Julian Betts and Eric Johnson
Abstract
Research which has tested for a link between school resources and students' earnings based on
samples of workers educated in the first half of the century tends to find stronger elasticities than do studies which have examined workers educated in the last twenty to thirty years. The paper
investigates whether diminishing returns could account for this pattern. A simple model of how
school inputs affect achievement suggests that term length is likely to affect achievement in a
different way than inframarginal expenditures such as spending to reduce class size or to increase
teachers' salaries. The theoretical case for diminishing returns is easier to make for inframarginal types of spending than it is for extensions to the length of the school year. Based on 1980 Census data on wages and state-level information on school resources, we find evidence of strongly diminishing returns to the teacher-pupil ratio and moderately diminishing returns to teacher salary. In both cases, the wage elasticity with respect to these inputs is predicted to have fallen over time. In contrast¸ we find evidence for increasing returns to length of the school year. The presence of diminishing returns to expenditures on class size and teachers' salaries suggests that an equalization of school expenditures across states would moderately increase average earnings among adult workers.