenders because a flat value of each gallon of ethanol is essentially paid P*1 for by the government in the form of a tax credit. The result is that fuel blenders will supply more E10 to fuel retailers and less E0 despite consumer preference for E0. This phenomenon 0 Q*1 Q*2 Q takes place at the marginal level, however, by making it less costly to fuel blenders to supply E10 on a per unit basis since the tax credit is also on a per unit basis. Therefore, the slope If the consumer abides by their preferences, they will purchase of E10, though it was previously steeper than the slope of E0, more gallons of E0 and fewer gallons of E10. In the factor will become more and more shallow as the tax credit increases market for gasoline, the equilibrium price and quantity can be until it has an identical slope to the supply curve of E0, making called Pg*1 and Qg*1, respectively. Likewise, in the factor fuel blenders indifferent to whether they supply E0 or E10. market for ethanol, the equilibrium price and quantity can be E0 and E10 Supply and Demand corn. Because of the tax credit, a supply decrease of the same magnitude (if any) is not required for the equilibrium price of P D D(E10) corn to be reached. The increase of the consumption of E10 ultimately increases the gross revenues generated from corn production, but the costs are deferred to the government (taxpayer) and to fuel consumers. E0; E10 Now that the basic premise of corn price behavior in response to ethanol incentivization is understood, the responsiveness of corn prices can be further deduced through P*2 the relaxation of the assumption of fixed miles travelled. The demand for consumer-grade vehicle fuel under the relaxed P*1 assumption becomes dependent not on the exogenous constraint, but on the equilibration of fuel supply and demand. Relaxing the mileage constraint suggests that demand will react to the changes in supplier behavior. With respect to the 0 Q*1 Q*2 Q two demand curves under the constraint, the notion that the consumer will prefer E0 to E10 on account of its greater fuel Fuel blenders will inevitably diversify supply from purely E0 efficiency remains. In this case, it can be estimated that under these conditions because total revenue is bound to consumption behavior for E0 will move in a direction opposite increase by moving the quantity demanded for fuel towards a to that of E10. Essentially, two separate markets for fuel are point at a higher price level for each fuel so that revenues are created after the enactment of the 1978 ethanol tax credit, maximized with respect to marginal costs. If fuel economy one in which consumer demand shifts outward to raise prices were identical between E0 and E10, this would be a point (E0) and one in which demand shifts inward to lower prices where proportionately adjusted quantity demanded and price (E10). created a perfect square. Since this is not the case, and it is equally preferable for fuel blenders to supply E10 as E0, they E0 and E10 Supply and Demand can take advantage of the 10,000-mile consumer requirement P D(E10) D(E0) in making their supply decisions. Although E10 may have an identical supply curve to E0, because of the lower fuel economy of E10 compared to E0, consumers will have to purchase a higher quantity of E10 than E0; E10 they did E0 in order to meet the requirement of 10,000 travelled miles. This creates an outward shift in the consumer demand curve for fuel so that consumers purchase a higher P*1 quantity of E10 than they would have of E0 to travel 10,000 miles. The magnitude of the shift must be such that the P*2 intersection of the new demand curve with the E10 supply curve without the tax credit is at a quantity Q greater than the quantity Q*1 created by the intersection of the E0 supply curve and the original demand curve in order to emphasize the 0 Q*2 Q*1 Q consumer mile requirement. The new equilibrium price and quantity is at the intersection of the E10 supply curve with the tax credit (equal to the E0 supply curve) and the new, Whereas correlation between the prices of ethanol, corn, and outwardly shifted demand curve. At this point, the consumer gasoline may have incidentally correlated prior to the is purchasing at both a higher price and quantity of fuel than enactment of the tax credit, the model suggests that this they would have if all marketed fuel were of the E0 blend. To correlation decreases after the tax credit. While the magnitude employ the idiom, the tax credit would be “forcing the hand” of the tax credit fluctuates over time between 1978 and 2011, of consumers so that they would have to expend a larger it is the effect of the tax credit’s existence that is of primary proportion of their disposable income on fuel in order to travel import to this study. 10,000 miles in a marketing year. The enactment of the Renewable Fuel Standard in 2005, The increased demand for ethanol is also realized as an however, suggests a different effect in the behavior of fuel outward shift in the demand curve for corn, causing corn prices supply and demand than that of the tax credit, and thus a to increase. If the equilibrium price for corn in a market different effect on the price of corn. With the advent of the without a tax credit for ethanol was below marginal cost, ethanol mandate per the RFS, the new legal requirement was supply would have decreased in order to raise the price of that retail fuel generally have a minimum proportion of fuel be comprised of ethanol. This requirement precludes any widespread development of alternative fuel markets to those This change strengthens the correlation between the price of markets which have an ethanol blend. Since the mandate’s corn and the price of gasoline after 2011 because the indiscriminate applicability is with respect to suppliers of fuel counteracting effect is removed. and not consumers, the costs associated with the transfer of moderate fuel volumes containing ethanol to nearly all fuel Methodology volumes containing ethanol are incurred by suppliers. This shifts upward the supply curve for E10 while all but abolishing To describe the counteracting phenomena from the tax credit the supply curve for E0. Without E0, demand for E10 restores and the Renewable Fuel Standard postulated by the to a level proportional to its relative fuel economy to E0. theoretical model, a simple linear regression model approach was employed. While it may be altered in future study to capture the exact effects of each of the different forms of E0 and E10 Supply and Demand ethanol policy, functional form was left relatively unchanged from its default arrangement of level variables and P D corresponding parameters. The sign, comparative magnitude, E10 and statistical significance of each of the variables of special interest was of the immediate essence here rather than the values of the estimators themselves in succinctly describing E0 the precise effect of ethanol policy decisions on corn prices. P*2 To inform the statistical model as to what specific variables of interest would best aid the definition of explanatory factors, the changes in both corn and gasoline prices were compared P*1 throughout several time intervals, specifically those intervals where the arrangement of public policy on ethanol production was altered. Their correlations are enumerated in the table below: 0 Q*2 Q*1 Q Year Range Level Log 1970-1978 0.905621 0.870111 1979-2005 0.171992 0.204697 The increase in the price of fuel as a result of the RFS is met 2006-2011 0.829768 0.755167 with a proportional increase in the demand for ethanol, which 2012-2017 0.857493 0.863973 in turn proportionally increases demand for corn. Therefore, under the RFS, the price of corn can be interpreted as being In addition to the several variables of special interest codependent with the price of gasoline, experiencing a general intended to represent changes in ethanol policy, several increase after 2005, assuming that the RFS effect dominates controlling variables were included to accommodate for other the tax credit effect. With the repealing of the tax credit in major factors predicted to be instrumental in the 2011, the supplier price support is taken away, so a reverse determination of the price of corn. These variables are not effect of what happened to the supply curve for fuel after the exhaustive and were included specifically to help shed light on tax credit is expectable. the potential effects of the variables concerning ethanol policy. These variables account for average national gasoline prices, real gross domestic product, an agricultural E0 and E10 Supply and Demand productivity index, and public and private research and development investments into the agricultural sector across P D E10(’11) the time period from 1970 to 2014. E10(’05) The variables intended to account for the effect of ethanol P*2 policies on corn prices are multifold. The two different forms of ethanol policy, namely the tax credit and the mandate, are independent from one another and create their own separate P*1 effects, yet these effects take hold in the market for corn specifically when they interact with the price of gasoline, as suggested by the theoretical models. Therefore, two of the variables of special interest are an interaction term between a time dummy for the ethanol tax credit and the price of gasoline and an interaction term between a time dummy for the Renewable Fuel Standard and the price of gasoline. The inward shift in consumer demand for E10 per the theoretical model 0 Q*2 Q*1 Q also is an effect that takes hold because of the existence of the tax credit itself, so in addition to the interaction term, a time dummy accounting for the time during which the ethanol tax where: credit was in effect was also included. Finally, the high correlation between gasoline prices and corn prices in the year 𝑃𝑐 = corn price range of 1970 to 1978 in the immediately preceding table 𝛼 = constant suggested that a placeholder time dummy interacted with the 𝑃 = gasoline price (parameter 𝛽) price of gasoline also be included to account for a time period 𝑌 = real GDP (parameter 𝛾) of economic tumult and tenuousness that marked the decade 𝐸 = agricultural productivity index (parameter 𝛿) of the 1970’s, where because of periodic oil shocks, the prices 𝐼 = R&D investments (parameter 𝜃) of commodities not inherently correlated like corn and 𝑇 = tax credit dummy (parameter 𝜀) gasoline may have been each correlated with the price of oil. 𝐷 ∗ 𝑃 = oil shock dummy & gas price (parameter 𝜇) With the obvious inclusion of a variable for the price of corn 𝑇 ∗ 𝑃 = tax credit dummy & gas price (parameter 𝜋) itself, the variables were organized into a simple linear 𝑀 ∗ 𝑃 = RFS dummy & gas price (parameter 𝜌) regression format: Where the sample regression model is concerned, an error 𝑃𝑐 = 𝛼 + 𝛽𝑃 + 𝛾𝑌 + 𝛿𝐸 + 𝜃𝐼 + 𝜀𝑇 + 𝜇(𝐷 ∗ 𝑃) + 𝜋(𝑇 ∗ 𝑃) + 𝜌(𝑀 ∗ 𝑃) term 𝑒 was also included. Results The tabular results of this study have been parceled out into two illustrations— one with all of the data from the regression and one with selected data considered most relevant to the question. Complete Regression Results Regression Statistics Multiple R 0.957045291 R Square 0.915935689 Adjusted R 0.897254731 Square Standard Error 0.446790216 Observations 45 ANOVA df SS MS F Significance F Regression 8 78.30024684 9.787530855 49.0304452 4.55787E-17 Residual 36 7.186373883 0.199621497 Total 44 85.48662072 Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Lower 95.0% Upper 95.0% Intercept -1.894106199 0.761353645 -2.487813923 0.01762011 -3.43820296 -0.3500094 -3.43820296 -0.350009438 Gas Price (USD) 1.305594245 0.309955456 4.212199595 0.00016156 0.676975445 1.93421305 0.676975445 1.934213045 rGDP ($bil) 9.46748E-05 5.00798E-05 1.890481239 0.06676443 -6.8916E-06 0.00019624 -6.8916E-06 0.000196241 Agricultural -0.905882846 1.024734909 -0.884016772 0.38255316 -2.98414157 1.17237588 -2.98414157 1.172375876 Productivity Index R&D Investment 0.01401323 0.117337691 0.1194265 0.90560145 -0.22395864 0.2519851 -0.22395864 0.251985096 ($bil) Oil*P -0.136648778 0.118406788 -1.154062028 0.25607884 -0.37678888 0.10349132 -0.37678888 0.103491319 Tax credit 2.247916856 0.984486483 2.283339481 0.02841643 0.251285725 4.24454799 0.251285725 4.244547986 dummy Tax Credit -0.980908544 0.320201622 -3.063409043 0.00412793 -1.63030753 -0.3315096 -1.63030753 -0.331509556 interaction RFS*Pg 0.251057056 0.140461997 1.78736642 0.08230013 -0.03381308 0.53592719 -0.03381308 0.535927189 Selected Regression Results Adjusted R Squared 0.897254731 Coefficients P-value Intercept -1.894106199 0.01762 Gas Price (USD) 1.305594245 0.000162 rGDP ($bil) 9.46748E-05 0.066764 Agr Productivity Index -0.905882846 0.382553 R&D Investment ($bil) 0.01401323 0.905601 Oil*P -0.136648778 0.256079 Tax credit dummy 2.247916856 0.028416 Tax Credit interaction -0.980908544 0.004128 RFS*Pg 0.251057056 0.0823 As for the controlling variables, there was statistically each of these terms statistically significant, but their signs and significant, positive explanatory power shown in gas prices and magnitudes reflected what the hypothesis anticipated and real GDP as expected. The agricultural productivity index, what the theoretical economic model explicated. Two reflecting increasing productive efficiency over time, was counteracting corn price effects resulted from the tax credit unsurprisingly negative. Its statistical significance was tenuous and a general corn price boost from the Renewable Fuel and was not reflective of a satisfactory degree of confidence in Standard. the variable’s explanatory power. The R&D investment control variable, however, was not statistically significant, suggesting Discussion that, at least during the observed time period, the study did not detect a conclusive effect of agricultural research and There are two key conclusions to be drawn from this study. The development investment efforts on corn prices. first is that an ethanol tax credit has a two-sided effect on corn As for the variables of specific interest, a counteracting prices. While by way of fuel suppliers an ethanol tax credit effect was found between the tax credit binary variable and raises derived demand for corn and therefore its price, by way the tax credit interaction term, just as was suspected from the of fuel final fuel consumers the demand for fuel shifts, and a theoretical model. Each of those variables was overwhelmingly cap manifests on the degree to which ethanol’s inclusion in statistically significant, up through a 95% level of confidence. consumer-grade fuel blends impacts the corn market. This cap The RFS interaction term was significant for 90% confidence is contingent on trends in fuel consumption, such that ethanol and showed a positive relationship between the instatement demand increases as fuel consumption increases. As a of the Renewable Fuel Standard and corn prices. The adjusted conclusory byproduct, the economic deduction from these R-squared further suggests that the estimators are highly counteracting effects is that instantaneous equilibrium explanatory of the data. . quantities of E0 and E10 likely exist in the market for Overall, the coefficients were reasonable in their consumer-grade fuel under an ethanol tax credit, suggesting magnitude with respect to the general expectations of the the same may be true in the derived market for corn holding regression as compared to the theoretical model, and each all else constant. showed its expected sign, and the regression’s estimators The overall result of the tax credit only creates modest were generally significant. The oil shock interaction term was growth in the real price of corn in the long run. The ethanol not significant with a satisfactory confidence level, but this fact mandate via the Renewable Fuel Standard, however, does not come a surprise that alters the interpretation of the consistently raises corn price over time in a way that data. The oil shocks that marked the 1970’s —which served as supersedes the floundering effect of the tax credit. Perhaps a placeholder phenomenon through a dummy variable the most certain and insightful extrapolation from this study, interaction term— though correlating gas price with corn price though, is that different economic approaches to ethanol through common price connections in the market for crude oil, policy, even if they have the same economic goals, can were not systematic in the market for corn. In this case, a produce different economic effects. longer time series would likely be needed to retest their significance. The three terms that were of especial note in this Acknowledgements study were the tax credit dummy variable, the tax credit and gasoline price interaction term, and the Renewable Fuel I would like to personally extend my thanks to Dr. Benjamin Standard and gasoline price interaction term. Not only were Rashford, Head of the Department of Agricultural and Applied Economics at the University of Wyoming. Dr. Rashford served as primary research advisor for this study and provided invaluable insights into the timely and methodical release of this document and its accompanying presentation. Furthermore, his mentorship was a resource of lasting impact, as I am convinced that the practical and personal knowledge he imparted will be instrumental in improving the quality of any and all research I have the privilege of conducting in the future. Sources Bryce, R. (2015). The Hidden Corn Ethanol Tax: How Much Does the Renewable Fuel Standard Cost Motorists? Manhattan Institute for Policy Research (32). https://media4.manhattan-institute.org/pdf/ib_32.pdf de Gorter, H., Drabik, D., Just, D. R. (2015). The Economics of Biofuel Policies: Impacts on Price Volatility in Grain and Oilseed Markets. Palgrave Studies in Agricultural Economics and Food Policy. Federal Reserve Bank of St. Louis. (2020). Real Gross Domestic Product. [Data set]. Federal Reserve Economic Database. https://fred.stlouisfed.org/series/GDPC1#0 Macrotrends. (2020). Corn Prices – 59 Year Historical Chart. [Data set]. Macrotrends LLC. https://www.macrotrends.net/2532/corn-prices-historical- chart-data Greene, D. L. (1978). Econometric analysis of the demand for gasoline at the state level. United States Department of Energy Office of Scientific and Technical Information. Retrieved from https://www.osti.gov/biblio/6713363 Li, Y., Miao, R., Khanna, M. (2018). Effects of Ethanol Plant Proximity and Crop Prices on Land-Use Change in the United States. American Journal of Agricultural Economics. Retrieved from https://academic.oup.com/ajae/article- abstract/101/2/467/5231793 McNay, A. D. (2008). An Estimation of the Demand for Gasoline in Montana, and Projections of Future Gasoline Consumption. Montana State University Department of Agricultural Economics and Economics. Retrieved from https://scholarworks.montana.edu/xmlui/bitstream/handle/ 1/1842/McNayA0808.pdf?sequence=1&isAllowed=y Peters, M., Langley, S., & Wescott, P. (2009). Agricultural Commodity Price Spikes in the 1970s and 1990s: Valuable Lessons for Today. United States Department of Agriculture Economic Research Service: Amber Waves. https://www.ers.usda.gov/amber- waves/2009/march/agricultural-commodity-price-spikes-in- the-1970s-and-1990s-valuable-lessons-for-today/ USDA Economic Research Service. (2020). Indices of farm output, input, and total factor productivity for the United States, 1948-2017. [Data set]. United States Department of Agriculture. https://www.ers.usda.gov/data- products/agricultural-productivity-in-the-us/agricultural- productivity-in-the-us/#National%20Tables USDA Economic Research Service. (2019). Agricultural Research Funding in the Public and Private Sectors. [Data set]. United States Department of Agriculture. https://www.ers.usda.gov/data-products/agricultural- research-funding-in-the-public-and-private-sectors/