concentration in these two regional markets results in refinery profits. We did not find this
pattern in our panel data model. Similarly, gasoline imports have a significant negative effect on
the profit margin in both East Coast and Midwest regions, possibly because these two regions are
more heavily dependent on imported refinery products to meet their regional demand. Ethanol
production has a significant negative effect on the refiner’s profit margin in all five PADD
regions.
Conclusions
We employ pooled OLS regression, a fixed effect panel data model, and a panel FGLS
estimation method to quantify the possible impact of ethanol on regular gasoline in the U.S. as a
whole and in five regions of the U.S. The models control for gasoline imports, refinery capacity,
capacity utilization rate, hurricanes, market concentration in the refinery industry, stocks, and
seasonality.
Estimation results show that over the period 1995 to 2007, ethanol production had a significant
negative effect of $0.29 to $0.40 per gallon on retail gasoline prices. The results suggest that this
reduction in gasoline prices came at the expense of refiners’ profits. These results are statistically
significant across a range of model specifications and across all regions.
Results for individual U.S. regions indicate that the largest impact of ethanol on gasoline is
found in the Midwest region where gasoline prices were reduced by 39.50 per gallon. The Gulf
Coast region is found to have experienced a 24.60 reduction in the retail gasoline price, while for
the West Coast and East Coast, the average price drop is about 23.30. The smallest impact, a
17.10 reduction, is found in the Rocky Mountain region, mainly because of its comparatively
low gasoline consumption.
These reductions in retail gasoline prices are surprisingly large, especially when one considers
that they are calculated at their mean values over the sample period. The availability of ethanol
essentially increased the “capacity” of the U.S. refinery industry and in so doing prevented some
of the dramatic price increases often associated with an industry operating at close to capacity.
Because these results are based on capacity, it would be wrong to extrapolate the results to
13