Quantile regression in lower bound estimation

In this paper, I illustrate the additional information that can be prodivided in estimating the lower bound (Sutton 1991, 1998) by using quantile regression. Quantile regression allows us to invesigate the influence of outliers. Previous lower bound have been performed using the simplex method. In this paper, the lower bound estimates are obtained using both methods for sectors belonging to a 'control group' and sectors belonging to an 'experimental group' forItalian manufacturing sectors in 1995. The data employed are drawn from the ISTAT (National Institute of Statistics, Italy) dataset. The results suggest that Sutton's predictions are robust.