Applications of econometric analysis to forecasting in international relations

Abstract

The apparent neglect of quantitative methodology in political analysis can be explained partly by the absence of a common paradigm or frame of reference for political inquiry and partly by the lack of experience with experimental analysis of empirical data. The absence of general theory poses considerable difficulties for analysis and for specifying the nature of expected relationships or outcomes. For example, without a good theory of war, it is difficult to explain, account for, and predict wars among nations as well as to forecast the probable range of casualties, the extent or duration of violence, geographical scope, and so forth. And the absence of sufficient experience with quantitative analysis poses equally numerous difficulties bearing upon our ability to go beyond purely descriptive modes of inquiry. For example, without sound analytical and computational tools it is difficult to develop empirical models, or simulations, or forecasts of such dynamics. This paper examines some key issues and difficulties encountered in the course of applying econometric analysis to forecasting in international relations. We will note the problems involved and the solutions adopted, and indicate the consequences of faulty analysis, analytical bias, or measurement error.