Heterogeneous technology and panel data: The case of the agricultural production function

详细信息	查看全文 | 推荐本文 |

• 作者：Yair Mundlak^a ; ^{mundlak@agri.huji.ac.il} ; Rita Butzer^b ; ^{ritabutzer@gmail.com} ; Donald F. Larson^c ; ^{dlarson@worldbank.org}
• 关键词：C23 ; D24 ; Q10 ; O13
• 刊名：Journal of Development Economics
• 出版年：2012
• 期刊代码：62_03043878
• 类别：soc
• 出版时间：September, 2012
• 卷：99
• 期：1
• 页码：139-149
• 文件大小：281 K

摘要

Economic growth involves reallocating resources from traditional to new techniques of production, creating new relationships between particular resources and productivity. The paper analyzes the implications of this process on the estimation of agricultural production functions using a panel of countries. The data includes a measure of capital in agriculture absent from most studies. We employ a heterogeneous technology framework where implemented technology is chosen jointly with inputs to interpret information obtained in the empirical analysis of panel data. In this framework, estimates depend upon the economic environment, which is represented by state variables. It turns out that the old problem of identifying the production function cannot be resolved through the use of instrumental variables, but can be resolved using the allocation error. The paper discusses the scope for replacing country and time effects by observed state variables. The empirical results differ from those reported in the literature for cross-country studies, largely in augmenting the elasticities of capital and land and reducing those of fertilizer and labor. The evaluation of the marginal value productivity accounts for the flow of capital and fertilizer to agriculture and the flow of labor to other sectors, thereby contributing to overall economic growth.