Returns to scale in the production of taxi services: An empirical analysis

This paper applies a log-linear simultaneous equation model to estimate the returns to scale in a Cobb-Douglas-type bilateral searching and meeting function which characterises the search frictions between taxis and customers. Number of taxis, taxi fare, population, average occupied taxi journey time and disposable income are used as exogenous variables; customer demand, average customer waiting time and average taxi waiting and cruising time in search of each customer are used as endogenous variables. These variables are coupled together through the simultaneous equation model whose parameters are estimated from the Hong Kong Annual Taxi Survey data as well as data from Traffic and Transport Digest and Annual Digest of Statistics. Our calibration results show that urban taxi services in Hong Kong exhibit an increasing return to scale. The established simultaneous equation model can be used to obtain regulatory information useful for decision-making, such as the impact of the restriction over the number of taxi licenses and the fixing of taxi fare structure.