Marginal costs of instruction at two-year higher education institutions.

Abstract

The purpose of this study was to determine the extent to which marginal costs of instruction vary by curricular area at two-year higher education institutions in the U.S. A subordinate objective was to determine the extent to which these costs vary for similar curricular areas among (1) "public" institutions versus "private, not-for-profit" versus "private, for-profit" institutions, and (2) "junior colleges" versus "community colleges" versus "technical colleges." The curricular areas targeted for marginal cost generation in the study were (1) arts and sciences, (2) business and data processing, (3) health, (4) service, (5) trades and (6) technical. Marginal costs were estimated in the study by means of regression analysis. Three different classes of cost function and quasi-cost function regression models were utilized: (1) enrollment-based models, (2) degree-based models, and (3) award-based models. These differed as to the types of measures utilized to account for "output" at institutions. In all models, some version of "total expenditure on instruction" served as the dependent variable. Variables pertaining to costs of input factors and to technologies of production served, along with "output" data, as independent variables in the study. Virtually all data used in the study were obtained from 1989-90 Integrated Postsecondary Educational Data System (IPEDS) surveys and were cross-sectional in nature. Four different regression model specifications were attempted in the study: (1) linear, (2) double-log, (3) linear with second order and interaction terms ("linear translog-like"), and (4) logarithmic with second order and interaction terms ("translog-like"). The effectiveness of each of these functional forms varied by institutional sector. Statistically significant marginal cost estimates were obtained in the study for all six curricular areas in most institutional sectors, but conflicting sets of estimates often were observed. As a result, the investigation was only partially successful in determining whether marginal costs of instruction for similar curricular areas varied systematically by institutional sector. In addition, a lack of data regarding quality of output rendered cross-sectoral comparisons tenuous. It was concluded that disaggregation of institutions into more homogeneous units of analysis in future studies held potential for the generation of more definitive marginal cost results.