Search Results
Report
Robust inference in models identified via heteroskedasticity
Identification via heteroskedasticity exploits differences in variances across regimes to identify parameters in simultaneous equations. I study weak identification in such models, which arises when variances change very little or the variances of multiple shocks change close to proportionally. I show that this causes standard inference to become unreliable, outline two tests to detect weak identification, and establish conditions for the validity of nonconservative methods for robust inference on an empirically relevant subset of the parameter vector. I apply these tools to monetary policy ...
Report
Identifying shocks via time-varying volatility
An n-variable structural vector auto-regression (SVAR) can be identified (up to shock order) from the evolution of the residual covariance across time if the structural shocks exhibit heteroskedasticity (Rigobon (2003), Sentana and Fiorentini (2001)). However, the path of residual covariances can only be recovered from the data under specific parametric assumptions on the variance process. I propose a new identification argument that identifies the SVAR up to shock orderings using the autocovariance structure of second moments of the residuals, implied by an arbitrary stochastic process for ...
Working Paper
Mean Group Distributed Lag Estimation of Impulse Response Functions in Large Panels
This paper develops Mean Group Distributed Lag (MGDL) estimation of impulse responses of common shocks in large panels with one or two cross-section dimensions. We derive sufficient conditions for asymptotic normality, and document satisfactory small sample performance using Monte Carlo experiments. Three empirical illustrations showcase the usefulness of MGDL estimators: crude oil price pass-through to U.S. city- and product-level retail prices; retail price effects of U.S. monetary policy shocks; and house price effects of U.S. monetary policy shocks.
Working Paper
Estimating Impulse Response Functions When the Shock Series Is Observed
We compare the finite sample performance of a variety of consistent approaches to estimating Impulse Response Functions (IRFs) in a linear setup when the shock of interest is observed. Although there is no uniformly superior approach, iterated approaches turn out to perform well in terms of root mean-squared error (RMSE) in diverse environments and sample sizes. For smaller sample sizes, parsimonious specifications are preferred over full specifications with all ?relevant? variables.
Working Paper
Mean Group Distributed Lag Estimation of Impulse Response Functions in Large Panels
This paper develops Mean Group Distributed Lag (MGDL) estimation of impulse responses in large panels with one or two cross-section dimensions. Sufficient conditions for asymptotic consistency and asymptotic normality are derived, and satisfactory small sample performance is documented using Monte Carlo experiments. MGDL estimators are used to estimate the effects of crude oil price increases on U.S. city- and product-level retail prices.
Working Paper
Estimation of Impulse Response Functions When Shocks are Observed at a Higher Frequency than Outcome Variables
This paper proposes mixed-frequency distributed-lag (MFDL) estimators of impulse response functions (IRFs) in a setup where (i) the shock of interest is observed, (ii) the impact variable of interest is observed at a lower frequency (as a temporally aggregated or sequentially sampled variable), (iii) the data-generating process (DGP) is given by a VAR model at the frequency of the shock, and (iv) the full set of relevant endogenous variables entering the DGP is unknown or unobserved. Consistency and asymptotic normality of the proposed MFDL estimators is established, and their small-sample ...