Modulated renewal processes, suggested by D. R. Cox, give a flexible w
ay to introduce dependencies into point processes. We discuss the asym
ptotics of partial likelihood inference for modulated renewal processe
s when the random covariate for the process involves its history. We s
how, in some generality, that the estimators of the regression paramet
er and the cumulative hazard have the same asymptotic distributions th
at they would have under the usual proportional hazards model, even th
ough the martingale justification for partial likelihood no longer app
lies because of a reordering of the time-scale. An example is given to
illustrate the ideas. A simulation study is presented to confirm the
theoretical results.