A COMMENT ON SOME RECENT ESTIMATES OF ECONOMIES OF SCALE IN HIGHER-EDUCATION

A simulation of the response of teaching labour inputs to a variety of course characteristics results in scale elasticities which are greater than one in the ceteris paribus case, but which are sensitive to course characteristics. Support is given to Nelson and Hevert (1992) who argue that cost functions omitting class size are mis-specified, but the argument is extended to other characteristics. For both Nelson and Hevert, and Lloyd, Morgan and Williams (1993) it is suggested that estimated scale elasticities may be distorted by departmental or institutional decisions to 'spend' some of their cost savings from increased scale.