Recently Huque and Katti (1976) have claimed that the maximum conditional likelihood estimator for the shape parameter of a gamma distribution while consistent is asymptotically inefficient. It is shown that their argument is in error and that in fact asymptotic efficiency is obtained. Furthermore it is argued that the conditional likelihood in this situation contains all available information for the shape parameter.
|Number of pages||3|
|Journal||Sankhyā: The Indian Journal of Statistics, Series B (1960-2002)|
|State||Published - 1983|