The prices for feedstocks are critical for the economically viable production of bio­fuels. In addition to raw material prices, crude oil as key competitor product also influences the profitability of biofuels. Prices for both are interrelated. Increasing oil prices tend to fuel demand for alternative sources of energy and thus the prices for raw materials. A positive correlation between the prices for crude oil and global grain commodities has been demonstrated in a model by Chen et al. (2010).

In order to project raw material prices for biofuels, we analyse the relation between the price of biofuel raw materials (pB) of type k (maize, wheat, rapes oil, palm oil and wood) and past crude oil prices (pO) while also considering other major drivers of raw material prices, including a price index for agricultural prod­ucts (pA), growth in world population (POP), growth in wealth (per capita income: GDP/POP), change in energy consumption per capita (EN/POP) and global infla­tion (pGDP). The linear regression model to be estimated reads as follows:

pBk, t = a + ei, kpOt + P2,kPAt + e3,kpGDPt + xi, k POPt + X2,k GDP/POPt

+ X3,kEN/P°Pt + &k, t > (1)

with t being a time index for months, a being a constant, в and x being parameters to be estimated and є being a time and k-specific error term. We take the following monthly price data for five different biofuel raw materials k.

• Maize: US No. 2 Yellow, FOB Gulf of Mexico ($/t)

• Wheat: No. 1 Hard Red Winter, ordinary protein, FOB Gulf of Mexico ($/t)

• Rapes oil: Crude, fob Rotterdam ($/t)

• Palm oil: Malaysia Palm Oil Futures (first contract forward) 4-5 % FFA ($/t)

• Wood: average price ($/m3) for softwood (average export price of Douglas Fir, U. S. Price) and hardwood (Dark Red Meranti, select and better quality, C&F UK port)

The data for crude oil prices were obtained as an average of Dated Brent, West Texas Intermediate and Dubai Fateh (Euro/barrel). Raw material prices were taken

All prices are average prices per year Ein barrel Rohol sind 159 L

Die Dichte von Rohol schwankt zwischen 0.8 bis 1 kg/l—beim Vergleich mit Rohol rechnet man

im Allgemeinen mit einer Dichte von 0.883 kg/l

Als mittlere Dichet von Holz wurde 600 kg/m3 angenommen from www. indexmundi. com. Table 1 shows average annual prices for the five biofuel feedstocks as well as for crude oil, based on monthly data from April 1982 to April 2010. The historical price overview shows significant differences in price developments for the different types of raw material. For example, the palm oil price has doubled between 2006 and 2010, while during the same time prices for wood remained almost stable.

Annual data on population, GDP, energy consumption, inflation and agri­cultural prices were taken from the ‘World Development’ and converted into monthly data through linear interpolation. We measured all prices, GDP and

energy consumption in US Dollars and converted them into Euros using monthly exchange rate averages.

An ARMAX (Harvey 1993) modelling approach with a one month autoregres­sive term of the structural model disturbance and additive annual effects was used (see the results in Table 2). It is obvious that the price of crude oil is significantly correlated to prices for biofuel feedstock. Crude oil has the weakest impact on prices for wheat and maize, while rapes oil and palm oil prices are heavily influ­enced. The influence on wood is in between these two groups. The results indi­cate that both rapes and palm oil have been used as energy inputs to a significant degree in the past and are therefore more closely related to oil price changes than wheat and maize. These are still predominantly used as input for food production.

Future prices for biofuel feedstock in 2015 and 2020 are based on the estimation results in Table 2. For the calculation, projected values for all independent vari­ables are necessary. In regard to prices for crude oil, we refer to oil price scenarios that have been published by IEA (2007) and the International Energy Outlook. We then investigate the effects of crude oil prices per barrel of Euro 50, Euro 100, Euro 150 and Euro 200 in 2020. For 2015, crude oil prices are calculated through linear interpolation of the 2011 value and the 2020 scenario. In regard to the other vari­ables, we assume a 1 % p. a. increase in world population, a 2.5 % p. a. increase in GDP per capita, a 1.25 % p. a. increase in energy consumption per capita, a 5 % p. a. increase in agricultural prices and a global inflation of 6 % p. a. These assump­tions are close to the average rate of change of each variable during 1982 and 2010. For simplicity reasons, business cycle effects are not taken into consideration.

Dependent on different crude oil price developments, biofuel raw material prices for 2015 and 2020 are determined. Table 3 reports projected prices for 2015 and 2020 as well as actual and predicted prices in 2010. Prices are expressed in Euros per tonne, as production cost scenarios use tonne units for all material inputs. For crude oil, we assume a mass density factor of 0.883 kg/L and for wood a mass density factor of 0.6 kg/dm3. Except for palm oil, predicted 2010 prices are higher than they really were. This indicates that the price level in 2010 was lower than one would have expected if prices had followed the typical development of the past three decades. A calming effect on commodity due to the economic crisis may be one of the main reasons for that. The 2010 price level for all raw materials, except palm oil, was below the peak of the pre-crisis level in 2006 and 2008, while crude oil prices in 2010 were close to the pre-crisis peak. As mentioned before, we refrain from considering any type of business cycle effects on prices but focus on longer term trends in raw material prices. For this reason, we do not adjust projected prices for 2015 and 2020 to the ‘prediction error’ in 2010 but consider the higher predicted prices for 2010 (and consequently for 2015 and 2020) as reflecting an upcoming upwards trend of commodity prices in case the world economy recovers.

Wheat, rapes oil and maize prices are expected to undergo the largest rise until 2020. In the Euro 50 scenario, prices for these three types of biomass will increase by 89, 85 and 66 %, respectively, compared with the actual prices in 2010, which were rather low. In the Euro 200 scenario, price advances will be significantly higher. Changes are all above 100 %. In regard to palm oil, prices are expected to

Table 2 Results of ARMAX model estimations

remain stable in the Euro 50 scenario but increase substantially in the Euro 200 scenario. This reflects the stronger link between crude oil and palm oil prices. As for wood, all scenarios except the Euro 200 scenario expect tumbling prices. The latter estimates constant prices for the time period between 2010 and 2020.

Waste material is another important group of raw material for biofuels. However, there are no world market prices available, due to waste rarely being traded internationally, because of high transport costs per unit and small unit val­ues. In our scenario analysis, we assume that the prices for waste lignocellulosic material are constantly 1/4 of the price of maize and the price for waste oil is 1/2 of the price of palm oil. At this point, we assume that producers are price takers and that production functions are linear homogenous.