Abstract:
Benhabib, Schmitt-Grohe, and Uribe (2003) argue that if you relied solely on local analysis you would be led to believe that aggressive, backward-looking interest rate rules are sufficient for determinacy. But from the perspective of global analysis, backward-looking rules do not guarantee uniqueness of equilibrium and indeed may lead to cyclic and even chaotic equilibria. This comment argues that this result is premature. We utilize a discrete time model and make two observations. First, compared to their continuous time model, the cyclic equilibria under a backward-looking rule are much less likely to arise in a discrete time model. Second, pure backward-looking rules are less likely to suffer from these global indeterminacy problems than rules that also include current or future inflation.