A Statistical Analysis on Assortative Matching

We consider the problem of statistical analysis on assortative matching. The random vector of interests consists of the age and the history of the first passage time of each of the newly-formed pairs. Complete data are not often available, however. Observations in each variate of the random vector maybe uncensored, right-censored, left-censored, interval-censored or missing. We propose a model that allows such kind of data. We estimate the joint distribution function of the random vector by the generalized MLE (GMLE). We use a self-consistent algorithm to obtain the GMLE and establish its asymptotic properties. We apply the estimation method to study age at first marriage using a panel data