Forecasting the demand for tertiary education using econometric and Markovian models

In this paper a model for the demand for tertiary education in the state of Victoria by qualified students leaving secondary school is developed and estimated. The objective of obtaining short-term forecasts is hampered by projection difficulties associated with one of the main independent variables, the eligibles (i.e., the pool of potential tertiary students). In order t o overcome these difficulties an absorbing Markov chain incorporating changing transition proportions is derived and applied. This Markov model, coupled with more conventionally obtained projections of the remaining independent variables, is then used by the demand model to obtain the required forecasts.