Diminishing marginal returns (DMR) to school inputs could explain a wide variety of findings in the research literature. One important example is the influential finding by Heyneman and Loxley that school inputs are the 'predominant influence' on achievement in developing nations, where input levels are low, even though the same school inputs have relatively little influence in developed nations, where input levels are higher. However, few studies of education production, including those related to the Heyneman-Loxley hypothesis, use functional forms that allow for DMR, and common tests for DMR appear to be invalid. Various tests are implemented using data from 32 countries. As is commonly found in the literature, the marginal effects of school inputs are frequently negative, precluding DMR. In those cases with positive marginal effects, there is more evidence for DMR than for increasing returns, but constant returns rarely can be rejected. DMR therefore does not appear to explain the differences in results between developing and developed nations. Variable definitions are appended. (Contains 6 tables, 2 figures, 7 formulas, and 17 notes.)