Causal reasoning in econometric models

摘要

Propagation of change based on causal ordering is a central element of causal reasoning in economic models. While causal reasoning has most often been applied in qualitative models, we demonstrate a technique for causal reasoning that offers explanations of structure and behaviour in quantitative, econometric contexts. Given a matching of equations with endogenous variables, causal reasoning can be applied to both static and dynamic system models. By propagating the disturbance of one or more exogenous variables, impact or static multipliers of the model can be derived along with a causal explanation. Dynamic analysis is achieved by propagation of lagged endogenous variables carried from the previous time periods. Two versions of Keynesian macro-econometric models, Klein's Model I, and the Klein-Goldberger Model are used as examples.