Working papers // University of California, Department of Economics 05,35
Queuing theory may be useful for analyzing economic phenomena involving count and duration data. We develop maximum likelihood estimators for the time-varying parameters of a simple queuing system based on two kinds of data: complete interarrival and service times (IST), and number of units in service (NIS). The IST estimator dominates the NIS estimator, in terms of ease of implementation, bias, and variance. The model is useful for many empirical applications in economics.