Re‐examination of production, cost, and restricted profit functions using quantile regression approach
用分位数回归估计了柯布-道格拉斯形式的生产、成本和受限利润函数，发现规模报酬和技术进步在不同分位数上存在差异，传统方法可能高估或低估。
Abstract This article estimates quantile regressions of production, cost, and restricted profit functions using a Cobb‐Douglas functional form with non‐Hicks neutral technology change. In contrast to previous studies, quantile regression estimates reveal the relationship between the independent (production, cost, and restricted profit) and dependent (input quantities and prices) variables at each quantile of the distribution. An empirical application using data from 48 states in United States from 1960 to 2004 indicates the returns to scale and aggregate technology not only differ across production, cost, and restricted profit functions but across states in different quantiles of the distribution. This suggests the traditional measures of returns to scale and aggregate technology are under‐ and overestimated in states at upper and lower quantiles, respectively.