Bootstrap Confidence Intervals for Ratios of Expectations

this paper is to examine how (ii) can be addressed via the bootstrap approach [Efron 1979; L'eger et al. 1992; Hall 1992; Efron and Tibshirani 1993]. The basic idea is to use the observations (X i ; Y i ) to estimate F , the joint distribution of X and Y , in order to simulate observations from that estimated distribution to estimate the distribution of T and therefore its quantiles which are used in constructing a confidence interval. The usual estimator of F i