Gunter Festel, Martin Bellof, Martin Wurmseher, Christian Rammer and Eckhard Boles

Abstract The current taxation benefits for biofuels are only temporary. Therefore, biofuel production costs need to be able to compete with those of conventional fuels in order to gain market share in the future. However, highly complex influ­encing factors make a comparison of biofuel production costs with those of fossil fuels challenging. This chapter has three major goals: (1) a projection of future feedstock prices for biofuels based on the development of the price for crude oil, (2) a simulation of the effects of likely economies of scale from scaling-up produc­tion size and technological learning on production costs and (3) a scenario analysis comparing different biofuels and fossil fuels. European biofuel production costs for 2015 as well as 2020 are projected based on a calculation model for biofuel production. Our scenarios assume prices for crude oil between Euro 50 and Euro 200 per barrel for both reference years. Our results indicate that mid — to long-term, second-generation biofuels are very likely to achieve competitive production costs, if technological learning and economies of scale are factored in. Bioethanol made

G. Festel (*)

Festel Capital, Mettlenstrasse 14, 6363 Fuerigen, Switzerland e-mail: gunter. festel@festel. com

G. Festel • M. Bellof

Autodisplay Biotech GmbH, Merowinger Platz 1a, 40225 Dusseldorf, Germany G. Festel • M. Wurmseher

Butalco GmbH, Mettlenstrasse 14, 6363 Fuerigen, Switzerland G. Festel • M. Wurmseher

Department of Management, Technology, and Economics, Swiss Federal Institute of Technology Zurich, Zurich, Switzerland

C. Rammer

Centre for European Economic Research (ZEW), Mannheim, Germany E. Boles

Institute of Molecular Biosciences, Goethe-University Frankfurt,

Frankfurt am Main, Germany

A. Domingos Padula et al. (eds.), Liquid Biofuels: Emergence, Development and Prospects, Lecture Notes in Energy 27, DOI: 10.1007/978-1-4471-6482-1_5, © Springer-Verlag London 2014

from lignocellulosic biomass and biodiesel from waste oil promise the highest cost-saving potential in all crude price scenarios and are capable of outperforming fossil fuels and first-generation biofuels in the future.

Keywords Biofuels ■ Production costs ■ Scale effects ■ Learning curve effects

1 Introduction

The economic dependence on fossil fuels and the potential replacement of crude oil by biomass have been investigated in research chapters. Biomass use for bio­fuels competes with residential applications, heat/power generation and the pro­duction of food, animal feed and other industrial products. This is the reason for negative influences of first-generation biofuel production on global food prices. Whenever crude oil prices rise, the positive correlation between the production scale of biofuels and food prices becomes clearly visible, due to arbitrage effects (Chen et al. 2010). For example, a rising oil price has significantly influenced production volumes, and prices of agricultural grain on a global basis, as the pro­duction of biodiesel and bioethanol from soybeans and corn, respectively, have grown accordingly. The fact that biomass can serve as raw material for chemicals and numerous other applications is not solely the fuel industry that drives prices (Swinnen and Tollens 1991; Hermann and Patel 2007). Due to substitution effects, the price for raw materials is not only dependent on the development of biomass markets but also on the cost of fossil raw materials, such as crude oil. Conversion costs are driven by scale effects as well as time-dependent learning effects.

In order to better understand complex energy production systems under various policy objectives, numerous different calculation models have been developed. Both technical bottom-up approaches as well as macroeconomic top-down approaches have been utilised to describe the entire energy system (de Wit et al. 2010). Other authors evaluate whole supply chains for bio-based products (Stephen et al. 2010; Kim et al. 2011), biorefinery concepts (Fernando et al. 2006; Clark 2007; Francesco 2010) as well as the potential of biofuels for individual countries (Martinsen et al. 2010). For example, a mixed integer linear programming model that allows the selection of fuel conversion technologies, capacities, biomass locations, as well as the logistics of transportation from the raw material locations to the conversion sites and then to the final markets has been established by Kim et al. (2011).

Numerous research chapters have evaluated biofuels, such as biodiesel (Zhang et al. 2003; van Kasteren and Nisworo 2007; Araujo et al. 2010), or simulated biofuel processes with specialised software, such as Aspen HYSYS (West et al. 2008). By contrast, comparisons of one biofuel production process with other production processes (biofuels or fossil fuels) that take scale and learning curve effects into account are rare, even though production costs are imperative to the demand of biofuels. Some studies focus on a single process step, such as enzymes (Tufvesson et al. 2011; Klein-Marcuschamer et al. 2012), while others compare biofuel types through production cost analysis (Bridgewater and Double 1994; Giampietro and Ulgiati 2005; de Wit et al. 2010; NREL 2011). de Wit et al. (2010), for example, show that biodiesel is the most cost competitive type of fuel, dominating the early market of first-generation biofuels. Lower oil crop feedstock prices compared to those of sugar — or starch-containing crops are one of the rea­sons for biodiesel’s better cost performance compared with first-generation bioeth­anol. In addition, capital and operating expenditures for the transesterification of oil to biodiesel are below those for hydrolysis and fermentation of starch to bioeth­anol (de Wit et al. 2010). This chapter intends to calculate the production costs for various types of biofuel in Europe. It will also compare them with production costs of fossil fuels. Raw material, conversion and capital costs are taken into account as well as different scenarios of price development for raw materials and crude oil.

Four steps are central to the developed calculation model in order to analyse and compare biofuel production costs: (1) the definition of biofuel production scenarios in 2015 and 2020, (2) an estimation of future raw material prices based on assump­tions for crude oil price development and the relation between crude oil price and prices for biofuel raw materials which has been observed in the past, (3) the model­ling of scale — and time-dependent costs for capital expenditures and the conversion of biomass and (4) a calculation of the total production costs as a total of raw material, capital and conversion costs. Our model is based on publicly available data for single production process steps and the whole production process. The input data have been collected in expert interviews and intensive literature research during the past 5 years (Festel 2007, 2008). As model output, we have chosen production costs in Euro-Cent per litre, as this is a measure which end users, such as car drivers, can refer to.

Within the next 5-10 years, estimates do not see biofuels gaining a market share larger than 15 % globally (Gnansounou et al. 2009; Bagheri 2011). European Union (EU) targets support this estimation. The EU has set a target market share of 10 % in terms of all petrol and diesel transport fuels by 2020 (EU Commission 2003). That is why future fuel markets prices will still be driven by fossil fuels. Today, it is government regulations and subsidies that enable biofuels to compete with fossil fuels. However, our hypothesis is that government incentives will have a decreasing influence on biofuel demand medium to long term and that demand will be more and more driven by cost competitiveness with fossil fuels through, e. g., new tech­nologies, reduced costs in the production process, improved logistics. In the case that production costs of biofuels will be lower than those of fossil fuels, we expect demand to be high enough to absorb all produced volumes of biofuel. In our model, we do not take the connection between biofuel demand and biofuel prices into account, as the market share of biofuels is determined by its production costs.

In our model, we neglect the option of biofuel import from outside Europe but rather assume that all demand for biofuels within Europe will be met by European biofuel producers. The more developed production infrastructure, economies of scope to other production activities and a close proximity to end users may be a benefit for European production sites. Our input data for production costs are solely focused on Europe. However, our model could easily be adapted to other regions if input is changed accordingly.

Smith and Searchinger (2012) argue that existing life cycle assessments (LCAs) pertaining to biofuels seriously overestimate carbon absorption on the part of bio­energy crops and do not take sufficient account of GHG emissions resulting from the cultivation and harvesting of these crops.

The type of land, i. e. unfertilized grassland, forest land or traditional cropland, used for biomass feedstock is an important determinant of GHGs emitted from the soil (EPA 2006). Preparing fallow or underutilized land for agricultural production usually requires clearing off the majority of the animal and plant species. This can destabilize the soil by releasing significant amounts of stored carbon (EPA 2006). Some studies conclude that conversion of native land such as forest, grassland and abandoned land for biofuel crops leads to carbon debts[15] ranging from one to sev­eral 100 years (Fargione et al. 2008; Gibbs et al. 2008; Fritsche 2008). For exam­ple, Fargione et al. (2008) estimated the carbon debt of producing palm oil on forest land (releasing 3452 tCO2/ha) to be approximately 423 years. Table 5 below provides an overview of the estimated payback periods for a range of biofuels.

In contrast, biofuel crops grown on traditional croplands are less threatening to the environment since they have less embedded soil organic carbon (SOC) (Englund et al. 2011). However, intensive biofuel cultivation, especially if using annual crops, could lead to a substantial release of SOC. This is due to frequent disturbance to the soil (i. e. via tillage), which exposes protected organic matter and increases the rate of mineral decomposition, thereby resulting in lower SOC storage (Grandy and Robertson 2007).

Aside from tillage, farming and irrigation practices could also affect the net carbon balance of biofuels. Mechanized farming or the use of fossil-fuel-powered machinery for soil preparation, sowing, planting, weeding and harvesting activi­ties releases GHGs. Likewise, water for irrigation of biofuel crops is often sourced from rivers, lakes, canals, dams and groundwaters. While this reduces water avail­ability for other uses, it also leads to soil salinization when the irrigation process is poorly managed (Englund et al. 2011). These impacts can be mitigated where rain harvesting systems such as terraces, bunds and small dams are available.

Another issue related to harvesting is mono-cropping, or the planting of only a single species or cultivar. While harvesting a particular biofuel crop on a large-scale over several years makes the process more economical, it can also increase the environmental footprint. Repetitive harvesting of a single variety of crop results in a

lack of biodiversity and a decline in soil fertility. To control pests and maintain yields in such environments, more chemical input and fertilizers are generally applied (Englund et al. 2011), which can lead to serious ecological impacts (more in Sect. 5 in A Comparison Between Ethanol and Biodiesel Production: The Brazilian and European Experiences). However, as Dale et al. (2010) report, such impacts can be minimized by adopting sustainable land management practices.[16]

Studies of LCAs have shown that GHG emissions can vary substantially between biofuels, but are mostly lower than those associated with conventional fossil fuels. Through a meta-analysis of LCA literature, Davis et al. (2008) found that the results range between -89 MgCO2 per hectare per year for corn-based biofuel (Farrell et al. 2006) to 9.6 MgCO2 per hectare per year[17] for biofuel pro­duced from switchgrass (Searchinger et al. 2008). Results also varied between authors for biofuels produced from the same crop. For example, Shapouri et al. (2002) found that corn ethanol reduces CO2 emissions by 1.2 Mg per hectare per year, while Delucchi (2006) determined that it increased CO2 emissions by 5.14 Mg per hectare over the same period. Some studies reported the results in terms of change in GHG emissions compared to fossil fuels. The variation in this case was once again large and ranged between -114 % for switchgrass (Adler et al. 2007) to 93 % for corn (Searchinger et al. 2008). LCAs, however, have often overlooked the impacts of LUC on overall GHG emissions. When Bailis and Baka (2010) compared biodiesel from Jatropha in Brazil with conventional biodiesel without considering LUC, they noted a 55 % reduction in GHGs. In contrast, when they included LUC, the net emissions were estimated to increase by 59 %.

Despite providing a cradle-to-grave assessment, LCAs therefore reach varying conclusions on any biofuel depending on the methodological approach adopted. While using an LCA should ideally be an ongoing process for handling and prior­itizing information as new data comes to hand, it is worth noting the “seven grand challenges” that McKone et al. (2011) identified for undertaking a comprehensive LCA of biofuels. These are

• Understanding farmers, feedstock options and practices.

• Predicting biofuel production technologies and practices.

• Characterizing tailpipe emissions and their health consequences.

• Incorporating spatial heterogeneity in inventories and assessments.

• Accounting for time in impact assessments.

• Assessing transitions as well as end states.

• Confronting uncertainty and variability.

A proper understanding of these issues will have profound implications with respect to what feedstocks should be used for biofuel production, together with what lands are most suitable for environmentally sustainable feedstock produc­tion. Any conclusion reached from an LCA must consequently be tempered by the knowledge that the same assessment could provide a different result at another point in time.

The cost of producing biodiesel depends on a number of factors, including the feedstock used in the process (i. e., the production cost of biomass), the capi­tal and operating costs of the production plant, the current value and sale of by­products, and the yield and quality of the fuel and by-products. Table 8 provides total and unit production costs of a representative European biodiesel plant (Italy) using rapeseed oil as feedstock (2010), which is a good example that includes the average characteristics of Italian plants, on the base of the information collected through firm survey (Finco 2012). The plan has capacity for 150,000 tons and pro­duces 150,000 tons of biodiesel.

Table 8 shows that the major economic factor to consider for input costs of bio­diesel production is the feedstock, which is about 80 % of the total production cost. This means that the market trend commodities prices highly influence the result of the biodiesel industry. In particular, feedstock costs can vary significantly from region to region due to their availability and market fluctuations, which can also make biodiesel production costs vary over time. Vegetable oils prices have changed significantly in the last 5 years. The prices have been rather stable until end of 2006, while from 2007 to 2008, they are more than doubled, declining again in 2009 reaching the 2006 level. In the second semester of 2010, the price registered another increase followed by a slight fall in 2012 (OECD-FAO 2012).

Table 9 shows the net margin of our representative plant. Nowadays, our plant perceives a negative economic result because revenues do not cover production costs. This result is mainly driven by the biodiesel price that is fixed by the refiner­ies and it is not connected with the production costs.

There are two components that influence the value of biodiesel: the diesel price on Platts and a premium price. The premium is determined by the refinery indus­try, and it depends on the vegetable oils price and the contractual power of the biodiesel plant. Technically, the premium price should correspond to the difference between the production costs and the diesel price on Platts, which biodiesel pro­ducers widely call the ‘business margin.’

However, according to the data from biodiesel plants, the premium price per­ceived corresponds to approximately 65 % of the ‘business margin.’ Moreover, this percentage depends on the policies adopted by the Governments, such as tax excise reductions or subsidies.

It is important to underline that biodiesel plants use a blend of vegetable oils and, consequently, the price can probably be lower than the rapeseed oil price that was used in the Table 9. Taking this into account, the results present an accurate representation of the Italian biodiesel industry.

However, the increased price of vegetables oil, the economic crisis, and policy changes at European level had negative impact on biodiesel production. For exam­ple, in Italy, the reduced tax exemption in 2009 and the subsequent abolition has diminished the profitability of the biodiesel plant realizing losses.

Lastly, the third attribute refers to the specificity of the assets involved in the trans­action. Assets are specific if their return depends on the continuity of a specific transaction. The more specific the asset, the greater the agents’ dependence on achieving the negotiation and therefore the greater the loss from an opportunistic behavior by one of the parties.

Williamson (1985) also proposes classifying the different ways a given transac­tion is performed, starting with the spot market, continuing with long-term con­tracts and concluding with the hierarchy (a single firm securing the transaction in question). If the asset specificity is null, the TCs are negligible, requiring no con­trol over the transaction; therefore, the spot market would be more efficient than other organizational forms. If, instead, the asset specificity is high, the costs asso­ciated with breaching the contract will be high, which would imply greater control over the transactions.

Also according to Williamson (1981), asset specificity is the most important critical dimension, as it is related to the type of investment. Thus, after perform­ing the specific investment, the seller and the buyer will operate in a bilateral exchange relationship for a considerable period of time (irreversibility cost). Williamson (1991a, b) discriminates six types of asset specificity:

a. locational: those whose application in a given transaction generates cost sav­ings in transport and storage, meaning specific returns to these productive units;

b. physical: those more suitable for a specific purpose (e. g., specific inputs for the production of a specific product);

c. human: related to the use of specialized human capital for an activity. This type of specificity is related to accumulated knowledge by the continuous execution of a particular activity;

d. dedicated: specific assets for a given transaction (e. g., to service a specific customer);

e. brand: refers to capital—not physical or human—manifested in a company’s brand, which is particularly relevant in the franchising world; and

f. temporal: refers to the value of the assets related to the period when the trans­action is processed. Thus, this asset becomes especially relevant in the case of negotiating perishable products.

According to Azevedo (2000), as it is not possible to determine a relationship that contains all eventualities, in some cases, renegotiation is inevitable. However, as an opportunistic behavior is a possibility, this renegotiation is subject to one of the parties taking advantage of the gains, which in turn results in losses to the other party. Thus, in economic transactions, based on the issue of opportunism, one side could try to take advantage of the other due to the impotence of predicting future events. Hence, agents often have to resort to safeguard contracts, which in turn contribute to increase some TCs.

There are some forms reported in the literature that enable controlling the prob­lems of post-contractual opportunism, namely increase the resources to monitor transactions, reduce information asymmetry, and adopt contractual incentives rewarding the agents’ compliance or good performance. The vertical integra­tion itself can eliminate conflict of interest, especially in transactions between an organization and its suppliers, reducing TC, though this integration could increase operating costs (Bonfim 2011).

On account of the intrinsically qualitative competitive process, the literature generally does not address the governance structures and the theory of competi­tiveness. This supposes, mistakenly, that the coordination of supply chains occurs efficiently or that more efficient structures through mechanisms associated with competitive rivalry are used (Farina 1999).

Coutinho and Ferraz (1995) pointed out that strategies are the basis of the dynamics of competitiveness, which seek to expand and renew the companies’ capacity required by the standards of competition (or “rules of the game”) in the market they are embedded.

Buainain et al. (2007) deem that competitiveness will only be achieved by including practices that encourage cooperation between the economic agents of a supply chain, including the government. According to the authors, considering that a company’s competitiveness is linked to the system it is inserted in could mean significantly changing the way such company views and manages its business. Thus, the authors emphasize the importance of vertical and horizontal manage­ment within a system to gain competitiveness. According to Buainain et al. (2007), a serious problem is the lack of works and experiences that report the problems of internal management in the family farmers’ network, as well as the relationship between them and their customers and suppliers.

Thus, competitiveness is reflected by these companies’ greater or lesser ability to adopt governance structures that reduce TC, enable greater integration with the agricultural production, and set conditions for systemic competitiveness (Batalha and Souza Filho 2009).

Since 1850, corn and beets crops have been used as raw materials for ethanol pro­duction in the USA. Ethanol was a popular fuel for lighting during the first half of the nineteenth century, and in 1860, 13,157,894 gallons of ethanol were burned in the USA for lighting (Herrick 1907). The ethanol tax initially imposed a fee of 20 cents per gallon in 1862 and reached $2.08 per gallon in 1864 (Herrick 1907).

Between 1919 and 1933, ethanol was forbidden to increase the demand for products such as gasoline. During World War II, the production of ethanol rose to 600 million gallons per year. The US interest in ethanol fuel has grown since the oil crisis of the 1970s. The USA started using 10 % ethanol blended into gasoline at the end of 1970. The Energy Tax Act of 1978 (ETA) officially defined 10 % as the required volume of a non-fossil-fuel blend with gasoline (Solomon et al. 2007).

In addition, the demand for ethanol produced from corn has increased since the main product that was added to gasoline at one time (namely, methyl tert-butyl ether, or MTBE) was revealed as a contaminant of groundwater. The use of MTBE in gasoline was banned in almost 20 American states in 2006, and since then, etha­nol has become its main substitute (RFA 2011).

Today, the sharp growth in the production and consumption of ethanol is asso­ciated with federal legislation that was created to reduce oil consumption, increase energy security, and reduce CO2 emissions in the country. From 1983 to 2005, the production costs for making ethanol from corn decreased by 65 %, and further­more, the industrial processing costs decreased by 45 % (Hettinga et al. 2009).

The use of ethanol has been expected to expand since the Energy Policy Act of 2005 established a production target of 7.5 billion gallons of renewable fuel by 2012. The Security Act of 2007 raised this target and required the annual use of 36 billion gallons of renewable fuels until 2022 (RFA 2011). With these incentives and the maturity of the industry, the USA is currently the world’s largest producer of ethanol and it represents approximately 60 % of the world’s production.

In addition to ethanol, biodiesel is currently used as a biofuel in the USA. The two major feedstocks for biodiesel production are soybeans and rapeseeds. The fed­eral government plays a key role in determining the course and especially the scale of biodiesel development, and it gives incentives such as tax exemptions, price con­trols, production targets, and direct subsidies (Lin et al. 2011).

Advances to stimulate biodiesel were proposed by Congress, and President Bush signed the Energy Independence and Security Act of 2007. The scale of production has grown significantly, and furthermore, plants are now distributed in various parts of the country. Today, the total production of biodiesel is nearly 1 billion of gallons.

In some parts of the USA and Canada, camelina (Camelina sativa (L.) Crantz) is emerging as an oilseed feedstock for biodiesel that is intended for use in avia­tion fuel, as it can be grown on wheat fields that would otherwise be left fal­low without harming the soil. In fact, growing camelina in these fields usually improves their fertility. The USA has announced plans for using algae as a feed­stock for future generations of biofuels and is promoting the biofuel industry by providing grants and sponsorships.

Our initial analysis examines ethanol production volumes between now and 2040. We consider four different scenarios altering the presence of the EPA RFS 2 man­dates (EPA 2009) and market penetration costs. Figure 2 presents ethanol production

volumes with and without penetration costs under a mandate. When a drop-in type fuel is produced avoiding market penetration costs, we find increased ethanol mar­ket supply. Our second comparison analyzes the impact of penetration barriers when mandates are not present (Fig. 3). There without a mandate, less crop ethanol is pro­duced than with mandates and more cellulosic ethanol is produced after 2020, mainly due to lower processing costs. Furthermore, the removal of market penetration costs has a stronger impact on ethanol volumes reflecting the greater flexibility allowed.

Next, we examine the impact of adding carbon prices. First, we consider the case with no mandates in place, but with market penetration barriers (Fig. 4). There total ethanol production volume increases slightly under increasing car­bon price, and ultimately reaches about a 10 % increase in total production. Simultaneously, cellulosic ethanol replaces crop ethanol production due to its enhanced GHG emission offset efficiency (see McCarl and Sands (2007) for esti­mated offset rates).

We also examine the projections of future ethanol production under “no mandates and no penetration costs in place” scenario for three points of time (Fig. 5). It can be observed that removal of market penetration costs drives ethanol volumes up. In 2020, the amount of total ethanol produced fluctuates between 25 and 30 billion gal­lons per year (depending on the GHG price); in 2030, the amount of total ethanol is in the range of 30 and 35 billion gallons per year; and in 2040, the total etha­nol amount reaches 40 billion gallons per year under GHG price of $50 per ton of CO2e. GHG payments provide additional revenues and increase ethanol volumes. At the same time, we see that only under scenarios with a carbon payment and no penetration costs does the total volume of ethanol produced reach the RFS2 biofuel mandate levels. Thus, it appears that in the absence of carbon trading schemes or a drop-in fuels it is highly unlikely that the EPA RFS2 mandate will be ever met.

Finally, we examine impact of penetration costs removal on the total volume of ethanol produced. At the same time, we assume that scenario with no penetration costs could be a case of all drop-in biofuels which do not require adjustment in infra­structure before their distribution in the market. Some innovative liquid biofuels, like butanol or methanol, are free from corrosive properties and they could be distributed and sold to the end-consumer through the currently existing distribution networks and pumping stations. As one can observe in the Fig. 6, removal of penetration bar­riers raises the total ethanol production by around 5 billion gallons per year unde

Fig. 5 Projection of future crop and cellulosic ethanol production under varying GHG prices (at three points of time) for “no mandates and no penetration costs in place” scenario all considered GHG prices in 2020. In 2030, the situation looks slightly different. Under $0 carbon price and scenario with no penetration costs the amount of ethanol produced is around 6-7 billion gallons higher than under the scenario with penetra­tion costs in place. However, under $100 carbon price this difference between two scenarios amounts to 10 billion gallons per year. Some discrepancies could also be noticed in the projections for 2040. Under $0 carbon price, the total amount of etha­nol under no penetration costs exceeds the total amount of ethanol under scenario with penetration costs by 10 billion gallons per year. However, once the carbon price reaches $100 per ton of CO2e, the gap between both scenarios amounts to almost 15 billion gallons per year. In general, these projections display the pattern which reflects the impact of penetration costs removal on the amount of total ethanol pro­duced. Clearly, removal of penetration barriers enables ethanol to be absorbed by the market and encourages growing consumer ethanol demand. On the other side, exist­ence of penetration barriers and lack of investments aiming at their reduction might hamper further development of the ethanol industry.

As experience has shown, the production of lignocellulosic ethanol in the USA has not been launched on the industrial scale so far. It is believed that more techno­logical progress is needed to lower processing costs of cellulosic ethanol production. Enzymes used for fermentation have to become cheaper and biochemistry of reactions

needs to become more efficient. So far, cellulosic ethanol production is limited to operations in pilot plants; therefore, it is difficult to estimate processing costs of cellu — losic ethanol per gallon. Uncertainty related to potential location of cellulosic ethanol plants makes it challenging to assess feedstock and other materials costs, transporta­tion costs, and capital costs related to cellulosic ethanol production. Until now, one of the available projections is the estimation made by the National Renewable Energy Laboratory (NREL) which presents $3.29 per gallon as a viable unitary process­ing cost (see Fig. 1 for the specific estimation of cost). In our next analysis, we try to analyze the impact of further cellulosic ethanol processing cost reductions to see how much the processing costs have to fall from $3.29 per gallon level for cellulosic ethanol production to become economically profitable. We also look at what level of processing costs the volume of ethanol produced approaches volumes contemplated by the Energy Independence and Security Act and the EPA RFS2 mandates. In this analysis, we hold crop ethanol production costs constant. Figure 7 shows the effect of decreasing processing costs on production volumes for three different points in time.

Decreasing processing costs are found to reduce crop ethanol and increase cel — lulosic ethanol volumes in the absence of mandates. For example, we find that 50 % cost drop in processing costs of cellulosic ethanol causes amount of crop ethanol produced in 2020 to drop to around 5 billion gallons per year and in 2040 to drop to around 1 billion gallon per year (Fig. 7, panel a). These amounts are much smaller compared to current levels of crop ethanol of around 13-14 billion gallons.

At the same time, the 50 % cost drop causes the volume of cellulosic ethanol to increase from 13.3 billion gallons per year to 19 billion gallons per year in 2020. It is also worth mentioning that RFS2 mandate schedule requires cellulosic ethanol to be produced at the level of 16 billion gallons per year by 2022. From the projec­tion of cellulosic ethanol production in 2020 in Fig. 7 (panel a), we observe that this volume is only achievable under 25 % decrease in processing cost. When it comes to the total ethanol volume, the EPA 2022 mandate of 31 billion gallon per year is never achieved, even if the processing costs drop by 60 % (Fig. 7, panel a).

2 Conclusions

We find that ethanol mandates create volumes that are generally higher than would occur in the free market and that market penetration costs and carbon prices are big influences in ethanol market penetration. Namely:

• positive carbon prices, lower infrastructure costs, or some other cost reduction are needed to provide economic incentive for second-generation liquid biofuels production if they are to reach mandated levels;

• technological progress is essential to reduce processing costs and, thus, produc­tion costs of the second-generation biofuels and to make the second-generation biofuels cost competitive.

Biofuels have often been seen as a way to enhance the agricultural sector. This is especially the case in the developed world, where locally produced food crops find it increasingly difficult to compete at a global level because developing and underde­veloped nations produce the same at a much lower cost. In these cases, governments provide considerable subsidies, promote low-interest loans and impose various trade barriers to incentivize farmers to produce these crops at a competitive price and thereby sustain their agricultural sector. Given that biofuels, especially first-gener­ation biofuels, rely on edible crops as a feedstock, they create an alternative market for such agricultural products as a valuable input for the energy sector. In this sec­tion, we look at the degree to which rural economies, where farming is the livelihood for most people, are influenced by the burgeoning biofuel industry.

One of the central arguments in favour of biofuels is its contribution to rural development through increased employment opportunities and higher income. It has been estimated that the biofuel industry requires approximately 100 times more labour than the capital-intensive fossil fuel industry to produce the same energy output (Renner and McKeown 2010). This is because there is a wider array of jobs associated with biofuel production. These positions can relate to farming through to biotechnological research. Scaramucci and Cunha (2007) estimated that more than 5 million jobs could be generated in Brazil by the year 2025 if 5 % of global gasoline demand is replaced by sugarcane-based bioethanol from Brazil. Jobs also result from indirect employment, such as those involved in the sales of biofuels and transport of biomass. In 2006, all types of biomass operation in the United States employed about 136,999 people directly and another 310,000 across the supply chain (Domac et al. 2005). While the numbers are substantial, rational­izing pro-biofuel policies simply based on potential job creation can be problem­atic. This is because the net economic benefits depend on a multitude of factors.

For example, production capacity and level of mechanization can influence the scope for job creation. While a heavily mechanized production system increases labour productivity, it also minimizes employment opportunities. Likewise, a large refinery may achieve higher economies of scale, but the number of workers required per unit of output is low. Brazil’s policy to control the rate of mechaniza­tion and provide support for small-scale refineries has assisted with controlling unemployment and poverty in the region (APEC 2010). In 2006, 351 plants were able to provide employment for approximately 700,000 people to produce 17,900 million litres of ethanol from 5.9 million hectares of land. In this context, the Brazilian Social Fuel Seal (Selo Combustfvel Social)[6] initiative, which supports biofuel producers through tax incentives, is worth mentioning here as it promotes diversification of jobs within biofuel-producing regions and encourages the ongo­ing participation of family-based feedstock production firms in the nation’s biofuel industry (Padula et al. 2012). However, large-scale production is crucial for biofuels to compete with fossil fuels (DfID 2007). This may negate the expectations of regional development emanating from the biofuel industry. Indeed, potential bene­fits from new or expansion of existing biofuel facilities are often overestimated. This is because refinery building or expansion provides construction-related jobs to those generally living outside the local area. As a result, most of the initial impact is not felt locally (APEC 2010; Hillebrand et al. 2006; Moreno and Lopez 2008).

Net employment may also vary depending on the land displacement effect. Switching from existing food crops for biofuel production does not always result in additional employment (Jaeger and Egelkraut 2011). Rather, it simply exchanges one market for another. With regard to the impacts of biofuel policy on employment, analysis based on dynamic and long-term general equilibrium adjustments, includ­ing shifts in jobs in agriculture among biomass-producing regions, has found that biofuel policies would not provide any additional economic activity. This is because the increase in bioethanol output would be offset by a reduction in livestock pro­duction (Dicks et al. 2009), especially because land-use changes take effect. Furthermore, de Gorter and Just (2010) claim that higher fuel prices induced by bio­fuel subsidies magnify the inefficiency of the preexisting wage tax by reducing real wages and thus discouraging work. This would reduce labour supply and generate deadweight costs because the tax base becomes eroded as consumers move away from the taxed good and use substitutes. On the contrary, if the land used for bio­fuel production was not in use or was abandoned, any job created would potentially increase net employment and foster economic growth (Diop et al. 2013).

As with employment expectations, it is perceived that biofuels increase the income levels of those engaged in the industry. Parcell and Westhoff (2006) found that, in 2006, the average annual salary of ethanol-related salary was much higher than the average US salary. However, this may not always be the case as earnings and job security can vary significantly across a number of factors. Skilled labour working in technical roles has a much higher income potential than unskilled labour working in the field or in the refinery. In fact, there are fewer white-collar jobs compared to blue-collar jobs. Depending on the type of feedstock, employ­ment opportunities may vary. In the case of Brazil, the high seasonality of sug­arcane production means that the ratio between the number of temporary and permanent workers is significant (DfID 2007). As a result, many workers do not have a biofuel job throughout the year. Failures of biofuel projects are becoming increasingly common, and these failures adversely affect the livelihood of many vulnerable farmers in regional areas (APEC 2010).

While one objective of biofuel policies is to help farmers, landowners stand to benefit the most from increases in crop prices. Crop growers who lease land there­fore only benefit until higher profits associated with rising feedstock prices are captured by higher land values and land rents. Take corn for example. Though dis­puted by Ajanovic (2010), as corn prices rise, domestic pork and poultry producers reliant on this crop to feed their livestock will potentially reduce their international competitiveness, thereby causing a reduction in production levels if higher prices are not absorbed by consumers (Brown 2008). Although the flow of profits from these facilities may initially stimulate rural economies, a rise in crop prices over time owing to demand has the potential to minimize these benefits. There will also potentially be a reduction in livestock farming in these same areas (Dicks et al. 2009), especially as land-use changes take effect. This could eventually work to offset this advantage.

To understand how the biofuel industry has influenced rural development, we look at the employment data of three major biofuel markets, these being the United States, Brazil and the EU (it must be understood, however, that income may vary significantly within the sector itself). If one takes into account that abso­lute numbers of employment may only tell part of the story, unemployment and employment data in the agricultural sector are presented in the form of percentage of total labour force and of total employment, respectively. As can be observed from Fig. 4, bioethanol production/consumption does not seem to have increased employment in agriculture in the United States. Employment in agriculture is relatively stable during the observed period, despite the substantial increase in domestic biofuel production, and has even slightly declined. With respect to the overall impact on employment, the unemployment rate has increased in recent years. Figure 5 illustrates the case for Brazil. Once again, bioethanol production/ consumption has not had the effect of increasing employment in the agricultural sector. Indeed, the employment in agriculture has declined significantly in recent

bioethanol production (‘0000 barrels per day )

times, even though biofuel production/consumption has increased sharply. The reason may be that a greater use of mechanical harvesting has resulted in fewer jobs being generated. Yet there seems to be some positive impacts on overall

employment as a drop in the unemployment rate has been observed since 2006. As in the United States and Brazil, biodiesel production/consumption does not increase employment in agriculture in the EU. Like the United States, employment in agriculture has also slightly declined, despite a significant observable jump in biofuel production and consumption. Furthermore, biofuels seem to have a neutral impact on overall employment (Fig. 6).

So, despite the fact that first-generation biofuels use crops currently grown by farmers within the respective domestic biofuel markets investigated, there is no clear overall benefit with respect to the number of people employed in the agricul­tural sector. While jobs are obviously being created in terms of biofuel processing, the same positive effects do not seem to flow through to the agricultural sector in the economies discussed.

The observations made above have significant implications. As it is eventually realized that more sustainable forms of biofuel production beyond first-generation processes are necessary, this will arguably also have significant impacts on local or regional economies reliant on the growing and processing of particular feed­stocks. In many cases, food crops currently being used for biofuel production will not be optimum for later-generation bioethanol production, which can use all man­ner of biomass (Blottnitz and Curran 2007). Once demand for biofuels grows, the cost equation of producing biofuels from these less energy-intensive crops will undoubtedly force producers to look for crops that can produce the most energy at the least cost (McCormick-Brennan et al. 2007). In many cases, this might mean that regions currently producing biofuel feedstocks will not be well placed to grow the preferred types of biofuel crops. This will clearly have detrimental impacts on economies that are closely tied to long-held agricultural traditions, especially if market conditions continue to militate against their ability to compete with other economies in the open food market. Yet this might be precisely the reason why governments continue to support first-generation biofuels, for moving to later — generation processes brings with it the spectre of moving from labour-intensive to more technology-based production.

When production costs are analysed, scale economies and technological learning play an essential role (de Wit et al. 2010). A commonly used concept to meas­ure and quantify effects of technological learning implies fixed percentage cost declines every time when the cumulative production doubles. This concept is called the experience curve approach (Hettinga et al. 2009) and has been imple­mented in numerous energy models. However, thus far, only few authors have developed models with particular focus on biofuels, e. g. de Wit et al. (2010). Just as in studies on biomass-integrated gasification/combined cycle (BIG/CC) plants for the production of electricity (Faaij et al. 1998; Uyterlinde et al. 2007), learning curve effects can be implemented into our model. Therefore, progress ratios for distinct process steps of biofuel production were estimated.

When investigating the cost-efficiency for biofuels, production scale size (scale effects) and technological improvements of the process (learning effects) need to be considered. While the learning effects are very dynamic and will improve over time (typically with decreasing pace), scale effects are rather static. However, the latter may also have a dynamic component in the case that production capaci­ties increase over time. Numerous studies have attempted to differentiate between static scale economies and dynamic learning effects (Stobaugh and Townsend 1975; Sultan 1975; Hollander 1965; Preston and Keachie 1964) and, in general, the studies have discovered static scale economies to be statistically significant but small in magnitude relative to learning-based effects (Lieberman 1984).

Scale effects are based on scale law, which describes an inverse correlation between decreasing production costs resulting from increasing plant sizes (Blok 2006; Haldi and Whitcomb 1967). Up to a point, larger production scales are asso­ciated with decreasing marginal costs per unit and thus decreasing average costs per unit of biofuel outcome. However, transport costs have to be considered, which leads to an optimal production scale for each production facility.

In order to determine the optimum plant size, specific characteristics of differ­ent types of biofuels need to be considered. For bioethanol, Nguyen and Prince (1996) show that capital costs per unit of product can be reduced if mixed crops are used in order to extend the length of the processing season. This leads to lower production costs and results in a smaller optimum plant size.

Further, cost-reduction potential can be realised through technological advance­ments and other learning benefits related to the production process. de Wit et al. (2010) state examples, such as a more efficient organisation of production and transportation processes, the use of advanced materials and lifetime prolongation of catalysts. Various studies have examined and proven the significance of these learn­ing-based cost-reduction components. Hettinga et al. (2009), van den Wall Bake et al. (2009) and Hamelinck et al. (2005) have shown that this type of cost-reduction potential for bioethanol made from corn or sugarcane ranges from 25 to 50 %.

The examples show that a top-down model can be useful for producers, investors and policy makers, because it helps to easily understand production costs for fos­sil fuels and biofuels without a strong focus on technical details. Based on different

feedstocks and conversion technologies, this chapter intends to contribute to the discussion through the development of a simple top-down calculation model which compares the production costs of biofuels to fossil fuels in Europe for 2015 and 2020.

Agriculture has brought about widespread environmental degradation, with this degradation intensifying as it became increasingly mechanized and reliant on man-made inputs. It is therefore important to bear in mind the potentially negative impacts that intensified farming practices will have on ecosystems in areas where organic material for biofuel production is being grown, often where less-intensive agricultural methods have previously been employed, such as in the developing world. This is because the exploitation of biofuels, being made from organic mate­rial, has the potential to have a number of significant negative effects on ecol­ogy. In this section, the impact of biofuels from an ecological perspective will be explored, particularly as a result of unregulated use of (1) herbicides, pesticides and fertilizers, and (2) land-use changes.

4.1 Broader Impacts of Herbicides, Pesticides and Fertilizers

The growing of crops is made problematic by pests, which can destroy whole fields of crops. To enhance productivity and ensure saleable crops in sufficient volumes, pesticides are regularly used to keep unwanted organisms under con­trol. Yet pesticide run-off from agricultural land has the potential to pollute local watercourses and can result in a loss of biodiversity when food supplies for higher organisms are reduced (Charles et al. 2009). This is especially the case with respect to insects, many of which are highly detrimental to agricul­ture, but are of vital importance to creatures higher up in the food chain. The presence of harmful chemicals found in pesticides can also flow throughout food chains, thereby leading to chemical build-up in higher organisms, espe­cially avian fauna, and raptors in particular, the strength of whose eggs is affected by chemicals such as DDT, thereby leading to greater infant mortality (Sodhi et al. 2011).

Furthermore, production processes and distribution relating to pesticides, her­bicides and fertilizers can also contaminate water supplies, something of especial importance in developing nations such as India, where clean, disease-free reticu­lated water might not be readily available to all citizens (Rajagopal 2008). With respect to water, Eisentraut (2010, p. 10) notes that, while water requirements for second-generation crops could be less than for first-generation biofuels, depending on the crop type and local environmental conditions, the total demand for water could be higher owing to additional water treatment steps in the production pro­cess. In addition, run-off from nitrogen-rich fertilizers can profoundly increase the incidence of algal bloom in freshwater aquatic environments. Such outbreaks can result in these watercourses becoming starved of oxygen, while the presence of thick layers of light-seeking algae at the surface of the watercourse can block out sunlight and impede the ability of other plants to photosynthesize effectively (Bergkamp et al. 2000). The result can be a catastrophic loss of biodiversity in affected watercourses, particularly when organisms at the lower end of the food chain are threatened. In many cases, watercourses have become depleted of oxy­gen owing to algal blooms, thereby creating hypoxic or ‘dead’ zones where only a few organisms can survive (Dale et al. 2010).

The use of fertilizers can also result in the acidification of soils (Eisentraut 2010). The phosphorus, nitrogen and potassium contents in fertilizers, when they become dissolved in water, alter the pH balance of the soil. High acidity reduces the functioning of nitrifying bacteria responsible for the breakdown of organic matter into ammonium and nitrate for plant uptake. This might not be problem­atic for biomass cultivation, depending on the species being cultivated, but could have disastrous impacts in adjacent areas affected by fertilizer run-off. Finally, the very production of fertilizer and its distribution are generally energy inefficient. In most countries, it is reliant on carbon-based forms of energy and therefore contrib­utes to GHG proliferation, all of which, as noted previously, needs to be taken into account in an overall LCA.

The concept of sustainability is derived from ‘sustainable development,’ which has been defined in the Brundtland report as ‘development that meets the needs of the present generation without compromising the ability of future generations to meet their own needs’ (WCED 1987: 45). The concept of sustainable development has traditionally focused on three pillars (i. e., social, environmental, and economic), and in recent years, it has evolved including other components such as policies and institutions (Diaz-Chavez, 2011).

The EU, since first announcing its intention to set a mandatory biofuels target, has maintained that any production or use of biofuels must be sustainable (European Federation for Transport and Environment 2009). The Renewables Directive (2009/28/EC) aims to ensure this ambition is met through the use of mandatory sustain­ability criteria. The criteria set out three main requirements which biofuels must meet in order to be counted toward the target or to be eligible to receive tax rebates or subsidies:

• The greenhouse gas emission savings from the use of biofuels and bioliquids must be at least 35 % (rising to 50 % in 2017) compared to fossil fuels;

• The feedstock of biofuel is not to be derived from land with high biodiversity value such as high biodiversity grassland; and

• The feedstock of the biofuel is not to be derived from land with a high carbon stock. These criteria apply to biofuels and bioliquids and for both imported and domestically produced feedstock.

A significant part of biofuels debate since 2009 focused on indirect land — use change and its exclusion from the EU sustainability criteria. ILUC is not accounted in the Renewables Directive, and therefore, the emissions resulting from ILUC are not included in the greenhouse gas life cycle analysis calculations (Amezaga et al. 2010).

For biofuels, the length and complexity of the supply chains make the sustainabil­ity issue very challenging. Biofuels’ pathways include several successive segments over the fuels life cycle (e. g., feedstock production, conversion of the feedstock to biofuels, wholesale trade, retail, and use in engines) and multiple actors (e. g., feed­stock suppliers, biofuels producers, biofuels consumers, and public authorities).

In order to be sustainable, biofuels should be carbon neutral, especially consid­ering the necessity of fossil fuel substitution and global warming mitigation. Also, biofuels should contribute to the economic development and equity. Moreover, they should not affect the quality, quantity, and use of natural resources as water and soil, should not affect biodiversity, and should not have undesirable social consequences (Lora et al. 2011).

Several authors have recently raised concerns about the environmental costs benefits and social implications of biofuels production such as underlying uncer­tainties over the life cycle emissions of greenhouse gas emissions (GHG), pos­sible deforestation for feedstock production, degradation of soil and air quality, increased water consumption, possible loss of biodiversity, possible competi­tion with food production, and other potential social imbalances (Ajanovic 2011; Gnansounou 2011; Finco et al. 2012; Padella et al. 2012).

Land-use change is considered one of the most important environmental impacts to address, mainly because of its impacts on GHG and wider ecosystems. Recently, many studies are working on land use, direct and indirect (LUC, ILUC). For exam­ple, the research studies of Brazil show that the amount of new land required for sugarcane production would be relatively small (Arima et al. 2011; Macedo et al. 2012). In the same way, the LUC module based on a transition matrix developed by Ferreira Filho and Horridge (2011) and calibrated with data from the Brazilian Agricultural Censuses of 1995 and 2006 shows how land use changed across different uses (crops, pastures, forestry, and natural forests) along those years. The results obtained by general equilibrium models approach show that the ILUC effects of ethanol expansion are of the order of 0.14 ha of new land coming from previously unused land for each new hectare of sugarcane. This value is higher than the values found in Brazilian literature (Ferreira Filho and Horridge 2009, 2011).

Careful assessment of these impacts has given rise to criticisms from economists, ecologists, NGOs, and international organizations, who call for additional analysis of biofuels’ effects. Furthermore, the European Union and several countries have adopted certification schemes to biofuels to respond to these growing concerns and to address the sustainability issues derived from the expanding production of biofuels.

Current and future biofuels production could have important environmental and ecological impacts. One of the major reasons for producing biofuels is to reduce greenhouse gas emissions and to mitigate the effects of global warming produced by fossil fuels. However, some unintended impacts of biofuel production are land, air, water, and biodiversity.