This paper proposes some tests for choosing between two estimated models using some Pearson type statistics. We allow for arbitrary ?n-asymptotically normal estimators to be used in forming these statistics. This provides some flexibility in practice. Then Large Sample theory and bootstrap...

We conduct a performance evaluation in terms of technical efficiency of large law firms based on a unique sample of 163 large law firms in the USA in 1994. To estimate technical efficiency for individual firms, we use the stochastic frontier production function with composed errors where output...