Search Results
Working Paper
Doubts about the Model and Optimal Policy
This paper analyzes optimal policy in setups where both the leader and the follower have doubts about the probability model of uncertainty. I illustrate the methodology in two environments: a) an industry populated with a large firm and many small firms in a competitive fringe, where both types of firms doubt the probability model of demand shocks, and b) a general equilibrium economy, where a policymaker taxes linearly the labor income of a representative household in order to finance an exogenous stream of stochastic spending shocks. The policymaker can distrust the probability model of ...
Working Paper
Optimal Fiscal Policy with Recursive Preferences
I study the implications of recursive utility, a popular preference specification in macrofinance, for the design of optimal fiscal policy. Standard Ramsey tax-smoothing prescriptions are substantially altered. The planner overinsures by taxing less in bad times and more in good times, mitigating the effects of shocks. At the intertemporal margin, there is a novel incentive for introducing distortions that can lead to an ex-ante capital subsidy. Overall, optimal policy calls for a much stronger use of debt returns as a fiscal absorber, leading to the conclusion that actual fiscal policy is ...
Working Paper
A Note on the Finite Sample Bias in Time Series Cross-Validation
It is well known that model selection via cross validation can be biased for time series models. However, many researchers have argued that this bias does not apply when using cross-validation with vector autoregressions (VAR) or with time series models whose errors follow a martingale-like structure. I show that even under these circumstances, performing cross-validation on time series data will still generate bias in general.
Working Paper
A Note on the Finite Sample Bias in Time Series Cross-Validation
It is well known that model selection via cross validation can be biased for time series models. However, many researchers have argued that this bias does not apply when using cross-validation with vector autoregressions (VAR) or with time series models whose errors follow a martingale-like structure. I show that even under these circumstances, performing cross-validation on time series data will still generate bias in general.