Inference in Structural Vector Autoregressions when the Identifying Assumptions are not Fully Believed: Re-evaluating the Role of Monetary Policy in Economic Fluctuations
Journal of Monetary Economics,
Point estimates and error bands for SVARs that are set identified are only justified if the researcher is persuaded that some parameter values are a priori more plausible than others. When such prior information exists, traditional approaches can be generalized to allow for doubts about the identifying assumptions. We use information about both structural coefficients and impacts of shocks and propose a new asymmetric t-distribution for incorporating information about signs in a nondogmatic way. We apply these methods to a three-variable macroeconomic model and conclude that monetary policy shocks are not the major driver of output, inflation, or interest rates.
Does Machine Learning Help us Predict Banking Crises? ...
Impulse Response Analysis in a Misspecified DSGE Model: A Comparison of Full and Limited Information Techniques
Applied Economics Letters,
In this article, we examine the effect of estimation biases – introduced by model misspecification – on the impulse responses analysis for dynamic stochastic general equilibrium (DSGE) models. Thereby, we use full and limited information estimators to estimate a misspecified DSGE model and calculate impulse response functions (IRFs) based on the estimated structural parameters. It turns out that IRFs based on full information techniques can be unreliable under misspecification.
Much Ado About Nothing: Sovereign Ratings and Government Bond Yields in the OECD
IWH Discussion Papers,
In this paper, we propose a new method to assess the impact of sovereign ratings on sovereign bond yields. We estimate the impulse response of the interest rate, following a change in the rating. Since ratings are ordinal and moreover extremely persistent, it proves difficult to estimate those impulse response functions using a VAR modeling ratings, yields and other macroeconomic indicators. However, given the highly stochastic nature of the precise timing of ratings, we can treat most rating adjustments as shocks. We thus no longer rely on a VAR for shock identification, making the estimation of the corresponding IRFs well suited for so called local projections – that is estimating impulse response functions through a series of separate direct forecasts over different horizons. Yet, the rare occurrence of ratings makes impulse response functions estimated through that procedure highly sensitive to individual observations, resulting in implausibly volatile impulse responses. We propose an augmentation to restrict jointly estimated local projections in a way that produces economically plausible impulse response functions.
Sign Restrictions, Structural Vector Autoregressions, and Useful Prior Information
This paper makes the following original contributions to the literature. (i) We develop a simpler analytical characterization and numerical algorithm for Bayesian inference in structural vector autoregressions (VARs) that can be used for models that are overidentified, just‐identified, or underidentified. (ii) We analyze the asymptotic properties of Bayesian inference and show that in the underidentified case, the asymptotic posterior distribution of contemporaneous coefficients in an n‐variable VAR is confined to the set of values that orthogonalize the population variance–covariance matrix of ordinary least squares residuals, with the height of the posterior proportional to the height of the prior at any point within that set. For example, in a bivariate VAR for supply and demand identified solely by sign restrictions, if the population correlation between the VAR residuals is positive, then even if one has available an infinite sample of data, any inference about the demand elasticity is coming exclusively from the prior distribution. (iii) We provide analytical characterizations of the informative prior distributions for impulse‐response functions that are implicit in the traditional sign‐restriction approach to VARs, and we note, as a special case of result (ii), that the influence of these priors does not vanish asymptotically. (iv) We illustrate how Bayesian inference with informative priors can be both a strict generalization and an unambiguous improvement over frequentist inference in just‐identified models. (v) We propose that researchers need to explicitly acknowledge and defend the role of prior beliefs in influencing structural conclusions and we illustrate how this could be done using a simple model of the U.S. labor market.