This paper develops a methodology to quantify the demand for taxicab services and to study the factors that are likely to influence such demand. Since owners and non-owners of private vehicles have different travel pattern characteristics, separate demand models are developed and estimated in this paper. Two modelling techniques are used in an integrated manner to quantify the demand for taxicab services. The first set of models relates to passengers' propensity to change modes. It is postulated in this paper that passengers with a high propensity to change mode are likely to be frequent users of taxicab services. Ordered probability models are specified and estimated to determine passengers' propensity to change modes. The second set of models relates to the frequency (demand) with which passengers use the taxicab. Departing from classical Poisson regression model structure, negative binomial models are specified and estimated. The dispersion parameter turned out to be significantly different from zero in both models, confirming the appropriateness of the specified negative binomial models. Empirical findings from this paper seem to indicate the need to develop separate taxicab demand models for each group of passengers. The demand for taxicab services appears to be influenced differently by the same set of external variables. The study illustrates the need to integrate passengers' propensity to change mode into the taxicab demand models. The developed integrated models seem to provide a more comprehensive explanatory framework. (a).
Samenvatting