Proper econometric specification of the gravity model
Proper econometric specification of the gravity model
复制标题
DOI：
10.1111/1467-9701.00074
复制
发表时间：
1997-05-01
期刊：
影响因子：
2.5
通讯作者：
Matyas, L
中科院分区：
文献类型：
--
作者：
Matyas, L
对外贸易流量的建模与预测一直是国际经济学的重要课题。将其形式化的最有成效的方法之一是使用重力型模型。这些模型也被用来显示贸易集团和经济区域的（统计）意义，如APEC、南方共同市场、北美自由贸易协定等。Polak（1996）在最近的一篇论文中批评了这种做法，他指出使用区域虚拟变量来检验贸易集团在解释出口量方面发挥重要作用的假设存在歧义，并提出了一种替代规范来挽救这一模型。我想对这个问题发表评论，并表明从计量经济学的角度来看，在这个领域使用的所有重力类型模型都是错误的。然后，我展示了这是如何导致对贸易集团虚拟变量的错误解释和不正确的经济推断的。
MODELLING and predicting foreign trade flows has long been an important task in international economics. One of the most fruitful ways to formalise this has been through the use of gravity type models. These models have also been used to show the (statistical) significance of trading blocs and economic areas like the APEC, MERCOSUR, NAFTA, etc. This practice was criticised in a recent paper by Polak (1996), who pointed out the ambiguity of using regional dummy variable (s) to test the hypothesis that trading blocs have a significant role in explaining the volume of exports and also, suggested an alternative specification to salvage this model. I would like to comment on this issue and show that all gravity type models used in this area are misspecified from an econometric point of view. Then, I demonstrate how this leads to the incorrect interpretation of the trading bloc dummy variable (s) and improper economic inference.