You are on page 1of 18

Available online at www.sciencedirect.


Energy Economics 30 (2008) 2073 – 2090

Toward an optimal U.S. ethanol fuel subsidy
Dmitry Vedenov, Michael Wetzstein ⁎
Department of Agricultural and Applied Economics, University of Georgia, Athens, GA 30602, USA Received 26 June 2006; received in revised form 11 December 2006; accepted 12 February 2007 Available online 27 March 2007

Abstract Enhanced environmental quality, fuel security, and economic development along with reduced prices of ethanol–gasoline blends are often used as justifications for the U.S. federal excise tax exemption on ethanol fuels. However, the possible effect of increased overall consumption of fuel in response to lower total price, mitigating the environmental and fuel security benefits, are generally not considered. Taking this price response into account, the optimal U.S. ethanol subsidy is derived. Estimated values of the optimal subsidy reveal that the subsidy's environmental and security benefits are questionable. However, positive environmental and security benefits from the ethanol tax-exemption subsidy may be obtained if the subsidy is combined with an increase in the excise tax on gasoline. © 2007 Elsevier B.V. All rights reserved.
Keywords: Economic development; Energy security; Environmental quality; Ethanol; Excise tax; Renewable fuels; Tax exemption

The U.S. ethanol federal excise tax exemption on vehicle fuels has existed in controversy since its inception in 1978. Designed to provide incentives for the establishment and development of a U.S. renewable fuels industry, proponents have praised its lasting and expanding benefits while critics have condemned it as deadweight loss corporate welfare. Supporters – including the American Coalition for Ethanol and American Corn Growers Association – argue that the ethanol fuel tax exemption (subsidy) helps to improve air quality, leads to energy security, spurs economic development, reduces greenhouse gases, and lowers the overall price of vehicle fuels. For example, a study by Cooper (2003) states that consumers could save $0.08 per gallon if more ethanol were introduced into U.S. fuel markets. In contrast, Taxpayers for Common Sense argue there is only an artificial demand for ethanol solely based on governmental subsidies and the subsidies are a corporate bailout for agribusiness conglomerates. For a more detailed discussion on the merits of
⁎ Corresponding author. Tel.: +1 706 542 0758; fax: +1 706 542 0739. E-mail address: (M. Wetzstein). 0140-9883/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.eneco.2007.02.004

5 billion gallons in 2012. The optimal subsidy is then estimated given published parameter values. The results of estimation and sensitivity analysis are presented in the following two sections. This possible increase in petroleum fuel use. The last section discusses implications of our results and presents concluding comments. ethanol market. Vedenov. 2010.S. roads in 2005 (Lundegaard. the bulk of the ethanol fuel has been used in E10 fuels. Reduction in ethanol price due to subsidy then translates into lower price of blended fuels. Our objective is to derive the socially optimal U. as a side effect. This optimal subsidy is based on utility maximization which explicitly incorporates the effect that an ethanol subsidy not only stimulates the use of renewable fuel. only five million of these so-called flex cars were on U. but. Surprisingly there is little or no acknowledgment of this indirect price response effect. may then mitigate the other beneficial effects of the subsidy. Analyses behind both of these popular counter positions tend to miss the (somewhat unintended) effect of an indirect demand response to the ethanol subsidy. 1. The next section presents a brief overview of the U. M.2074 D. we analyze the hypothesis that this increased consumption of fossil fuels can mitigate and possibly completely offset the environmental and security benefits of the subsidy. Specifically.51 per gallon is extended through December 31. In particular. sets a new goal for expanding domestic fuel supplies with renewable fuels. the impact the elasticity of total fuel use with respect to the ethanol subsidy has on the environmental and security benefits of the ethanol subsidy should be investigated prior to reaching such a conclusion. Until recently. ethanol subsidy incorporating a comprehensive evaluation of the environmental. while eliminating the oxygenate requirement. followed by a discussion of parameter values used to evaluate the subsidy. such price responses should be taken into account. More importantly the ethanol tax incentive at $0. In particular. A theoretical model for the optimal ethanol subsidy is developed next. and economic development benefits. The ethanol–fuel market received a further boost from the 2005 Energy Bill which. the environmental benefits of improved air quality and reduced greenhouse gases along with fuel security benefits may be reduced or completely offset by the increased use of fossil fuels. due to an ethanol subsidy. ethanol . Wetzstein / Energy Economics 30 (2008) 2073–2090 and concerns with the ethanol subsidy. The remainder of the paper is organized as follows. the renewable fuels standard sets a national minimum usage requirement of four billion gallons in 2006 with a mandated increase to 7. also provides a price incentive for agents to increase their use of fossil fuels. Thus. the 2005 Energy Bill will boost the total market for ethanol by mandating minimum usage and provide a price subsidy for achieving this level of usage. which contain only 10% ethanol and 90% conventional gasoline. mainly ethanol and biodiesel. By far. The sensitivity of the optimal ethanol subsidy to elasticities and the marginal welfare gains from improvements in the environment and security are also analyzed.S. One reason for the lack of interest may be an assumption that the consequences are second-order small and thus of limited concern. much less an investigation of it. 2006). the reader may refer to the National Center for Policy Analysis (2005).S. Although vehicles that can use both conventional gasoline and the E85 fuel (blend of 85% ethanol and 15% petroleum gasoline) are commercially available. security. If the total effect of an ethanol subsidy is to be determined. causing higher demand for the blended fuels in general and their primary component – fossil fuels – in particular. However. This represents a fundamental shift in the pattern of ethanol fuel use. Current ethanol tax exemption provisions and benefits The force behind the recent rapid expansion of the ethanol–fuel market is the current phasingout of MTBE (methyl–tertiary–butyl ether) as a fuel oxygenate and its replacement with ethanol.

The 2005 Energy Bill removed this disproportionate impact by supplementing the trust fund for ethanol subsidy revenue losses. the ethanol subsidy does have some social costs. 2.D. Vedenov. As outlined by the Renewable Fuels Association's Ethanol Industry Outlook 2005 (2005b). this created a disproportionate impact across states given the uneven distribution of ethanol usage. and ρ are quasi-concave. Theoretical model As outlined by Parry and Small (2005) a number of previous studies have investigated the external costs of vehicle transportation in an attempt to develop an optimal gasoline tax. and the economy. . The other variables are miles of vehicle travel M. the ethanol subsidy still reduces government revenue which impacts overall governmental allocations. M. government highway spending G. Limited if any studies have attempted to model the optimal ethanol fuel subsidy. As addressed by Rask (2004). P. Parry and Small (2005) in their analysis identify a Pigovian and Ramsey tax along with congestion feedback as components of an optimal gasoline tax. a theoretical model for the optimal ethanol subsidy is developed. environmental degradation P. For a number of states the benefits of the ethanol subsidy stimulating economic development did not totally compensate for the loss in economic development from reduced highway funding. The functions u. Incorporating all these factors into a model for deriving the optimal ethanol subsidy will provide insights into how these benefits interact with the cost of the subsidy. In this regard. whereas δ is weakly convex representing the disutility from pollution. changes in ethanol prices now directly translate into changes in prices of ethanol blends thus affecting demand for blended fuels overall and fossil fuels in particular.S. After MTBE was banned as a fuel oxygenate by a majority of states. security. the government incentives have established a renewable fuels industry with production from homegrown resources. M Þ þ gðGÞ−dðPÞ þ qðAÞ. Using the past studies on deriving the optimal gasoline tax. ethanol has emerged as an oxygenate of choice and is rapidly becoming a complimentary good to conventional gasoline. as well as government cost of an ethanol subsidy are embedded in (1). Prior to the 2005 Energy Bill the Highway Trust Fund was reduced by the lost revenue from the ethanol subsidy. provide a greater understanding of the subsidy's impacts on the environment. The Outlook also indicates gasoline prices would increase over 14% without ethanol fuels.. E10 fuels reduce tailpipe fine particulate matter emissions and carbon monoxide and can reduce greenhouse gas emissions. The external benefits of reduced pollution (both local air quality and greenhouse gases) and increased fuel security. Variables G. This ensures a safe and reliable source of fuel which also generates economic and environmental benefits. However. Therefore. U. ð1Þ where X denotes a composite consumption good with associated numeraire price pX = 1. as a guide. γ.1 billion.000 jobs yielding a potential GNP increase of $25. assume a static model with many agents. and fuel security A.. in the U. so the agent perceives them as exogenous. 2006). and thus. associated with using a vehicle U ¼ uðX . Wetzstein / Energy Economics 30 (2008) 2073–2090 2075 was primarily a substitute to MTBE and changes in ethanol prices typically resulted in changing the balance between quantities of ethanol and MTBE demanded (Zhang et al. However. and A are features of the agent's environment. Let a representative agent have the quasi-linear utility function. These environmental and fuel security benefits are also accompanied by economic benefits with the creation 147. in particular Parry and Small (2005).

2076 D. pollution. is decomposed into air quality. which is reflected in (3′). As a consequence. H ¯ ¯ A ¼ Að F 0 . An agent then attempts to maximize utility (1) subject to the budget constraint X þ ð p−saR ÞF þ H ¼ IðsÞ. . A. PF. is produced according to a linear homogeneous function M ¼ M ðF. and government spending issues. As a federal gasoline tax exemption. MTBE (cf. M. the ethanol subsidy negatively affects the government highway budget by either decreasing funds available for highway development or requiring a redirection of funds from other governmental programs. A major component of fuel security is the possible reduced ¯ welfare loss from trade disruptions. with fuel blenders continuing to use ethanol even in the face of prices rising as high as $4. ð3Þ ¯ ¯ where ∂A/∂F 0 b 0 and ∂A/∂H N 0. the substitutability between ethanol and other gasoline additives is being rapidly replaced by a strong complementarity relationship between ethanol and conventional gasoline. Wetzstein / Energy Economics 30 (2008) 2073–2090 Following Parry and Small (2005) vehicle travel. agents ignore the effect of their own driving on fuel security. ð2Þ where F represents fuel consumption and H is money expenditure on fuel efficiency. It is assumed air quality depends on miles driven and greenhouse gases on fuel consumption ¯ ¯ P ¼ PM ð M Þ þ PF ½ F ð1−aR aq ފ. ð6Þ 1 Prior to 2005. liability issues with other gasoline additives and anti-backsliding provisions of the 2005 Energy Bill regarding air quality essentially left fuel blenders with ethanol as the only oxygenate alternative. and aggregate ¯ fuel efficiency. P. and the per-gallon subsidy s reduces these tax collections ¯sαR. ¯ dPF =d F N0. The federal gasoline tax is used for funding the governmental expenditures on highway development and improvements along with mass transit systems. PM. Vedenov. F 0. ¯ dPM =dM N0. 2003.g. However. Gallagher et al. ð4Þ ¯ where M represents aggregate miles driven by all agents and αq denotes the percentage reduction in greenhouse gases from renewable fuels.. improved vehicle fuel efficiency leads to a higher sticker price. ethanol as an oxygenate was primarily a substitute for other gasoline additives e. HÞ.50 in 2006 (McKay). M. 2006). H Þ. The relation of F 0 to aggregate fuel consumption is ¯ ¯ F 0 ¼ ð1−aR Þ F . ð3′Þ ¯ where F represents aggregate fuel consumption and αR denotes the renewable fuel (ethanol) share in total fuel consumption. the total tax collections are F t. ð5Þ where t represents the excise tax rate and s denotes the per gallon ethanol subsidy. This assumes a tradeoff exists between vehicle cost and fuel efficiency.1 The environmental effect of driving. i. ¯ Fuel security. is based on the aggregate level of fossil fuel consumption. and greenhouse gas emissions. The demand for ethanol is now much more inelastic.e. by F Given the presence of externalities. With no ¯ subsidy. Formally ¯ G ¼ F ðt−saR Þ.

where the subscripts denote first partial derivatives of the respective functions. The latter. M. H/M.C. H environmental quality. M. 2. P. holding the fuel price constant. This total cost is the sum of the fuel cost per mile. HÞ ¼ max u½X . the Euler theorem states M ¼ MF F þ MH H: ð9Þ Substituting (8) into (9) and solving for the agent's monetary benefit of driving per mile. sÞ: . AL=AF ¼ uM MF −kðp−saR Þ ¼ 0. uM/λ. uM/λ. p − sαR is the effective price per gallon of fuel paid by the agent. Welfare effects As noted by Parry and Small (2005). aggregate fuel consumption. Then from (10) pM ¼ pM ðsÞ.D. where λ is the Lagrange multiplier. G. (10) can be interpreted in the sense that agents equate their monetary marginal benefit of driving per mile. F. agents ignore their own impact on fuel security. H).e. 2. and H ¼ HðsÞ ¼ aHM ðsÞM ð pM . in turn. G. and A then become the model's parameters along with p which is suppressed because it is not varied with s. Expenditures on consumption. is influenced by the level of subsidy. pM.O. F . Thus. The terms s. yields uM ð p−saR ÞF H þ ¼ pM : ¼ M M k ð10Þ Eq. Hފ þgðGÞ−dðPÞ þ qðAÞ þ k½IðsÞ−X −ð p−saR ÞF−HŠ. Wetzstein / Energy Economics 30 (2008) 2073–2090 2077 where p represents the total price per gallon of fuel. P. Agent's cost per-mile The optimal subsidy is determined from the indirect utility function V ðs. sÞ. M ðF. fuel. ð7Þ obtained by maximizing (1) subject to (6). the ratios of variables become the functions of the subsidy only. since the subsidy promotes economic development which enhances the agent's income. and I is disposable income. M ¼ Mð pM . with the total per mile cost of driving.2. AÞ ¼ max LðX . ¯. Vedenov. αFM = F/M = αFM(s) and αHM = H/M = αHM(s). From these MF k ¼ ¼ MH : p−saR uM ð8Þ With the vehicle travel function (2) specified as linearly homogeneous. AL=AH ¼ uM MH −k ¼ 0. (p − sαR)F/M and non-fuel costs per mile. sÞ.1. and aggregate fuel efficiency. The F. due to the homogeneity property of M(F. and non-fuel driving cost are equal to the agent's income. i. The determination of the agent's cost per mile (10) is based on the assumption that agents ignore the effect of their own ¯ ¯ driving on aggregate mileage. and treat those as constant. these impacts should be taken into account in determining the social welfare effects of the ethanol subsidy. F ¼ FðsÞ ¼ aFM ðsÞM ð pM . However.s for (7) are AL=AX ¼ ux −k ¼ 0. and government spending.

from the definitions of G. Vedenov. the renewable fuel share. so their partials with respect to s are the partials of M. aggregate mileage. and fuel efficiency.2078 D. fuel ¯ ¯ consumption. ∂V/∂P = − δ′ b 0. and hence ∂M/∂s N 0 and dH/ds b 0 (cf. The indirect effects are further decomposed into the effect a subsidy has on the externalities of fuel consumption. Substituting (12a) − (12c) into (11) and dividing by λ results in the marginal monetary welfare effect of the ethanol subsidy s: ðdV =dsÞ=k ¼ ðg V =kÞf−FaR þ ðdF=dsÞðt−saR Þ−ðdaR =dsÞsFg −ðdV=kÞfðdPF =dFÞ½ðdF=dsÞð1−aR aq Þ−ðdaR =dsÞaq FŠ þ ðdPM =dM ÞðAM=AsÞg þ ðqV=kÞfðAA=AFÞ½ðdF=dsÞð1−aR Þ−ðdaR =dsÞFŠ þ ðAA=AHÞðdH=dsÞg þ dI=ds þ aR F þ ðdaR =dsÞsF: ð13aÞ Following Parry and Small (2005). The direct effect of a change in the subsidy is the change in fuel expenditures. respectively. F. and rewrite (13a) as ðdV =dsÞ=k ¼ aR F þ ðdaR =dsÞsF−ðg V=kÞFaR þ ½ðg V=kÞðt−saR Þ−E PF F þ E AF ŠðdF=dsÞ þ ½−ðg V=kÞsF þ E PF a −E Aa ŠðdaR =dsÞ −E PM ðAM =AsÞ þ E AH ðdH=dsÞ þ dI=ds: An interpretation of (13b) is that the marginal welfare effect of ethanol subsidy is decomposed into the direct and indirect effects. dP=d s ¼ ðdPF =d FÞ½ðdF=d sÞð1−aR aq Þ−ðdaR =dsÞaq FŠ þ ðdPM =dM ÞðAM =AsÞ. (4). Noting that ∂V/∂s = λ[dI/ds + αRF + (dαR/ds)sF] (by the envelope theorem). and A in (5). the social welfare effects. F . we define the externality effects as E PF F ¼ ðdV=kÞðdPF =dFÞð1−aR aq ÞN0. Wetzstein / Energy Economics 30 (2008) 2073–2090 The welfare effects of an incremental change in the ethanol subsidy may then be determined by totally differentiating the indirect utility function (7) with respect to the subsidy level s. E PF a ¼ ðd V=kÞðdPF =dFÞaq FN0. ∂V/∂A = ρ′ N 0. yields dV =d s ¼ gVdG=d s−dVd P=ds þ q VdA=ds þ k½dI=ds þ aR F þ ðdaR =dsÞsFŠ: Next. H are no longer constant. . E ¼ ðq V=kÞðAA=AFÞFb0. (9)). E PM ¼ ðd V=kÞðdPM =dM ÞN0. and E AH ¼ ðq V=kÞðAA=AHÞN0. dA=ds ¼ ðAA=AFÞ½ðd F=dsÞð1−aR Þ−ðdaR =dsÞFŠ þ ðAA=AHÞðd H=dsÞ: ð12aÞ ð12bÞ ð12cÞ ð11Þ ¯ In determining (12a) − (12c). miles driven. and fuel efficiency. dF/ds N 0 (cf. dG=d s ¼ −FaR þ ðd F=dsÞðt−saR Þ−ðdaR =d sÞsF. M. M . and H. P. and ∂V/∂G = γ′ N 0. (10)). and (3). Since higher levels of subsidy result in lower overall prices and thus higher consumption of fuel. along with the income effect. AF Aa E ¼ ðq V=kÞðAA=AFÞð1−aR Þb0.

Fs. αFH. and again negatively affecting MEB. αFM. The indirect marginal benefits are changes in air quality. respectively. so that MEB b 0 if ξ b 0. it must provide a sufficient positive stimulus toward enhancing the share of ethanol in overall fuel consumption. In contrast. for the ethanol subsidy to result in a net positive MEB. Vedenov. Wetzstein / Energy Economics 30 (2008) 2073–2090 2079 2. Define MEB as the net marginal external benefit of renewable fuel use MEB ¼ ð−E PF F þ E AF Þ−EPM b=aFM þ E AH f=aFH þ ðE PF a −E Aa Þn=aFa . Recall that the first three terms in (14) are unambiguously negative. These and other properties of MEB are summarized in the following propositions. in turn. fuel efficiency. (EPF α − EAα)ξ/αFα N 0. which. ðdF=dsÞH Fs ðdaR =dsÞF as n¼ ¼ N0. Recall that αR = e/F. MEB can be positively augmented if the subsidy increases the share of renewable fuel in the total fuel consumption. The proof of this proposition is based on the discussion above. and renewable fuel share with respect to the subsidy. M. EAHζ/αFH. where the parameters β. creating a disincentive to invest in fuel efficiency (EAHζ/αFH b 0). where e is the quantity of ethanol used.D. ξ b 0 if and only if αs b 0. and αs denote the elasticities of mileage. ð14Þ and Ms. The subsidy will also provide a positive incentive to increase miles traveled (∂M/∂s N 0) thus reducing air quality (− EPMβ/αFM b 0). if ξ N 0. ξ. An increase in the subsidy will stimulate additional fuel consumption (dF/ds N 0). Taking the derivative of αR with respect to s yields daR ðde=dsÞF−ðdF=dsÞe ¼ : F2 ds . Hs. fuel. it is more convenient to express the marginal welfare effects (13b) in terms of elasticities. ζ.3. aFH ¼ F=H. Therefore. fuel efficiency. per unit change in fuel consumption. will add to the greenhouse gas effect (− EPFF b 0) and decrease fuel security (EAF b 0) resulting in lower MEB. and αFα are defined as b¼ ðAM =AsÞF Ms ¼ N0. (EPF α − EAα)ξ/αFα. Since Fs N 0. EPMβ/αFM. − EPF + EAF. ðdF=dsÞM Fs ðdH=dsÞF Hs ¼ f¼ b0. aFa ¼ F=aR . Marginal external benefits For further analysis and interpretation. The MEB will be negative if the quantity of ethanol used is less responsive to a subsidy than the overall fuel consumption. The direct marginal benefits are the effects of fuel use on greenhouse gas emission − EPFF and fuel security EAF. F MEB is composed of the direct benefits of fuel use. ðdF=dsÞaR Fs aFM ¼ F=M . and renewable fuel share. and the sign of the last term is determined by the sign of ξ = αs/Fs. Proposition 1. From (14) an increase in renewable fuel share will retard greenhouse gases emission and enhance fuel security. and indirect net external marginal benefits from a per unit change in fuel consumption.

Since AMEB ðE PF a −E Aa Þn ¼ N0. then αs b 0 and thus MEB b 0. is to an ethanol subsidy. M. αR. i. the higher is the MEB. and αs. Vedenov.e.2080 D. Hs. AHs aFH Hs ð16Þ ð15Þ higher responsiveness of mileage or fuel efficiency to the ethanol subsidy implies lower MEB.e. Since AMEB E PM b ¼− b0. the positive benefits of renewable fuels are more than offset by negative externalities caused by the increased use of fossil fuel associated with the increase in overall fuel consumption. For the ethanol subsidy to result in positive net marginal external benefits. ¼ ds aR ds e ds F i. The proposition can be proved directly by taking the first partial derivatives of (14) with respect to the elasticities Ms. Proposition 2. M. If the subsidy stimulates overall fuel consumption more than the use of ethanol. to the subsidy results in higher MEB. s. a level of subsidy that increases overall fuel consumption by 1% will result in less than a 1% increase in mileage driven. The derivative of (9) with respect to s is dM AM dF AM dH ¼ þ : ds AF ds AH ds ð18Þ . yields as ¼ daR s de s dF s − ¼ es −Fs . The most important consequence of Proposition 1 is that higher ethanol subsidies do not necessarily result in positive net welfare benefits. Corollary. is to s. Proof of Proposition 3 is based on the linear homogeneity of M = M(F. AMs aFM Ms and AMEB E AH f ¼ b0. Proposition 3. if the quantity of ethanol is more inelastic than the overall fuel consumption. The more responsive mileage. Hence. or fuel efficiency. The more responsive the renewable fuel share. H. the lower is the MEB. it is necessary that the quantity of ethanol used is more responsive to the subsidy than the total fuel consumption. the elasticity of renewable fuel share is equal to the difference between the elasticities of ethanol quantity and total fuel consumption. Wetzstein / Energy Economics 30 (2008) 2073–2090 Dividing both sides by αR and multiplying through by s. β = Ms/Fs b 1. Directly from the proof to Proposition 1 is the following corollary. H). αR. Aas aFa as ð17Þ a higher positive response of renewable fuel share.

M. times the percentage change in the renewable fuel share. yields 0 ¼ MEB þ aR s=Fs þ ðIs =Fs Þ=aFI þ ðas =Fs ÞsaR þ ðg V=kÞGs ðt−saR Þ=Fs : Solving for s. the optimal ethanol subsidy per gallon may now be restated as s⁎ ¼ MEBFs þ ðIs =aFI Þ þ ðgV=kÞtGs : ½ðg V=kÞGs −ð1 þ as ފaR ð23Þ . we have Ms bMF b1: Fs ð19Þ The last inequality follows from the linear homogeneous nature of M = M(F. Vedenov.4. H) resulting in MF + MH = 1 which implies MF = 1 − MH b 1. Proposition 3 reflects the fact that an increase in miles driven due to higher fuel consumption is partially offset by decreasing investment in fuel efficiency caused by higher ethanol subsidies. Optimal ethanol subsidy Setting first-order condition (13b) to zero and dividing by dF/ds yields 0 ¼ MEB þ ½1−ðgV=kފaR s=Fs þ ðIs =Fs Þ=aFI þ ½1−ðg V=kފðas =Fs ÞsaR þ ðg V =kÞðt−saR Þ.D. 2. Wetzstein / Energy Economics 30 (2008) 2073–2090 2081 Multiplying through by s/M and multiplying the first and second terms on the right-hand side of (18) by F/F and H/H. Proposition 4. sαR/(t − sαR). ð20Þ where Is = (∂I/∂s)(s/I) and αFI = F/I. respectively. The elasticity of government spending with respect to an ethanol subsidy. Rearranging (20) and substituting in Gs. Solving for s then yields the optimal ethanol subsidy per gallon of ethanol s⁎ ¼ MEBFs þ ðIs =aFI Þ þ ðg V =kÞtFs : ½ðgV=kÞðFs þ 1Þ−1−as þ ðg V=kÞas ŠaR ð21Þ As an aid in interpreting (21) consider the following proposition. Gs is Gs ¼ Fs −ð1 þ as Þ saR : t−saR ð22Þ This proposition may be proved by dividing (12a) through by G = F(t − sαR) dG 1 dF 1 aR daR s ¼ − − : ds G ds F t−saR ds t−saR Multiplying through by s and the last term by αR/αR yields (22). The elasticity of government spending is equal to the elasticity of fuel consumption minus the percentage of tax revenue loss due to the subsidy. yields the elasticity relation Ms ¼ MF Fs þ MH Hs : Since Hs b 0.

e. 134. occurs on both sides of the formula in (23) because it implicitly enters the elasticity of government spending with respect to an ethanol subsidy. and observed data are directly applied in (21). F. Recall that αR is the proportion of ethanol fuel consumption. Fuel efficiency. 2003). Wetzstein / Energy Economics 30 (2008) 2073–2090 The optimal ethanol subsidy. and dividing it by the total fuel consumed.75 billion total gallons of ethanol used.184 per gallon (Williams. 2005). Parameter values Benchmark values and parameter ranges used for estimating the optimal ethanol subsidy (21) are summarized in Table 1. is the sum of three parts.2082 D. times the responsiveness of this fuel consumption to a percentage change in the subsidy. (22)).S. currently set to $0. Note that the subsidy. 1/αFM = 22 miles per gallon. 2 .2 The resulting αR = 0. Vedenov. (γ′/λ)Gs less the percent increase in the renewable fuel share. 1 + αs. s⁎. The more elastic is income with respect to subsidy. 2004). Also. MEB.S.1 billion gallons. Dividing the Pigovian and Ramsey subsidies along with the government margin welfare by this proportion yields the subsidy per gallon of renewable fuel. Similar to a Ramsey tax. 3. Wang (2005) in a review of the literature on greenhouse gas emissions from corn ethanol relative to gasoline indicates most studies estimate greenhouse gas emission reductions of around 20%. The proportion of ethanol fuel used was determined from U. s. given (21) depends on parameter values at the social optimum and any observed values apply to the existing.51 per gallon of ethanol (RFA. from a change in the subsidy. was set at 550 gal per vehicle (U. passenger car fuel efficiency for 2000–2001 (US Department of Transportation.S. is based on the average U. with the range from 15 mpg to 30 mpg. s. s⁎. the higher the optimal level of the subsidy. The annual gallons of fuel consumed. and the government marginal benefits. Gs (cf. While 1 gal of ethanol contains only 2/3 of energy contained in a gallon of gasoline. The RFA has an interesting historical comparison of how this federal excise tax interacts with the ethanol tax exemption. FHWA. as noted by Parry and Small (2005).02 was used as a benchmark with the range set from 0.10. the reduced formula (21) is better suited for estimating the optimal ethanol subsidy. it is assumed the elasticities are constant. in this case Is. the Pigovian subsidy MEBFs. For our analysis. Considering the possible use of nitrogen fertilizer applications in the production of plant material for ethanol refining. M. (∂I/∂s)s. the optimal formula is only second-best. The federal excise tax on gasoline is set $0. While (23) is more convenient for interpretation. FHWA (2003) using 2003 estimates of 2. per gallon of fuel consumption. This ethanol–fuel proportion is augmented by the incremental loss of tax revenue from the subsidy. The Ramsey subsidy is the change in total income. (γ′/λ)tGs. While Wang's results indicate corn-based ethanol achieved moderate reductions in fossil energy use and greenhouse gas emissions. 2005).02 to 0. The Pigovian subsidy is the net external marginal benefits from a per unit change in fuel consumption. i. possibly non-optimal. The government marginal benefits are the marginal welfare effects times the tax rate times percentage change in revenue in response to percentage change in the subsidy. cellulosic ethanol with biomass feedstocks from fast-growing trees and switchgrass can have This assumes equivalent miles per gallon with ethanol-blended fuels compared with non-ethanol-blended fuels. Ramsey subsidy Is/αFI. the lower BTU content of ethanol blends is mitigated by more complete burning of the base fuel due to oxygenate properties of ethanol (American Coalition for Ethanol. 2004). [(γ′/λ)Gs − (1 + αs)]αR. the optimal subsidy depends on the elasticities. elasticity Fs. the process of ethanol production mitigates greenhouse gas emission reduction from using ethanol-blended fuels. The Pigovian and Ramsey subsidies along with the government marginal benefits are all discounted by the per dollar of subsidy change in welfare. ethanol-subsidy equilibrium.

1984 to 1987 and 1988 to 1993 and arrived at elasticity estimates of − 2. 3 Model estimates of this Tobit 2SLS are available from the authors. Fs Ethanol price. ep = − es. F (gallons/vehicle) Fuel efficiency. For the analysis. s (dollars/gallon) Fuel. Khazzoom (1997) surveyed the literature and determined the elasticities to be relatively small in the range of − 0.35 0.004 to 0.4 − 0.02 to 0. Parry and Small (2005) surveyed the extensive literature on gasoline and mileage own-price elasticities of demand. very limited analysis has been conducted on estimating the price elasticity of demand for ethanol. t (dollars/gallon) Ethanol subsidy. ζ Renewable fuel share/fuel.87 × 10− 3 0.02 0.08.473 1. αFH Fuel/proportion of renewable. Rask (1998) estimated the ethanol price elasticity for two time intervals.20 4. Hs. Since then.81.60 − 95.9 to − 0.3 and β = Ms/Fs = 0. Choosing − 0.55 0.1 0. Vedenov. EPFF (dollars/gallon) Fuel security.361 2.82 and − 0.68 × 10− 3 0.00 − 114. EPF (dollars) Fuel security. ξ Government welfare effect.059 − 0.51 550 22 0.09. EAH Renewable fuel proportion α Greenhouse gases. Is Mileage/fuel.3 This estimated elasticity is used as the benchmark with Rask's estimates as the range.870 27.0 to − 78.02 0. 1/αFM (miles/gallon) Elasticities Gasoline price. In contrast.7 to − 0. In terms of the elasticity estimates. limited if any empirical estimates were undertaken.59 to 8.1 as the benchmark results in ζ = Hs/Fs = − 0.169 0. M.18 0.37 to 2. αFI Benchmark value 0.2 to 0.3 to 0.09 − 0. was found to be − 0. αR Proportion of greenhouse gases. Using a Tobit 2SLS procedure for estimating the demand and supply of ethanol fuel.21 to − 0. mainly due to severe data limitations on quantities of ethanol demand and supply.250 to 0.9 0.0 much larger energy and greenhouse gas emission reduction benefits (Argonne.40 1 to 1.6.81 1.20 to 132. β Fuel efficiency/fuel. EAα (dollars) Ratios Fuel/engine efficiency.2 to 1 2083 15 to 30 0.82 0 to 8.4 with a range of 0. For the elasticity of vehicle fuel efficiency.018 to 0. EPM (dollars/mile) Engine efficiency.20 to 1 with 0. the range of αq is set from 0. ep. es Income response to subsidy. αFα Fuel/income. EAF (dollars/gallon) Air quality.55 with a range of − 0.37.10 Range 0.7 to −0.6 − 0.10 0.303 6.18 with a range of − 0. Wetzstein / Energy Economics 30 (2008) 2073–2090 Table 1 Parameter values Parameter Proportion of renewable fuel. This limited analysis prompted an empirical investigation of the ethanol market by the authors using a data set consisting of 765 observations cross-sectioned by state from 1988 through 2002.184 0.202 to − 0. the price elasticity of ethanol.126 0. γ′/λ Externality effects Fuel consumption Greenhouse gases. 2004). αq Excise tax.2 to 0. From this literature they adopted Fp = − Fs = − 0.0158 0.239 − 0.500 0.D.2 0.20 as the benchmark. .

for the next decade such as incremental military and other government resources allocated to Iraq. the elasticity of income with respect to the ethanol subsidy is determined by first calculating Ie and es.5 trillion (US.24 per gallon with a central value of $0. 2005a). but does not consider the impact on corn production. Rask (2004) investigates the effect ethanol subsidies have on the Highway Trust Fund and finds significant differences by state. Using these estimates results in EPF F ¼ ðd V=kÞðdPF =dFÞð1−aR aq Þ ¼ 0:06½1−0:02ð0:6ފ ¼ $0:059 and EPF a ¼ ðdV=kÞðdPF =dFÞaq F ¼ ð0:06Þð0:2Þð550Þ ¼ $6:60: In calculating the effect of driving on air quality. M.06. As noted by Parry and Small (2005). Department of Commerce.77 multiplier for food manufacturing.0 billion. and the costs of treating wounded soldiers. Their estimates include the direct economic cost to the U. the personal income. the regional effects of this subsidy can be substantial. Combining these estimates yields Ie = 2.5 billion lower if ethanol demand and production were zero.8 billion from ethanol plants.1. of $5. This includes using a 2. Associated with these costs are benefits (avoided costs) which include no longer enforcing U. the limited science of atmospheric dynamics.77 per gallon of ethanol. Considering Ie. With the benchmark level of income and given the production of ethanol in 1997 was 1.004 to 0. dI/ds. and the level of disposable personal income for the first quarter of 2004 was 8.02 to $0. With these estimates.5 billion in income is used with a range of zero to $10. lives lost. this gives an estimate of the increase in income.2084 D. wages. current estimates of greenhouse gases cost.869 × 10− 3 using the estimate of es obtained earlier. On a national scale the effect of ethanol production on gross domestic product (GDP) could be zero.S. they suggest a wide range of costs from $0. For analysis a benchmark of $7.02 as the benchmark value with a range of 0. the net marginal welfare effect from highway expenditure. With an increase in corn production and price enhancement. Wetzstein / Energy Economics 30 (2008) 2073–2090 Noting that Is = Iees. (δ′/λ)(dPF/dF). In estimating the contribution of highway capital to productivity. Nadiri and Mamuneas (1998) determine the current net social rate of return on highway expenditures to be 10% but in the 1950s and 1960s it was as high as 35%. Considering this regional effect. 2005). Wallsten and Kosec's (2005) analysis of the indirect cost of imported oil are one of the few investigations since 2003 (beginning of the Iraq conflict). are very speculative due to unknown long-run consequences. The amount of ethanol produced in 2004 was 3. The tax exemption just replaces Highway Trust Fund spending with spending on ethanol production.35. and salaries component of GDP is estimated to rise by $6. a leading study by Evans (1997) for the Renewable Fuels Association estimates an incremental annual addition to GDP of approximately $0. GDP would be $10.3 billion gallons. sanctions and having removed Saddam Hussein's regime.10 with a range of 1. Their estimates result in .0 billion. Vedenov. the opportunity cost of National Guard troops' lost civilian productivity. the change in annual income involves assessing the economic impact of ethanol plants.308 × 10− 3 and Is = 1. They estimate the cost of annual defense outlays to maintain the capability to defend the flow of Persian Gulf oil particularly from Iraq. is set at 1. and possible technology advancements. In reviewing the literature on these cost estimates. Using their estimates based on this assumption yields EPM = (δ′/λ)(dP/dM) = 0.4 billion gallons (Renewable Fuel Association.N. Considering a regional multiplier used by the Department of Commerce. Parry and Small (2005) assume air pollution from vehicles is proportional to miles traveled. In contrast. γ′/λ.00 to 1.

1 billion gallons and a level of disposable personal income of $8.22 per gallon of ethanol (Table 2). Associated with this central net cost is a range of $19.5 trillion. Is/αFI Governmental marginal benefits. the proportion of F to H as αFH = 4. MEBFs Fuel use Greenhouse gases. the last parameter in Table 1 is the proportion of fuel to income. 4. Vedenov. Results Applying the benchmark parameter values from Table 1 directly to (21) yields an estimated optimal ethanol subsidy s⁎ = $0.51 per gallon. − (E )ξ/αFα)Fs Ramsey subsidy. results in an incremental cost of fuel security (ρ′/λ)(∂A/∂F) = $0.75. − EPFFFs Fuel security. M. the negative marginal external benefits accounts for a large portion of this optimal/actual discrepancy.1 billion gallons. Wetzstein / Energy Economics 30 (2008) 2073–2090 2085 an annual cost of $34. which is less than the current highway tax exemption of $0. With an annual consumption of fuel of 134. αFI = 0.D.0932 − 0.00327 .303 with a range from 0. Using these parameters.22345 Total $0.1185 0. (γ′/λ)tFs Marginal costs Per dollar change in welfare. and hence EAF ¼ −0:173ð1−aR Þ ¼ −0:169. This is equivalent to annual expenditures on fuel efficiency of $113 and given F = 550.361 noting that EAH ¼ ðq V =kÞðAA=AHÞ ¼ ðq V=kÞðAA=AFÞðAF=AHÞ: Finally.250 to 0. − (EPMβ/αFM)Fs Fuel efficiency.2265 $− 0. (EPF ξ/αFα)Fs Aα Fuel security. we estimate ∂F/∂H as − 198/113 = − 1.0009 0. Dividing this cost by the annual consumption of fuel. a 198 gal in fuel savings.87.01462 0. and E AH = − 0.0 × 10− 5 0.173 per gallon of fuel.75) = 0. and E Aa ¼ −0:173F ¼ −95:2: The Alliance to Save Energy (2002) estimates the cost of fuel-economy improvements to increase fuel efficiency by 36% at $1693 over the 15 year life of a vehicle.173(− 1. As indicated from Table 2.7 billions yielding a net cost of $23. s⁎ $− 0. [(γ′/λ)(Fs + 1) − 1 − αs + (γ′/λ)αs]αR Marginal benefits/marginal costs Optimal ethanol subsidy. EAFFs Air quality.0 to $27.9 billion with benefits of $11.1113 0. 134. The ethanol fuel Table 2 Benchmark calculations of the optimal ethanol tax exemption (subsidy) Elements Estimates Components Marginal benefits Pigovian subsidy.0968 − 0.0387 − 0.0158. (EAHζ/αFH)Fs Renewable fuel share α Greenhouse Gases.0046 6.2 billion.7 billion.

these governmental marginal benefits along with the Pigovian and Ramsey subsidies are discounted by the marginal costs of the subsidy which include the loss of government revenue from increased use of ethanol fuels. a 1% increase in the subsidy yields a 0. will not automatically lead to positive marginal external benefits unless total fuel consumption is nonpositively responsive to the subsidy. Response of the optimal ethanol subsidy (dollars/gallon of ethanol) to a range of elasticity estimates. Attempting to increase the share of renewable fuels with a price subsidy. However. Note that the optimal subsidy is positive only because the benefits of economic development and increased government spending offset the negative marginal external benefits. 1. fuel security. air quality. subsidy does yield positive benefits from increased share of renewable fuel in the form of reduction in greenhouse gases and increased fuel security. The results also indicate that justification of a federal ethanol subsidy on environmental and fuel security grounds are questionable at best.e. The increased overall fuel consumption caused by the subsidy also augments the Highway Trust Funds which results in a fairly large contribution of the government marginal benefits to the total marginal benefits of the subsidy.52% increase in government revenue. which the subsidy in its current form stimulates by lowering the total fuel costs. The elasticity of government revenue with respect to the ethanol subsidy. The economic development benefits of the ethanol subsidy is the major justification supporting its social welfare benefits. estimated from (22) is Gs = 0. resulting from MEB b 0. the major argument in favor of the subsidy rests primarily on the economic development benefits generated by increased production of ethanol. and fuel efficiency are all negative. However. these positive benefits are overwhelmed by the negative externalities associated with the increased overall fuel consumption. The resulting marginal benefit to cost ratio yields the optimal ethanol subsidy of $0. Wetzstein / Energy Economics 30 (2008) 2073–2090 Fig. The net effects of the subsidy on greenhouse gases.52. M. such as the ethanol tax exemption. The negative Pigovian subsidy. If . is completely offset by the Ramsey subsidy and governmental marginal benefits. i. Instead.2086 D.22 per gallon. Vedenov.

EAH. 1 and 2 indicate that the optimal ethanol subsidy. the optimal subsidy is positively related to elasticities of ethanol fuel. and air quality cost. income. Fs. Hs. the optimal subsidy is negative over a wide range of this elasticity. s⁎. While sensitivity to the income elasticity. In particular. Is. In terms of the external benefits and cost. EPFF. . 5000 random draws of parameters in Table 1 were generated using two-sided power Fig. is very sensitive to the parameter values. γ′/λ. 2 illustrates the positive relations of the subsidy with government spending returns. both individual parameter variation and Monte Carlo analysis were implemented. is not as profound. Hs. 2. and the negative relations with greenhouse gases. In order to investigate the sensitivity of the optimal ethanol subsidy. and fuel efficiency.D. Is. Figs. EPM. and negatively related to the elasticity of fuel. Both Figs. es. This sensitivity of the optimal ethanol subsidy was also demonstrated by Monte Carlo analysis. Sensitivity analysis The wide range of parameter values in Table 1 suggests the benchmark optimal subsidy in Table 2 has an associated rather large variance. 1 and 2 illustrate the response of the optimal ethanol subsidy to a range of elasticities and external benefits and costs. s⁎. es. but is less affected by the elasticity of fuel efficiency. research leading to the narrowing of parameter ranges is warranted. Vedenov. The optimal subsidy is very sensitive to the elasticities of fuel. Fs. Wetzstein / Energy Economics 30 (2008) 2073–2090 2087 these economic benefits are mainly regional or confined to a particular set of states. 5. As indicated from Fig. M. Fig. 1. then state subsidies may be more appropriate. and especially to ethanol use. and cost of fuel security. Response of the optimal ethanol subsidy (dollars/gallon of ethanol) to a range of external benefits and costs. to ranges of the parameter values. For developing a more precise estimate of the optimal subsidy.

50 1. may be to increase the federal excise tax on gasoline so the socially undesirable effects of the ethanol Employing the power distributions had the advantage of capturing the asymmetric properties of parameter ranges relative to benchmarks.68 0. the solution. the probability that the optimal gasoline tax is $0.56 0.42 0. However. Table 3 lists the probabilities of optimal subsidy being below specific thresholds.00 Probability s⁎ b x 0. For the ethanol subsidy to contribute toward environmental and fuel security benefits.75 or less is only 29%. The effect the subsidy has on increasing total fuel consumption with associated negative environmental and fuel security external benefits underlies this outcome.S. tax rate is too low when considering the social welfare attributes.75 0. In the short run. and contributing to economic development. Vedenov. Implications It is very important in public policy analysis to consider the market effects of such policies. an increase in the gasoline tax in tandem with the subsidy may be required in order to offset the increased use of fossil fuels due to the subsidy.25 0. This result is consistent with Parry and Small's (2005) investigation of the optimal gasoline tax. however unpopular. As indicated from Table 3. increasing energy security. 2005 Energy Bill takes a step in this direction by mandating a certain percentage of renewable fuels to be used in all vehicles.0 0. the probability of the optimal subsidy being nonpositive is over 72%.00 − 0. a solution to this conflict may lay in increasing the number of E85-capable vehicles and overall shift to higher-proportion of ethanol blends so that ethanol subsidies stimulate use of ethanol rather than fossil fuels. The current ethanol tax exemption further amplifies this negative effect of a low gasoline tax by stimulating fuel consumption and overall retards the social welfare objectives. Ethanol fuels are considered a primary tool in meeting this mandate.00 2.S.92 probability distributions over respective ranges of parameters. and to create an empirical CDF for the optimal subsidy. reducing greenhouse gas emission. a statement that an ethanol subsidy will automatically result in benefits of improved air quality and fuel security along with reducing greenhouse gases is questionable given the impact this subsidy has on increasing overall fuel consumption. Wetzstein / Energy Economics 30 (2008) 2073–2090 Table 3 Monte Carlo results for optimal ethanol subsidy Level. In their analysis.00 − 1. 6. The objective of establishing a renewable fuel industry within the U.64 0. The results were generally robust with other distributions.4 The drawn parameters were then used to calculate the optimal ethanol subsidy in (21). The U.S. 4 . In the long run.50 − 0. In terms of the federal excise tax on gasoline. x (dollars/gallon) − 2. M. Parry and Small (2005) estimate the current U.2088 D.72 0.25 0.78 0.84 0. The government providing market incentives for the establishment of a renewable fuels industry is a case in point. is to increase social welfare by improving air quality.

International Journal of Production Economics 101. This suggests state – rather than national – level subsidies for ethanol may achieve higher social welfare benefits. Schamed. Energy and Emission Benefits of Fuel Ethanol. September. H. 29–44. Excess Profits.W. December. Gallagher. Does Britain or the United States have the right gasoline tax? American Economic Review 95. 2002. Nadiri. Small. Argonne. nationally the benefits may only be minor.. M.. Accessed December.S. e.S. Dept.html. 2003. American Coalition for Ethanol. Available online at http://www. J. 2006. Available online at http://www/ethanolrfa. 2005. Brubaker. Renewable Fuels Association Reports and Studies. Price...A. Care should be taken in distinguishing regional economic benefits from national benefits. Demand for Ethanol Aggravates Pain at the Pump. Transportation Technology R&D Center. Clean air and renewable fuels: the market for fuel ethanol in the U. US FHWA.. Federal Highway Administration. 2005a. Wetzstein / Energy Economics 30 (2008) 2073–2090 2089 subsidy are nullified.g. Energy Economics 20. Homegrown for the Homeland: Ethanol Industry Outlook 2005. 325–345. G. 2004. References Alliance to Save pg/061701/699475-28.N. 2005b. Bureau of Economic Analysis. This could result in positive marginal external benefits and associated Pigovian subsidy from the ethanol subsidy. Renewable Fuels Association (RFA). 2006. H. 2006. of Transportation. Accessed December. 2004. however. U. Ford.K. this income elasticity could be very large. Mamuneas. P. 1997. Without compensating for the increased fuel use from an ethanol subsidy. Ethanol subsidies and the highway trust fund. W. National Economic Accounts. M.S. Department of Transportation. 2005. Renewable Fuels Association (RFA). J. 1998. Bureau of Transportation Statistics. Federal Highway Administration. Combined with the positive Ramsey subsidy. National Transportation Statistics 2003. Will Increasing the Use of Ethanol in Gasoline result in Lower Prices and Cleaner Air? Parry.ethanolrfa. March. Vedenov. 2005. The “Volumetric Ethanol Excise Tax Credit” Eliminates the Impact of the Ethanol Tax Incentive on the Highway Trust Fund.P. McKay. M. 103–133. Contribution of Highway Capital to Output and Productivity Growth in the US Economy and Industries. ethanol US Department of Commerce. Argonne National Laboratory. Energy Journal 18. The Economic Impact of the Demand for Ethanol. Rask. 1276–1289.D. The results of our analysis also indicate that the optimal subsidy is sensitive to income elasticity given a change in the subsidy.24. Increasing America's Fuel Economy. 1998. Journal of Transport Economics and Policy 38. by increasing the federal excise tax on gasoline. Impact of pay-at-the-pump on safety through enhanced vehicle fuel efficiency. Shapouri. February. K. Ethanol Facts.H. Gallagher. M. U. Wall Street Journal 10.. Renewable Fuels Association (RFA). B1. Available online at http://www/ethanolrfa. 2003.. P. Journal of Policy Modeling 25.I. Oil Industry: Price policy/papers/view.S. Sioux Falls. Highway Statistics 2003. Price. SD (www. Business News (June 19) www. Welfare maximization. Lundegaard.. Consumer Federation of America. and allocation with a product performance or environmental quality standard: illustration for the gasoline and additives market.ethanol. K. Relatively small changes in the estimates used for this elasticity will result in the subsidy switching from positive to negative. 1997. February. US Department of Transportation.W. Shapouri. H. 230–245. Within a region. pricing. and Excuses. GM make big push to promote ‘flex-fuel’ vehicles.. 585–608.D.. T. J. 2005.. Position Papers. the anticipated social benefits of the ethanol tax exemption may not be achieved. National Center for Policy Analysis.stm. 2003. Cooper.. Spring Break in the U. 2004.. Khazzoom. 2004.. Rask. . P. Some long-run effects of growing of growing markets and renewable fuel standards on additives markets and the U. American Coalition for Ethanol. Fuel Economy Study: Comparing Performance and Cost of Various Ethanol Blends and Standard Unleaded Gasoline. this would yield net marginal benefits exceeding marginal costs. from 1984 to 1993.

Zhang.E. Kosec. Vedenov. D. Can the U.. Fiscal Facts... Argonne National Laboratory. Energy and Greenhouse Gas Emissions Impacts of Fuel Ethanol. 2005.S. Working Paper 05–19. Wetzstein. 2005. Tax Foundation. . Z. J. Wetzstein / Energy Economics 30 (2008) 2073–2090 Wallsten. The Economic Costs of the War in Iraq.2090 D. S. M. K.. M. M. Wang. Center for Transportation Research.. Williams.. Vedenov. 2005. AEI-Brookings Joint Center for Regulatory Studies. 2006. Ethanol Industry Compete in the Alternative Fuels' Market? Agricultural Economics (forthcoming 2007). Energy Systems Division..