"The usual Markov model of marriage permits informative experiments on the effect of alterations in the transition rules.  It can tell, for example, what difference it would make to the durability of marriage if there was no divorce and the other transitions were as observed. This is in addition to the capacity of the usual model to find the effect of small changes in the transition rates.  Canadian data for 1970-1982 permit comparisons over time, and show among other things not only that married men live longer than single, but that the difference is increasing; the increase in the 'marriage bonus' over time also appears for women." (SUMMARY IN FRE)