This paper analyses the technical efficiency of European railways in the period 1984–2005. Our results reveal first that the use of bootstrapping indicates a certain degree of sensitivity of the efficiency measurements to sampling variation. The efficiency decreases slightly and the DEA efficiency score overestimates the operational efficiency. Second, the corrected efficiency decreases from 1984 to 1989 and then increases until 2005 with some ups and downs. Third, the analysis of convergence indicates a worsening of efficiency, but in the sub-period 1995 to 2005 the efficiency improves. Fourth, controlling for environmental factors, the reforms generate increased railway efficiency.