WP 2010-22 November 2010 (Updated May 2011)

Preparing to load PDF file. please wait...

0 of 0
100%
WP 2010-22 November 2010 (Updated May 2011)

Transcript Of WP 2010-22 November 2010 (Updated May 2011)

WP 2010-22 November 2010 (Updated May 2011)
Working Paper
Charles H. Dyson School of Applied Economics and Management Cornell University, Ithaca, New York 14853-7801 USA
The Implications of Alternative Biofuel Policies on Carbon Leakage Dusan Drabik, Harry de Gorter and David R. Just
1

The Implications of Alternative Biofuel Policies on Carbon Leakage
Dusan Drabik, Harry de Gorter and David R. Just Charles H. Dyson School of Applied Economics and Management,
Cornell University, USA E-mail: [email protected], [email protected], [email protected]
First version: November 2010
Updated: May 2011
Abstract We show how leakage differs, depending on the biofuel policy and market conditions. Carbon leakage is shown to have two components: a market leakage effect and an emissions savings effect. We also distinguish domestic and international leakage. International leakage is always positive, but domestic leakage can be negative. The magnitude of market leakage depends on the domestic and foreign gasoline supply and fuel demand elasticities, and on consumption and production shares of world oil markets for the country introducing the biofuel policy. Being a small country in world oil markets does not automatically imply that leakage is 100 percent or above that of a large country. We show leakage due to a tax credit is always greater than that of a mandate, while the combination of a mandate and subsidy generates greater leakage than a mandate alone. In general, one gallon of ethanol is found to replace only 0.35 gallons of gasoline – not one gallon as assumed by life-cycle accounting. For the United States, this translates into one (gasoline-equivalent) gallon of ethanol emitting 1.13 times more carbon than a gallon of gasoline if indirect land use change (iLUC) is not included in the estimated emissions savings effect and 1.43 times more when iLUC is included.
Key words: biofuels, market leakage, indirect output use change, carbon leakage, emissions savings, domestic leakage, tax credit, mandate
JEL: Q27, Q41, Q42, Q54
2

The Implications of Alternative Biofuel Policies on Carbon Leakage
1. Introduction The issue of carbon leakage – where emissions reductions by an environmental policy
are partially or more than offset because of market effects – is often raised as an issue that will undermine environmental policies.1 Leakage has been extensively studied in the cases of cap and trade policies (e.g., Frankel 2009),2 reduced deforestation and land degradation REDD (e.g., Murray et al. 2004, 2009; Murray 2008) and indirect land use change (iLUC) generated from biofuels policies (e.g., Searchinger et al. 2008; Hertel et al. 2010; Tyner et al. 2010).3 Each source of leakage has created its own controversy. For cap and trade, green tariffs and producer rebates have been studied extensively as remedial measures,4 while the Kyoto Protocol has been reluctant to include REDD because of concerns over leakage and additionality (Murray 2008). In the case of biofuels, the issue has been whether or not biofuels fulfill a sustainability threshold (e.g., a 20 percent reduction in carbon emissions for U.S. corn-ethanol relative to gasoline it is assumed to replace). However, leakage has also been a criterion to determine the eligibility of biofuels for carbon offsets in the Clean Development Mechanism of the Kyoto Protocol.5
What has not been studied to date is the indirect output use change (iOUC) in the fuel market itself where the addition of biofuels always causes a reduction in world gasoline market prices.6,7 This paper develops a formal analytical framework to analyze the carbon leakage due to alternative biofuel policies, namely biofuel consumption subsidies (like the U.S. blender’s tax credit or a fuel tax exemption at the retail pump in many other countries) and mandates, and the combination of a subsidy and a mandate.8 In so doing, we identify two components of carbon leakage: the “market leakage effect” (also referred to as iOUC) and the “emissions savings effect”. The former refers to the resulting market effect of biofuels in displacing gasoline and other oil (domestic non-transportation and international oil)
3

consumption,9 while the latter refers to the relative carbon emissions of biofuels versus gasoline. A positive market leakage (which always occurs with a tax credit that expands fuel consumption) does not necessarily imply an increase in carbon emissions, but a negative market leakage (that is possible with a mandate) always implies a higher emissions reduction relative to what is intended.10
We distinguish “domestic” versus “international” leakage. Because world gasoline prices decline with either biofuel policy, international leakage is always positive, as is domestic leakage with a tax credit. But domestic leakage with a mandate can be negative under some market conditions, making it possible that total (domestic plus international) leakage can be negative. For plausible parameter values we, however, find that, in reality, this is not the case as international leakage is much bigger than domestic leakage.
Nevertheless, the level of market leakage for either policy depends on two key market parameters: (a) the elasticities of gasoline supply curves and fuel (gasoline plus biofuel) demand curves; and (b) consumption and production shares of the country introducing the biofuels. But leakage is found empirically to be more sensitive to elasticities than to market shares, and especially to changes in market parameters of the country not introducing biofuels.
Domestic leakage becomes more important relative to international leakage as the Home country consumes more gasoline and/or the relative demand elasticity of the Home country increases. Our empirical results show that domestic leakage is less important for total market leakage compared to the case of carbon leakage – a result driven by the emissions savings effect.
We show that a small importer (exporter) of oil facing a perfectly elastic excess supply (demand) curve does not automatically generate 100 percent market or carbon leakage. We also show that, under some market conditions, a country whose biofuel policies
4

have a smaller impact on world oil prices can see lower leakage compared to a country that lowers world oil prices more significantly.
The economics of a consumption mandate is shown to be more complex than that of a tax credit because the former generates a U-shaped fuel supply curve. However, for the same amount of ethanol, market leakage due to a tax credit is always greater than that due to a binding consumption mandate. We also find that the combination of a binding consumption mandate and a tax credit produces greater leakage than with a mandate alone. If in combination with a mandate, the leakage due to the tax credit alone is infinite.
If the tax credit is equal to the price premium that is necessary to generate the mandated amount of ethanol, then the tax credit exactly offsets the reduction in gasoline consumption due to the mandate if the country has no effect on world oil prices. However, if the country with the biofuel policy can affect world oil prices, then the tax credit more than offsets the reduction in gasoline consumption due to the mandate.
For most plausible elasticities and 2009 U.S. market shares, we find market leakage to be in the order of 60 to 65 percent for all three policy options (a tax credit, a mandate, and their combination), i.e., one (gasoline-equivalent) gallon of ethanol replaces only 0.35 to 0.40 gallons of gasoline and the rest (0.60 and 0.65 gallons, respectively) is displaced. This combined with the effect of iLUC makes one gallon of ethanol emit 1.43 times more carbon than one gallon of gasoline. Note that the EPA in its evaluation of iLUC using life-cycle accounting assumes a one-to-one replacement of gasoline with ethanol. On the other hand, the magnitude of carbon leakage is lower when iLUC is not taken into account, 20 to 25 percent, (because the emissions savings effect is strong) but significantly higher, 190 to 210 percent, when the effect of iLUC is considered that weakens the emissions savings effect. We show that iLUC is less important than iOUC.
5

Leakage under “autarky” can be interpreted as a measure for when all countries participate in the environmental policy (a biofuel policy in this case). We find that both the market and carbon leakage are lower by ‘expanding the coalition’ but not significantly so with the possible exception of carbon leakage with mandates but only when the emissions savings effect is sufficiently strong.
The remainder of this paper is organized as follows. The next section defines leakage and explains two components of carbon leakage – market leakage and the emissions savings effect. In Section 3, we analyze market leakage due to a blender’s tax credit. The discussion includes implications for how country size on world oil markets affects leakage. In Section 4, we investigate market leakage under a binding consumption mandate and discuss the leakage effects of adding a blender’s tax credit to the mandate. Numerical estimates of leakage and their sensitivity analyses are provided in Section 5. The last section provides some concluding remarks. 2. Market and Carbon Leakage Defined
Whenever a ‘clean’ biofuel is subsidized or mandated relative to a ‘dirty’ source like gasoline, carbon leakage occurs - the actual carbon savings may be more or less than the intended savings (from biofuel consumption). Carbon leakage is a result of two, typically, counteracting effects: the “emissions savings” effect and the “market leakage” effect (also referred to as “indirect output use change” (iOUC), see de Gorter and Just 2009b) in the fuel market.11 To define the former, denote carbon emissions per unit of energy from a dirty (e.g., gasoline) and clean (e.g., biofuel) source by ed and ec, respectively. Define the emissions savings effectξ to be the relative difference between ed and ec:
ξ = ed − ec ed
6

The interpretation of ξ is straightforward. A value of ξ = 0.20 means that a (gasoline equivalent) gallon of ethanol emits 20 percent less carbon relative to the same amount of gasoline.
While the emissions savings effect depends mostly on technical properties of the two fuel sources, the market leakage effect results from market forces in the fuel market after the introduction of biofuels. To show this, we write the initial world consumption of fuel, which is assumed to be all gasoline, as:

C0 = CH 0 + CF 0

where H and F denote Home and Foreign country, respectively. In the new equilibrium with E units of biofuels, world fuel consumption is given by:

C1 = E + CH1 + CF1

Market leakage (in absolute terms) due to the introduction of E units of biofuels is the change in world fuel consumption:

ΔC = C1 − C0 = E + CH1 + CF1 − CH 0 − CF 0 = E + ΔCH + ΔCF

where ΔCH and ΔCF represent a change in consumption of gasoline in the Home and Foreign country, respectively.
In relative terms, the market leakage effect is given by:

L = ΔC = E + ΔCH + ΔCF

ME

E

For example, if LM = 0.7, then one unit of biofuel replaces 0.3 units of gasoline, while total fuel use has increased by 0.7 units. Contrast this, for example, with life-cycle accounting which assumes gasoline is replaced by the biofuels gallon for gallon (gasoline equivalent).

7

We define carbon leakage in an analogous way to the market leakage: the change in global carbon emissions, due introduction of biofuels, is divided by the intended carbon reduction. The formulae for the market leakage and emissions savings effects can be combined to derive an expression for carbon leakage LC:

(1 − ξ )ed E + ed (ΔCH + ΔCF ) 1 − ξ ΔC + ΔC 1 − ξ ΔC − E

LC =

ξe E

=+ ξ

H

F=

+

ξE

ξ

ξE

d

= 1 − ξ + ΔC − E = 1 ΔC −1 ξ ξE ξE ξ E

which can be rewritten into a simple form:12

L = 1 L −1

(1)

C ξM

In deriving an expression for carbon leakage, we realize that the clean source only saves ξ [×100%] carbon relative to the dirty source.
Equation (1) clearly identifies the two driving forces of carbon leakage: the emissions savings and the market leakage (iOUC) effects. Depending on the relative value of the factors, three cases can be distinguished to determine the magnitude of carbon leakage. First, carbon leakage is zero, i.e., total emissions have not changed if LM = ξ . In this case, the carbon savings effect of biofuels is completely offset by an increase in fuel consumption, which results in higher emissions. Second, carbon leakage is positive whenever LM > ξ . In this instance, total emissions always increase. Notably, carbon leakage is more than 100 percent if LM > 2ξ . This means that an increase in global carbon emissions is higher that the intended reduction in emissions due to biofuels. Finally, in the event that LM < ξ , global emissions are reduced. In particular, they decrease by as much as intended if LM = 0, i.e., if ethanol replaces gasoline one to one. Formula (1) also reveals that introduction of biofuels

8

can reduce more carbon than initially expected only when the iOUC effect is negative (LM < 0), a situation possible only with a mandate as we show later.
It is the use of biofuels as a substitute for gasoline that gives rise to the many potential sizes and signs of carbon leakage. The magnitude of carbon leakage also depends critically on the value of the emissions savings effect. For example, total carbon emissions could increase if coal were replaced with oil, but very likely decrease were the former replaced with natural gas. To illustrate the sensitivity of carbon leakage to the size of the emissions savings effect, we note that the direct emissions of corn-ethanol (as measured by life-cycle accounting) are 52 percent less than emissions from gasoline (EPA 2010).13 In this case, the magnitude of market leakage is multiplied by two (i.e., 1/0.52, as per equation (1)). But if indirect land use change (iLUC) is taken into account, then corn-ethanol only saves 21 percent relative to gasoline (RFA 2010). The magnitude of the market leakage is multiplied by five (1/0.21) in this case (as per equation (1)).
The formula for carbon leakage given by (1) is also very general; it accommodates both autarky and international trade cases; allows for any type of policy that affects the introduction of biofuels on the market; and it requires some estimate of the emissions savings effectξ to determine the magnitude of carbon leakage. It also indicates that carbon leakage can only be positive, i.e., total emissions increase, if market leakage is positive. On the other hand, there can be situations when carbon leakage is negative, i.e., introduction of biofuels reduces total emissions, even though market leakage is positive. The latter outcome may occur when the market leakage (iOUC) is sufficiently small and/or a biofuel has substantially lower carbon emissions relative to gasoline.
There are competing methods to derive an estimate for the value of emissions savingsξ for a biofuel. First, one can compare the instantaneous amounts of carbon released when a fossil fuel and biofuel are combusted. In this case, the biofuel emits a different
9

amount of CO2 relative to the fossil fuel.14 Second, some argue that biofuels are by and large net zero as the amount of CO2 absorbed through the process of photosynthesis in growing the crop is equal to that when the biofuel is combusted.15 In this case,ξ = 1for any biofuel compared to a fossil-based fuel. Third, the life-cycle accounting approach measures all carbon emissions from “well-to-wheel” for fossil fuels and from “field-to-tank” for biofuels, which might yield yet another value forξ (e.g., 52 percent saving with corn-ethanol relative to gasoline (RFA 2010)). The fourth option is to add iLUC to the life-cycle value forξ (e.g., 21 percent savings with corn-ethanol relative to gasoline (EPA 2010)).
Implicitly embedded in equation (1) is the fact that the existence of positive market leakage undermines the emissions savings effect. Therefore, a question arises as to what the true emissions savings of ethanol compared to gasoline are when the iOUC effect is taken into consideration. A change in global carbon emissions due to the iOUC effect
is (E + ΔCH + ΔCF )ed . Therefore, the true carbon emissions of a gallon of ethanol introduced
in the market are given the sum of the own emissions of ethanol and the emissions due to the iOUC effect:

(1 − ξ )ed E + ed (E + ΔCH + ΔCF ) = (1 − ξ + L )e

E

Md

Substituting this expression into the definition of the emissions savings effect above, we arrive at:

ed − (1 − ξ + LM )ed = ξ − L
ed M This result is very intuitive: in the presence of positive market leakage, emissions savings of ethanol relative to gasoline are always lowered by the counteracting market leakage effect; hence never as high as supposed to be. In particular, if the market leakage

10
LeakageCarbon LeakageGasolineMarket LeakageBiofuels