Productivity and Efficiency of Educational Expenditures: A Quantitative Analysis Using a Semi-parametric Stochastic Frontier Model.

Despite the recent vigorous debate concerning the necessity for further increases in educational expenditure, the basic questions "Is Japan's spending on education really low?" and "Does money matter?" have never been examined analytically in Japan. The main purpose of this paper is to carry out a quantitative analysis of the effect and efficiency of educational expenditure in Japan and other developed countries. The paper consists of two parts. Firstly, using OECD expenditure data, I examine the conventional indices concerning educational expenditure and compare Japan and other developed countries using an alternative index. Secondly, I evaluate the effect of expenditures on secondary education and the inefficiencies of OECD countries by employing a stochastic frontier model. The following are the main results of the analyses. As to the comparison, there are some problems with the conventional macro indices such as educational expenditure per student, the educational expenditure- GDP ratio and the educational public expenditure-general public expenditure ratio. I employ the alternative index of Ram (1995) using the residuals of simple linear regression and compare the level of educational expenditures in OECD countries. In terms of primary and secondary education, the level of educational expenditure in Japan is close to the average of developed countries, and is therefore not much inferior to that of other developed countries, contrary to the recent claims concerning public education. The productivity of educational expenditure is evaluated in the light of two points: mean level and equity. The output indices of mean level and equity are estimated by the fixed-coefficient regression of the PISA data. The main analyses are conducted using a Bayesian semi-parametric stochastic frontier model which assumes no particular functional form concerning the effect of expenditures and the particular distribution concerning the inefficiencies. In terms of the mean level, an increase in educational expenditures has some positive effect. However, the production function is concave, and therefore there is a tendency for the effect of additional units of expenditure to diminish. The effect of educational expenditure on equity is neither monotonous nor strong. The estimated inefficiencies of each country tell us that there is a considerable gap between best practices and educational practices in many countries, and that the Japanese educational system is relatively efficient in mean level output and inefficient in equity output. These results imply that increasing educational expenditure requires further careful examination and discussion. [ABSTRACT FROM AUTHOR]