Models, Measurement, and “Universal Patterns”
回顾1950-60年代发展经济学先驱Tinbergen、Chenery和Adelman各自应对数据匮乏的建模策略，揭示其共同归纳方法论“无理论的测量”，对理解早期发展经济学实证研究有参考价值。
Development economics in the 1950s and 1960s, as Jan Tinbergen and Irma Adelman saw it, was a “groping in the dark.” Besides the limited knowledge of dynamic mechanisms, the lack of data was a severe problem for any attempt at modeling, whether macroeconomic, input-output, or in the tradition of national income accounting. As concerns the three main figures of this article—the empirical researchers Tinbergen, Hollis Chenery, and Adelman—each developed his or her own approach to modeling: modeling in stages, modeling to capture universal patterns, and factor analysis, respectively. Tinbergen, Chenery, and Adelman shared a common inductive methodology, which might rightly be called “measurement without theory,” in the sense that there was no economic theory that could help them in organizing the available messy and often unreliable data. Chenery saw a role for abstract mathematical models as helping inquirers to rise above such impediments, even if only temporarily, or to gain focus, as was the case with his making technical progress an exogenous variable. Tinbergen’s models of the “first stage” were also intended to make planning for development tractable, but without mixing planning with such focusing devices. Adelman, for her part, preferred to search for and identify factors of development.