Deepak Rajagopal, Gal Hochman, David Zilberman** (Deepak@berkeley.edu)
Fuel security and environmental considerations led governments worldwide to enact policies that support the development of alternatives to oil during the last decade. As a consequence of these policies, biofuels emerged as an alternative to conventional transportation fuels. These policies led to concerns about the unintended consequences to the environment and to food security. Policy makers are responding to these concerns through new types of environmental regulations that take a lifecycle view of emissions from fuels. We discuss whether the novel approaches to environmental regulation of bioufels (in the case of the U.S. Renewable Fuel Standard) and of fuels in general (in the case of the California Low Carbon Fuel Standard), are based on established sound principles of policy making.
In the U.S. biofuel mandates were first established through the Energy Policy Act of 2005. This policy mandated the production of 7.5 billion gallons of ethanol by 2012. At that time, the prevailing opinion about the environmental performance of biofuels such as corn ethanol in the U.S. and cane ethanol from Brazil was, despite some differences, generally that they have lower average carbon intensity than gasoline. It was also known that the carbon-intensity varies with the type of biofuel, and how it is produced (See Farrell et al. 2006 and Pimentel and Patzek 2005). For instance, although corn ethanol resulted in 20% less GHG emissions than gasoline, the net GHG benefits varied from less than 10% when it was produced using coal-based energy to greater than 30% when it is produced using natural gas. Yet, biofuel policies at that time treated all biofuels equal from an environmental standpoint.
These initial estimates of the GHG intensity of biofuels were based on assessments of biofuel production that was modest in scale relative to the scale of food production and also the scale of the fuel consumption. However, there was a growing recognition that further expansion of biofuels would lead to impacts thus far unaccounted in environmental lifecycle assessments (e.g., land use). The term 'lifecycle greenhouse gas emissions' refers to emissions related to the full fuel lifecycle, including all stages of fuel and feedstock production and distribution, from feedstock generation or extraction through the distribution and delivery and use of the finished fuel to the ultimate consumer, where the mass values for all greenhouse gases are adjusted to account for their relative global warming potential. The accounting technique used to calculate lifecycle emissions is called lifecycle analysis (LCA). Using a well-established partial-equilibrium model of international agricultural trade, Searchinger et al. (2008) calculated that taking into account emissions from land use change, caused by increase demand for cropland, results in GHG emissions that are on a per unit of fuel energy basis, twice as much as emissions from gasoline. Although the extent of land use changes estimated by Searchinger et al. is debatable, a consensus about the potential negative impact of land use change emerged.
This coupled with concern about food security, led to two major changes in biofuel policies. First, policy makers recognized that all biofuels are not created equal, leading to design of policies that encourage the development of better biofuels. Accordingly, the Energy Independence and Security Act (EISA) of 2007, which not only increased the volume of biofuel required to be blended into transportation fuel from 9 billion gallons in 2008 to 36 billion gallons by 2022, but also established different mandates for different types of biofuels. The largest mandate is for cellulosic biofuels derived from non-food sources and such biofuels are required to have a GHG intensity that is 50% lower compared to gasoline on a lifecycle basis. A similar but different type of policy the California LCFS policy mandates at 10% reduction in the average GHG intensity of all transport fuels sold within California. A second feature of both the RFS and LCFS policies is the inclusion of emissions from indirect land use change as part of the fuel lifecycle (the notion of indirect land use change is explained in detail later). The use of LCA and the consideration of indirect effects are novel aspects, which have not been attempted in environmental regulations especially since indirect effects are those that are not directly attributable to the supply chain of biofuel, but arise through the intricate web of interconnected markets.
A rationale for lifecycle accounting for GHG regulation
The motivation for the use of LCA stems from the fact that policies for controlling GHG, a global pollutant, are not global. Under a global policy, every polluter is responsible for his or her actions and every unit of GHG emission is accounted for. If this policy imposes or achieves a price on emission equal to the marginal social damage from emission then this policy achieves optimality. In the absence of a globally consistent policy, regional or sectoral policies that aim to reduce emissions from a subset of activities (say emission from fuel combustion) may cause emissions to increase in unregulated sectors. For instance, electric vehicles reduce emissions from transportation but increase emissions from the electricity sector. Similarly, while the carbon contained in the biofuel is eventually recycled through photosynthesis, it increases emissions in the agricultural sector due to increase in use of fertilizers, farm machinery and tilling. In the worst case the net effect may be an increase in total emissions. Thus, a lifecycle approach regulation, which takes into account for GHG emissions coming from activities in unregulated sectors, in addition to the regulated sector, may in principle be considered.
Issues in use of LCA
The key principle in regulation based on LCA is that one actor, usually the producer of the final good, is accountable for the activities of all actors within the supply chain.
The boundary for LCA? If one looks to assign accountability, a key question is the definition of the boundary of activities that the final producer is accountable for. This is a matter of intense debate. LCA has traditionally focused only on calculating emissions directly attributable to the lifecycle of the product. This refers to emissions that can be traced to activities directly related to the final product. For instance, a biofuel producer in addition to being responsible for emissions occurring at the biorefinery, is also responsible for emissions at the power plant that produced the electricity used by the biorefinery, the emission at the farm that produced corn, the emissions from the industry that produced the fertilizer used in the farm and so on.
There may also arise emissions that are not directly attributable but indirectly attributable to the activities related to the final product. One indirect link is through market-mediated effects. For instance, an increase in the demand for corn for ethanol increases the demand for land for corn. This may result in conversion of lands such as those enrolled in Conservation Reserve Program to re-enter corn production. Sometimes there may be longer and more complex chain of causation for land conversion. Such displacement of soybean by corn, displacement of grazing by soybean and displacement of forest by grazing.
Current regulations consider only indirect effects from land use change. However, there are multiple indirect effects. The change in agricultural commodity prices that cause land use change also may cause intensification of agriculture and the adoption of new technologies. The introduction of biofuels will also affect fuel markets. Biofuel policies that increase fuel supply will lower fuel prices and increase total fuel consumption and increases total emissions (See Rajagopal et al. 2007, Hochman et al. 2010), a phenomenon that can be termed indirect fuel use change. On the other hand biofuel policies that increase the price of fuel, say a biofuel mandates without a subsidy when the price of oil is low, reduce fuel supply and reduce total fuel consumption will have indirect effects that lower the effective GHG intensity of biofuel.
If we desire consistency in policy making then if indirect GHG emissions are land use change considered in evaluating biofuel, then they should be considered across the board unlike with current regulations such as the LCFS and RFS.
Ex post versus ex ante assessment: Calculations of lifecycle footprint are based on the technical relationship between inputs and outputs prevailing at the time of assessment. These relationships however do not merely reflect technical conditions in a firm or in an industry but also represent implicitly the prevailing economic and policy conditions that led a given firm to choose a certain technology and resulted in observed distribution of technologies across firms within a given industry. Policies on the other hand aim to affect outcomes in the future. When the technical, economic and policy conditions change, current relationships may not be indicative of future environmental performance. For instance, the historical data on agriculture shows remarkable variability in productivity. Periods of high rate of technology adoption such as during the green revolution resulted in large increase in agricultural output with little increase or even a decrease in global agricultural acreage. But during periods of low rate of innovation such as the 1990s before the adoption of genetically modified seed varieties, the land intensity of output was higher. Therefore projections about the future depend on the rate of technical innovation, which can be affected by policy. When historical data shows variability, relying on point estimates for making long-term future projections can be unsound.
Joint production: Another complexity in calculating LCA is under conditions of joint production of more than one output. For instance, the production of ethanol from corn is accompanied by the production of distiller grains a substitute for grain corn in animal feed operations. Similarly crude oil refining results in multiple products such as gasoline, diesel, jet fuel, naptha and other petro-chemicals. Allocating emissions across products is another complex task, which requires taking into account the supply and demand for each product and it substitutes. For a discussion of the different methods of allocating emissions see (Rajagopal and Zilberman 2007b).
Pollution shifts: LCA-based regulation is partial in the sense that it only regulates inputs entering the fuel supply chain. In other words it regulates emissions from corn used for fuel but not emissions from corn used to produce food. Therefore there is a potential for reallocation of pollution by shifting activities between regulated and unregulated activities. So while LCA-based regulation may ensure that “clean” inputs will be used in the production of the regulated product, it may result in no aggregate emission reduction because “dirty” fuel may now be used allocated to production of unregulated output. For instance, a lifecycle based policy that restricts the supply of gasoline from Canadian oil sands to the U.S. because of its higher carbon intensity may lead to diversion of gasoline from light sweet crude intended for China to the U.S. and the diversion of Canadian oil sands to China.
Albeit well-intentioned, the use of lifecycle-based upper-bounds and the accounting of emissions from market-mediated effects from fuels are difficult to implement in practice and for several reasons inconsistent with sound principles of policy making: (1) Calculation of lifecycle emissions rely on complex inter-linkages in the supply chain. These links are not always clear and the cost of acquiring the needed information is likely high. (2) LCA does not assign direct responsibility on polluters. One industry bears the burden for the presumed bad behavior in other industries. (3) These regulations do not weigh the benefit and costs of LCA based approaches vis-à-vis other approaches.
The conditions required for LCA-based fuel GHG regulation to be an effective tool for addressing climate change are stringent and unlikely to be fulfilled because of enormous information requirements. Moreover, accounting for indirect emissions addresses at best one externality of land use change but not others such as biodiversity and loss of ecosystem services such as flood control and water regulation. This suggests the importance of striving for policies that make people pay the social cost of their activities in a simple and transparent way. This involves implementing a carbon tax or cap and trade to address climate change along with payments for preservation of natural habitats or penalties on the contrary.
Regional policies may result in carbon leakages, simply because regulation raises costs in the regulated regions, and diverts trade of carbon intensive goods and commodities to unregulated regions. Then, if demand for these goods remains the same, supply of the carbon intensive commodities and goods may move abroad leaving global emissions unchanged. Carbon leakages also result from not pricing the environment correctly (Hochman et al., 2008). For instance, if land in the Amazon is underpriced, and its price does not incorporate its environmental benefits, then more land will be converted to agriculture than is socially desirable. This converted land emits to the atmosphere all the carbon that it sequestered.
This heterogeneity in policy across countries raises a more fundamental question: Suppose all the challenges from implementing indirect effects are resolved, but some countries elect not to regulate land use, and thus under price nature land. Land is underpriced in unregulated countries. Now, introduce an activity, say biofuel, which takes land away from crops for food and feed. To complicate the story further, assume demand for these crops increases due to growth in developing countries. The new opportunities to benefit from crop production are not left unnoticed, and production of these crops shifts to unregulated countries, at the expense of nature land. Such a shift is very profitable in these countries, because land is not correctly priced – land is cheap to convert. But then, who is responsible for the environmental cost from land use changes? The country that under-priced the environment or the productive activity that increased demand for this unregulated land? Should we penalize productive activity in the regulated country for reallocating land to production in unregulated countries, even if major parts are allocated simply because land is not priced correctly? Should we keep productive activity in the regulated country at bay, simply because other parts of the world do not price the environment? These questions are above and beyond the practical challenges in implementing indirect effect to LCA, and can be further complicated once we allow countries to have heterogeneous preferences regarding the environment.
Acknowledgements and Disclaimer
We thank the Energy Biosciences Institute for funding support. The views expressed herein are those of the authors and do not reflect the position of funders.
Hochman, G., Sexton, S., and Zilberman, D. (2008). “The Economics of Trade, Biofuel, and the Environment,” the International Agricultural Trade Research Consortium Annual General Meeting, Scottsdale, AZ, December 7, 2008.
Hochman, G., Rajagopal, D., and Zilberman, D. (2010). “The Effect of Biofuels on Crude Oil Markets,” AgBioForum
Farrell, A., Plevin, R., Turner, B., Jones, A., O'Hare, M., and Kammen, D., 2006. “Ethanol can contribute to energy and environmental goals.” Science 311 (5760).
Pimentel, D. and Patzek, T.W. (2005). “Ethanol production using corn, switchgrass, and wood; biodiesel production using soybean and sunflower” Natural Resources Research, 14(1).
Rajagopal, D., Sexton, S., Roland-Holst, D., and Zilberman, D. (2007a). “Challenge of biofuel: Challenge of filling the tank without emptying the stomach.” Environmental Research Letters 2(9).
Rajagopal, D. and Zilberman, D. (2007b). “Review of Environmental, Economic and Policy Aspects of Biofuels” World Bank Policy Research Working Paper Series # 4341
Searchinger, T., Heimlich, R., Houghton, R., Dong, F., Elobeid, A., Fabiosa, J., Tokgoz, S., Hayes, D., Yu, T. (2008). “Use of US croplands for biofuels increases greenhouse gases through emissions from land-use change.” Science 319 (5867).
* A Rube Goldberg machine is a deliberately over-engineered machine that performs a very simple task in a very complex fashion, usually including a chain reaction. (Source: Wikipedia)
** Rajagopal (Deepak@berkeley.edu) and Hochman (firstname.lastname@example.org) are with the Energy Biosciences Institute and Zilberman (email@example.com) is with the Department of Agriculture and Resource Economics, all at the University of California - Berkeley.
 Also referred to as Lifecycle analysis and Lifecycle accounting
 Importantly, the baseline for comparison of future biofuels is GHG intensity of gasoline in 2007. Critics argue that this places biofuels at a disadvantage with gasoline getting more carbon-intensive with time because of the transition to unconventional petroleum and lower grade crude oil such as heavy oils.
 These are termed indirect because land may not be converted for direct use in growing of a biofuel feedstock, but may be converted to grow any crop in response to increase in the price land due to biofuels. For instance, allocation of land for corn production for biofuel may occur at the expense of soybean cultivation. As a result returns to land suitable for soybean farming increase, which will lead farmers to switch non-soybean land to soybean. This sequence of events may eventually lead non-farm land to be converted to farmland.