Quantitative Economics 14(4), 2023, 1199-1220
Abstract: It is well known that Local Projections (LP) residuals are autocorrelated. Conventional wisdom says that LP have to be estimated by OLS and that GLS is not possible because the autocorrelation process is unknown and/or because the GLS estimator would be inconsistent. I show that the autocorrelation process of LP can be written as a Vector Moving Average (VMA) process of the Wold errors and impulse responses and that autocorrelation can be corrected for using a consistent GLS estimator. Monte Carlo simulations show that estimating LP with GLS can lead to more efficient estimates that generally have better coverage properties than estimation by OLS.
Published Version SSRN Working Paper Code and Replication Files
Abstract: Model selection criteria are one of the most important tools in statistics. Proofs showing a model selection criterion is asymptotically optimal are tailored to the type of model (linear regression, quantile regression, penalized regression, etc.), the estimation method (linear smoothers, maximum likelihood, generalized method of moments, etc.), the type of data (i.i.d., dependent, high dimensional, etc.), and the type of model selection criterion. Moreover, assumptions are often restrictive and unrealistic making it a slow and winding process for researchers to determine if a model selection criterion is selecting an optimal model. This paper provides general proofs showing asymptotic optimality for a wide range of model selection criteria under general conditions. This paper not only asymptotically justifies model selection criteria for most situations, but it also unifies and extends a range of previously disparate results.
Abstract: This paper investigates the utility of daily data in measuring high-frequency monetary policy surprises, comparing various announcement-day asset price changes with their intradaily (30-minute) counterparts. We find that both frequencies are similarly distributed and often highly correlated, particularly for longer-horizon measures. Testing daily surprises for systematic contamination from non-monetary policy news, we find no evidence to suggest that contemporaneous news releases bias their measurement. Empirical applications, including high-frequency passthrough to Treasury yields and proxy SVAR models, suggest that daily surprises produce results comparable to those obtained with intradaily data. Our findings suggest that while intradaily data remains invaluable for certain applications, daily data offers a practical and robust alternative for assessing monetary policy surprises, particularly when the event, or the reaction to it, extends beyond a narrow window, or when intradaily data is unavailable.
Abstract: Food inflation has been excluded from core measures of inflation under the reasoning that it is a phenomenon of the supply side of the economy, driven by stochastic supply shocks to agricultural production that can affect the availability of farm products and increase food price volatility. However, the share of food costs related to agricultural production has fallen over the years as food value chains have become more complex and food prices tied more closely to value added downstream in the supply chain. We calculate the magnitude and extension of agricultural price passthroughs to food prices in the United States after 2008. We leverage the results of simple models of food pricing under imperfect competition along the supply chain to identify possible sources of bias in the passthrough calculations. We argue that we can identify U.S. agricultural price passthrough to U.S. food prices in a structural vector autoregressive setting using weather instruments that shift supply of farm production but are excluded from demand. Our results suggest that understanding food inflation can benefit from focusing on factors affecting downstream segments of the supply chain.