Estimation of cost efficiency in Australian universities

Various studies have been carried out on estimation of cost functions for Australian universities, for example, Throsby (1986) and Heaton and Throsby (1997, 1998), and Lloyd et al (1993). While estimating a cost function with different purposes, all of the studies have estimated an aggregate cost function in the sense that the sampled institutions were assumed to have the same cost function. The present study differs from the previous studies in two aspects; first, it employs the stochastic frontier analysis for the specification of a cost function for Australian universities, which allows for the estimation of cost efficiency for each university. Secondly, a panel data set is utilised which enables not only comparisons of cost efficiency between universities, but also hypothesis-testing of assumptions about university cost functions. The data set for the present study are comprised of student, staff and expenditure statistics for 36 Australian universities over the period 1995-2002, obtained from the Department of Education, Science and Training (DEST). The student statistics enable a breakdown of student numbers by field of study and mode of attendance. Research expenditure has been excluded so the cost efficiency estimated in the study only reflects how efficiently resources were used for teaching. The findings from the study contribute to the formulation of an equity-efficiency based policy on university-financing. The main conclusion reached is that the hypothesis that universities are operating at minimum cost efficiency is rejected over the sample period 1995-2002. The period encompasses the peak of the debate on reform of higher education, including conflicting views on this dimension of university performance. The main direction of further research is to extend the model to include factors that explain the efficiency rankings and thereby provide scope for universities to strengthen their absolute and relative performance over time.