Table A7 (Appendix 5) shows the results of our OECD estimation. As the correlation matrix in Table A4 (Appendix 5) indicates, this sample does not suffer from the same collinearity problems as the full model. Therefore GDP is used in all estimations. Similar to the full model ,Land proves to be statistically significant in all columns. Land is slightly more statistically significant in columns 2 and 3 of the OECD model ,than columns 1 and 2 of the full model (e.g.,column2inthefullmodelhasa t-statistic of 3.06 ,while the same OECD regression in column 3 holds a t-statistic of3.24).Once again ,we observe that in OECD countries, more arable land leads towards biofuel policy ,which agrees with our intuition. The elasticities are also larger than