Killing two birds with one stone: US and EU biofuel programmes

1. Introduction

The recent impulse given to biofuel policies is likely to produce major side effects on highly regulated agricultural sectors, particularly in the European Union (EU) and in the United States (USA), where public regulation of the agricultural sector is prominent.¹ Since the new biofuel policies trigger feedstock price increases, biofuel policies will have interactions with agriculture policies aiming at supporting farmers' income. Additionally, as most of the energy crop production might lead to more intensive agricultural practices, they will interact with environmental policies directed at agriculture as well.

As pointed out by several studies (see, for example, Elobeid et al., 2006; Schmidhuber, 2006; Tokgoz and Elobeid, 2006; Gohin, 2007), subsidising the biofuel industry raises the price of agricultural feedstock.² These subsidies (and mandatory blending schemes) come in addition to the ‘traditional’ payments directed to agriculture. In the USA and in the EU, the support adds up to USD 4 billion per year (subsidies only). Even in Brazil (the most efficient producer in the world), the support is evaluated at USD 1 billion (Koplow, 2006; Kutas and Lindberg, 2007).³ In the EU, the development of biofuel production will allow the agricultural sector to benefit from a dual support: on the one hand, member states hand out decoupled payments to farmers (single-farm payments–SFPs) and on the other, they give large support to the biofuel industries whose production costs exceed the price at which they can sell their output. The increase in agricultural commodity prices raises the farmers' revenue and reduces the need for direct income support. Hence, for a given objective in terms of agricultural income, the regulator is able to operate a partial substitution between direct agricultural income support and subsidies to the biofuel industry. In that sense, biofuel subsidies and mandatory blending schemes could be considered a new element in the already wide range of instruments at the regulator's disposal to improve farmers' income. Owing to the importance of the Common Agricultural Policy (CAP) in the EU budget, the question of a partial substitution of biofuel subsidies for CAP payments could be on the EU political agenda very soon.⁴ In the USA, the ethanol programme might lead to a long-lasting price increase for maize, wheat and soybeans, which could temporarily stop the counter-cyclical and loan deficiency payments (Babcock, 2006).⁵ Of course, the competition for the same input with biofuel firms is harmful for the agro-food industry, which opposed policies favouring first-generation biofuels from the outset (see Unilever, 2006; Forbes, 2006; Confederation of the Food and Drink Industries of the EU, 2006). As early as 2006, Goldman Sachs (Financial Times, 2006) pointed to a possible decrease in agro-food firm profits owing to biofuel production. Hence, with rising revenues for farmers, decreasing profits for the agro-food industry and a reduced consumer surplus, the net effect on aggregate welfare is unclear.

We develop a model that disentangles these various effects in a framework related to the literature on the efficiency of agricultural programmes which considers the incidence of the opportunity cost of public funds in evaluating economic instruments aiming at supporting farmers' incomes (see, for example, Gardner, 1983; Alston and Hurd, 1990; Alston and James, 2002).

We show that even without any other motivation to support biofuel production (for example, to increase security of energy supply), the government may find it worthwhile to implement a biofuel programme to diminish the social cost of the farm support programme: indeed, it may be socially beneficial to implement such policies if the cost of public funds is high. This result might be one possible explanation as for why biofuel programmes have been in place in the EU and the USA for more than a decade.⁶ Three different biofuel-supporting schemes are considered in this article, namely subsidies, mandatory blending and subsidies in the framework of a dual energy market. We show that a mandatory blending scheme leads to a level of biofuel production higher than in the subsidy framework. When a dual energy market is considered, the quantity of biofuel produced may not necessarily be greater than in the two other frameworks: it depends on the distribution of the consumers' willingness to pay for biofuel and the level of the opportunity cost of public funds.

Considering the possibility of importing biofuels, the government may still take advantage of the substitution between the farm support programme and the biofuel support policies (subsidies or mandatory blending schemes). This effect leads to a higher domestic price of the agricultural commodity than the world price, relatively low import levels and the imported biofuels benefiting from a lower subsidy than biofuels produced from domestic input.

The effects of biofuels on environmental policies are double-edged. On the one hand, biofuels are presented as one of the main features of GHG mitigation policies in the transportation sector.⁷ On the other hand, sizeable production of energy crops will lead to increased local pollution, linked to the intensification in the use of fertilisers and pesticides notably, triggering concerns on rivers and ground water pollution as stressed by Marshall and Greenhalgh (2006) for the USA and Graveline (2006) for the EU. Farmers respond to the new market conditions by giving up environmental-friendly rotations and tillage in favour of agricultural practices that are more nitrogen-intensive.⁸ In addition, high levels of crop prices might push some farmers to opt out of the Conservation Reserve Program (CRP).⁹ More recently, a report from the National Academy of Science underlined the very negative consequences in the Mexico Gulf (eutrophication) linked to the increased production of corn (for ethanol) in the Midwest (National Research Council, 2008).

Hence, the environmental externalities may be positive for Greenhouse Gas (GHG) emissions (this externality is global by nature), but negative for agricultural production (local externalities). There is thus an essential contradiction between setting a prominent objective for biofuel production that will lead (through higher prices) to higher yields and to an intensification of agricultural production and adopting sound agricultural practices. The positive global environmental externalities of biofuels regarding GHG emissions ought to be weighed against the negative local externalities generated by the production of the agricultural raw material. We analyse the optimal trade-off between GHG mitigation and sound agricultural production practices. We show that because of the social cost of public funds, the optimal standard is stricter than the Pigouvian level. Indeed, by setting a stringent environmental standard, the regulator increases the marginal cost of production, hence the price of the agricultural feedstock, which reinforces the substitution effect between the biofuel policy and the farm support programme. Finally, we compare the monitoring policies associated with two types of monetary sanctions that the government can inflict in case of noncompliance, fines and cross-compliance provisions, and we show that for a high level of biofuels, cross-compliance provisions may prove less effective than fines.

Most of the literature that has investigated the links between agricultural policies and biofuel support schemes reaches conclusions pointing to the inefficiencies of these programmes (as in Gardner, 2007; Babcock, 2008; de Gorter and Just, 2009). In particular, Babcock (2008) finds that the US ethanol policy induces large welfare losses and results in transfers from taxpayers and non-ethanol corn users to corn and ethanol producers as well as blenders. Our results contradict this strand of literature due to premises that differ in three respects: (i) our model integrates a parity income programme that constrains the design of the agricultural policy, (ii) we take into account the opportunity cost of public funds and (iii) we compare biofuel support policies with decoupled agricultural policies, while most of this literature compares biofuel policies (subsidies) with coupled instruments (loan deficiency payments). Another original aspect of our work is that we consider the effect of biofuel policies on environmental policies directed to agriculture.

This paper is organised as follows. Section 2 introduces our model. The optimal production of biofuels and the optimal environmental standard are derived in a framework with a biofuel subsidy considering that biofuels and fossil fuels are perfect substitutes on the consumers' side. Section 3 deals with the mandatory blending policy. In Section 4, we relax the assumption of a perfect substitution between biofuels and fossil fuels and analyse the policy under a dual energy market. In Section 5, the model is extended to account for the possibility of biofuel imports. Lastly, Section 6 addresses the environmental consequences of increased agricultural production, notably pertaining to the enforcement of the environmental policies directed at agriculture.

2. Biofuel subsidy and the agricultural market

Consider an economy with an agricultural sector, a food sector and a biofuel sector. All agents in this economy are price-takers. Both the food and biofuel industries use the agricultural feedstock in order to produce their own outputs (food products and biofuels, respectively). The total quantity of the agricultural product is thus split between the food (x_F) and biofuel (x_E) sectors: X = x_F + x_E. The production of the representative firm in the food industry is given by y_F = f_F (x_F), with f′_F > 0 and f″_F < 0. For the biofuel sector, we have y_E =γ x_E, where γ < 1 is the biorefinery yield.¹⁰ We assume in this section and in the following that biofuels and fossil fuels are perfect substitutes for consumers and thus they are indifferent between the two (or any mix of the two) as long as they are priced the same. We relax this assumption in Section 4, where we consider a dual energy market.

2.1. Biofuel subsidy and the parity income programme

In order to give a hint on the value of λ_s, consider rapeseed production in the EU-15. With a price elasticity of the agricultural crop supply equal to 0.28 (see FAPRI elasticity database, 2006), a biofuel production cost twice as high as its market value: and a 10 per cent decrease in the consumption of rapeseed by the food industry, i.e. ⁠, we obtain λ_s = 0.18, which is below the lower boundary of the range of λ given in the literature (0.2 to 0.6). Therefore, a strictly positive quantity of biodiesel ought to be produced in the EU-15 on purely redistributive grounds.

Solving programme (15) leads to the following result:

Proposition 1

If a biofuel subsidy programme is implemented (x_E > 0), the biofuel subsidy satisfies

(16)

where α = 1/(1 + λ). The optimal environmental standard in agriculture satisfies

(17)

Proof: See Appendix A.

When a biofuel subsidy programme is beneficial, equation (16) defines the optimal subsidy as the weighted sum of two marginal benefits: GHG mitigation with weight α = 1/(1 + λ) and the increase in the value of agro-food crops with weight ⁠. With no fiscal distortion (λ = 0), we simply have σ_E = γB′ (y_E), i.e. the Pigouvian rule that the optimal subsidy should equalise the marginal benefit of GHG mitigation. Of course, if γB′ (y_E) is negative, the government must not subsidise the biofuel industry. With λ > 0, the optimal subsidy puts a lower weight on the GHG mitigation objective (α decreases with λ) and a larger weight on the increase in the value of agro-food feedstock. Hence, the larger the shadow cost of public funds, the less the GHG mitigation concern is the motive for the biofuel subsidy. Of course, if B′ (y_E) is positive, it is beneficial to implement a biofuel subsidy programme even if the shadow cost of public funds is lower than threshold discussed above. But if B′ (y_E) is negative, the government may still be willing to implement such a programme because of a large shadow cost of public funds (⁠ in that case).

Likewise, the right-hand side term of equation (17) is the weighted sum of two marginal losses of the agricultural environmental standard measures the marginal damage of the agricultural production and the decrease in the value of the agro-food feedstock due to a marginal increase in the environmental standard (see the appendix for the derivation of ⁠). Indeed, by increasing the environmental standard (i.e. by allowing farmers to pollute more), the government decreases the marginal cost of the agricultural input and thus the market value of crops. With no fiscal distortions, implying α = 1, we have the Pigouvian rule, which calls for an environmental standard that equalises the marginal decrease in the production cost with the marginal damage of agricultural practices. With fiscal distortions (λ > 0), the government puts a lower weight on the environmental motive (α < 1) and accounts for the price effect of the standard policy: by setting a stringent environmental standard, the regulator increases the marginal cost of production which reinforces the substitution effect between the biofuel subsidy policy and the farm support programme.

3. Mandatory blending

Hence, in addition to the loss of surplus on the food market as identified above in the case of the subsidy policy, the consumer is affected by a loss of surplus on the energy market due to the impact on energy price of the blending policy. The marginal loss of surplus on this market is made up of two terms: one corresponding to the (implicit) biofuel subsidy and the other to the effect of the blending requirement on the agricultural price. Solving programme (18) leads to the following result:

Proposition 2

If a mandatory blending programme is implemented (x_E > 0), the optimal level of energy crops, which satisfies

(21)

is greater than under a subsidy programme. The optimal environmental standard, satisfying

(22)

is more stringent than the environmental standard with the subsidy scheme.

Proof: See Appendix B.

The optimal biofuel production under a mandatory blending policy is greater than under a subsidy policy. There are two reasons explaining this result. First, the fiscal advantage of the blending policy not only affects the feedstock directed to the food sector, x_F, but also the one directed to the biofuel production, hence the entire agricultural production X, as reflected by the last term of equation (21). Second, as the consumer surplus is affected by a unit weight in the SWF while the taxpayer surplus is weighted 1 + λ due to fiscal distortions, the implicit subsidy level given in equation (21) indicates that the blending policy does not balance the environmental benefit of the GHG mitigation with the fiscal advantage of a larger agricultural price as in equation (16). As a consequence, this implicit subsidy is larger than the GHG marginal benefit whatever λ > 0. For the environmental standard, as it affects the agricultural production cost and thus the decoupled payment directed at farmers and financed by the taxpayer, the rule given by equation (22) shows a weighted sum of two marginal benefits very close to equation (17), reflecting the fact that the increase in the agricultural cost due to the environmental standard raises the government outlay. However, as in equation (21), the last term of equation (17) indicates that the fiscal advantage of the rise in the agricultural price affects the entire production of feedstock, which explains that this standard is more stringent than under a subsidy policy.

4. Dual energy market

In words, reducing the biofuel subsidy by EUR 1 (hence, increasing p_B by EUR 1) allows to raise [1 − F(θ_E)]y_M on the biofuel market, and thus the government is able to decrease the taxpayer charge by (1+λ)[1 − F(θ_E)]y_M, leading to a social net gain given by the first term of equation (25). However, this price increase also leads a fringe of biofuels consumers, equal to f(θ_E), to switch to fossil fuels and thus it leads to a decrease in the biofuel production equal to f(θ_E)y_M. The second term of equation (25) corresponds to the supplementary amount of tax needed to maintain the farmer revenue at the parity income level due to this marginal decrease in energy crops production. The result of these two counteracting effects depends on the distribution of the consumers' willingness to pay for biofuels and on the extent of the deadweight loss due to taxation, as stated formally in the following proposition:

Proposition 3

When biofuels and fossil fuels are imperfect substitutes, if a biofuel subsidy programme is implemented (x_E^* > 0), the optimal quantity of energy crops is larger than in the perfect substitution framework if

(26)

where θ_E is given by equation (24).

Proof: See Appendix C.

5. Imports of biofuels

The results of the previous section are limited to biofuels produced domestically. Buying biofuels on the world market could prove less expensive for society, even if governments must also account for the indirect benefits associated with a domestic biofuel support programme.¹⁶ To investigate this question, we assume that a quantity of biofuels z_I bought at price p_I is imported from a perfectly elastic world supply.¹⁷ The environmental GHG benefit linked to the use of the biofuel in replacement of fossil fuel is supposed to be identical whatever the biofuel origin, leading to total benefit B(y_E + z_I).¹⁸ We consider only a subsidy policy in this section and assume that biofuels and fossil fuels are perfect substitutes for consumers.

Denoting by and the optimal regulator choices, we have the following results:

Proposition 4

When the government can import the agricultural feedstock at pricep_I

• with no constraint on the biofuel production level, it is optimal to produce energy crops if λ is large. The optimal policy satisfies the following conditions

  (28)

  and

  (29)

  In particular, if the GHG marginal mitigation benefit is low, all the agricultural feedstock is produced domestically:and⁠.

• and has to fulfil an exogenous biofuel production objectiveQ > γx_E^*, it is optimal to produce energy crops domestically at level⁠. The internal price of the agricultural feedstock verifiesγp_A > p_Ileading to subsidiesγσ_E > σ_I.

  Proof: See Appendix D.

Condition (29) states that – in the absence of any constraint on the biofuel production – the optimal imported crop level equates the marginal social cost of subsidising the biofuel industry with the marginal environmental benefit of biofuel. If the marginal GHG mitigation benefit is large, it is optimal to import biofuels. Of course, if these benefits are low (or negative), the government does not allow imports: we have ⁠. Compared with equation (29), condition (28) entails the marginal social gain of the transfer of revenue from the food sector to the farmers, which eases the condition for a positive level of domestic biofuel crops. As equation (28) is similar to equation (16), if no biofuel objective is imposed to the regulator and GHG mitigation is low (or negative), a production of domestic energy crops will take place exactly as in the autarky framework, i.e. we have ⁠, and This result seems quite intuitive as the reason which had led to reach a ‘natural’ level of energy crops x_E^* essentially hinged upon the social benefit of transferring income from the agro-food industry and the consumer to the taxpayer, thus diminishing the distortions linked to taxation. The possibility of importing biofuels does not alter this rationale. However, when the regulator is faced with a biofuel objective, the situation changes with respect to the autarky framework. As the regulator goes beyond the ‘natural’ level of energy crops x_E^*, it trades off between two possibilities: producing domestically or importing. As long as the opportunity cost of blending domestic biofuels is smaller than the opportunity cost of blending imported biofuels, domestic production is preferred. Beyond this quantity level the remaining quantity to reach the constraint Q is imported.

6. Policy enforcement and cross-compliance

As stressed by Marshall and Greenhalgh (2006) for the USA and Graveline (2006) for the EU, in response to the increased demand of biofuel feedstock, farmers may be tempted to abandon environmental-friendly rotations and tillage. The alternatively adopted agricultural practices might well lead to water pollution problems due to a notable intensification in the use of fertilisers and pesticides as well as to increased soil erosion. In this section, we discuss the problem of enforcing environmental standards given the biofuel objectives of the governments. As noted by Bontems and Rotillon (2007), an efficient environmental policy is not a mere definition of the right level of tax or norm to abide by. To enforce a demanding policy, it is necessary for the State to control farms frequently, which is costly, and to be able to inflict sizeable penalties to have some effect on farmers' behaviour. The conventional rule which states that pollution should be reduced to the point where the marginal damage and the marginal cost of pollution reduction are equal does not apply if enforcing the rule is costly. The cost of pollution should be added up with the marginal cost of control. This ultimately reduces the level of environmental standards that must be imposed on farmers.¹⁹ Our focus is more particularly on the 2003 CAP reform, which imposed ‘cross-compliance’ provisions. This policy conditions the benefit from market support schemes to farmers abiding by environmental protection requirements. With the enactment of this policy, all farmers receiving direct payments must fulfil the requirements of 19 European legislative acts applying directly at the farm level in the domains of environment, public and animal health, pesticides and animal welfare. Farmers will face partial or total withdrawal of their SFP in case of non-compliance. As the SFP is a decreasing function of the farmers' revenue, the maximum fine in case of non-compliance is lower the larger the farmers' revenue. As the farmers' revenue increases with the biofuel production objective, it seems that this compliance policy may prove ineffective in the future given the high level of biofuel production forecast for 2020. We shall analyse this problem in a framework similar to Malik (1992), considering that inspecting farms is costly and that the government inflicts penalties that depend on the extent of the infringement. We analyse the two cases of an exogenous maximum penalty and a maximum penalty which corresponds to the farmer's decoupled payment, as is the case in the EU. We do not consider imports in this section.

Assume that whenever a farmer has chosen an emission level e that exceeds the standard ⁠, the agency is able to inflict a penalty that depends on the extent of the farmer's infringement, ⁠, and more precisely that the corresponding penalty is a fraction of a maximal penalty Ψ. The function f(·) is exogenously given (by an independent legislative body) and is assumed increasing and convex in the seriousness of the infringement, with f(0) = 0. The maximum penalty can either be a given amount (also determined by an independent legislative body), or the decoupled payment that the farmer should receive in case of compliance, i.e. The latter case corresponds to the current framework chosen by the EU to enforce environmental policies in agriculture. Let k be the intensity of control (the probability a farm is inspected), and μk the corresponding cost.

In this welfare function, the term in square brackets corresponds to public spending, which is made up of three components: the biofuel subsidies (C_X − γp_E)x_E, the decoupled payments to farmers and an additional term μk representing the cost of controlling farms.

Hence, cross-compliance may prove the most efficient policy if is low compared with the parity income and if agricultural production is low, while the government is more likely to choose an exogeneous maximal penalty policy for large biofuel objectives. Of course, for a given production level, the environnemental standard is different depending on the compliance policy. Denote by e^cc (Q) the optimal environmental standard under the cross-compliance policy given biofuel objective Q. As e^cc (Q) increases with Q, the farmer's profit also strictly increases with Q. Hence, for any biofuel production greater than ⁠, the government is able to implement the optimal environmental standard of the cross-compliance policy with an exogenous maximum penalty policy and thereby it can reduce its monitoring effort and thus the cost of the enforcement policy. Reciprocally, for production Q < Q_s, cross-compliance proves to be the most efficient policy.

7. Conclusion

The main results of this paper can be summed up as follows. First, we have shown that biofuel programmes may allow the regulator to operate a partial substitution between decoupled payments and the support for biofuels. This substitution is detrimental to the food industry (and to consumers). However, when the social cost of public funding is high, the regulator should finance a biofuel programme because of its redistributive property. Of course, this result rests on the existence of sufficiently high distortions in the tax system. Positive environmental externalities of the substitution of biofuels for fossil fuels tend to push the optimal biofuel quantity a step further. Conversely, if these externalities are negative, the support is decreased. The conclusions drawn in the case of a biofuel programme financed through subsidies can also be made when biofuels are promoted thanks to a mandatory blending scheme. The optimal level of biofuels that ought to be produced is even higher in that case. When a dual energy market is implemented, the optimal production of energy crops is not necessarily greater than with the other policies: it depends on the distribution of the consumers' willingness to pay for biofuels and on the opportunity cost of public funds. Taking into account the possibility of imports, the optimal level of energy crops produced domestically is set at a level where the domestic price exceeds the world price for energy crops, thanks to the saving of public funds permitted by biofuels.

The emergence of biofuel policies marks a profound change in the path of agricultural policy reform, which has mainly consisted in decoupling the support awarded to farmers from production decisions. This evolution has begun with the 1992 reform and was further reinforced during the Agenda 2000, Mid-Term Reform in 2002 and the ‘Health-Check’ which was adopted at the end of 2008. Clearly, the decision to design large-scale biofuel programmes has contributed to sharp increases in agricultural commodity prices, and this new biofuel policy, although formally taking place outside the CAP, has deep consequences on the logic of its future evolution. It could be argued that biofuel support policies have implications that look like re-coupling. Moreover, the enforcement of stringent environmental policies in agriculture may be jeopardised by the high price levels triggered by biofuel programmes. The reforms of the CAP had led to the implementation of cross-compliance provisions, with the possibility for the regulator to fine the farmers in case of infringement, the penalty paid by the farmer being proportional to the decoupled payment. With a new framework of high commodity prices and decreased decoupled payments, the cross-compliance scheme would be put at risk.

Many extensions of our framework could be investigated. First, we have considered that the subsidies were fine-tuned. This assumption could be criticised, as there are informational asymmetries between the regulator and the biofuel firms. In the mandatory blending framework, such informational asymmetries are not relevant but informational rents could well be replaced by monopolistic rents for biofuel producers. For instance, the major biodiesel firm in France covers more than 75 per cent of the market. We have also assumed a perfectly competitive agro-food sector. Relaxing this assumption may well lead to very stringent conditions for a socially valuable subsidy substitution effect between the farm support and the biofuel programmes.

Acknowledgements

Financial support received by the ‘New Issues in Agricultural, Food and Bio-energy Trade (AGFOODTRADE)’ (Grant Agreement no. 212036) research project, funded by the European Commission, is gratefully acknowledged. The views expressed in this paper are the sole personal responsibility of the authors and do not reflect those of the Commission, which has not viewed, let alone approved the content of the paper. The paper does not reflect the views of the institutions of affiliation of the authors either.

References

Appendix A: Proof of Proposition 1

Rearranging term gives equation (17).

Appendix B: Proof of Proposition 2

Rearranging terms and using gives equation (21). Similarly, equation (22) is derived from equation (B3) using (A6).

Appendix C: Proof of Proposition 3

Appendix D: Proof of Proposition 4

Author notes