Debate on the choice between maximum likelihood estimator and minimum chi-square estimator is still on. Rao’s school of thought advocates for the maximum likelihood estimator and Berkson’s school of thought for the minimum chi-square estimator. In this paper a numerical study has been made by the authors by computing the relative efficiency of these estimators together with that of minimum B estimator for the Exponential parameter. The study reveals that for few choices of ‘A’ and ‘B’ the constants in Statistic B introduced by Khan and Baily (1987), the minimum B estimator is more efficient than the other two. Also maximum likelihood estimator is more efficienct than the minimum chi-square estimator.