Agriculture and Economic Interdependence: Discussion
讨论了两篇论文，它们共同关注如何更好地理解农村发展、农业作用以及促进经济增长的政策设计，强调模型选择应基于问题性质，并指出在州或区域层面精确定义农业部门的重要性。
The two papers in this session are more similar in theme, focus, and conclusion than it would at first appear. The theme in both is to provide an improved conceptual and empirical basis for understanding rural development, the role of the agricultural and the design of policies that can foster economic growth and adjustment. Leones, Schlutter, and Goldman provide insights on the importance of appropriately defining the sector, given the intended analysis. Kilkenny and Otto suggest ways to make the analysis of rural or regional economies richer by emphasizing the possibilities for more fully introducing the spatial features of economic activity. The fact that Leones, Schlutter, and Goldman utilize an input-output framework for a context in their paper and Kilkenny and Otto utilize computable general equilibrium (CGE) as a framework for their paper should not be taken to suggest that we have a war of models. It is clear from both papers that the choice of model is conditioned by the nature of the problem to be addressed. The importance of more fully understanding the role of agriculture in both developed and developing economies is highlighted by the ongoing changes in agricultural policy. Systemic changes in domestic agricultural policy and in trade policy are underway. Developed nations are reducing subsidies to agriculture and decoupling subsidies from production decisions. Developing nations, as a part of financing discipline and restructuring, are improving the terms of trade for agriculture. And, in the formerly centrally planned economies, policies are being contemplated and introduced that imply large reductions in labor in agriculture. Consequences of these changes for the rural economy and national economic growth, and the choice of adjustment and compensation schemes, are far from settled. Leones, Schlutter, and Goldman show that the characterization of the agricultural sector can produce quite different estimates of its impact on the economy, and that care in defining and communicating this characterization becomes more critical at state and regional levels, largely because the states or regions are not miniature national economies. Implications are for the use of local data sources and the design of input-output and other models more specialized to the policy problems of focus. The era of cranking factored-down national models for analysis of state and regional economies is at an end. As Leones, Schlutter, and Goldman demonstrate, we can do better. And simply qualifying the results is not enough. Policy makers infrequently read or understand the qualifications. Even those who offer the qualifications often do not know exactly what they mean in the specific analytical context. The answer is simply to make the models fit the problems.