David E. Spencer

Developing a Bayesian vector autoregression forecasting model

Abstract In recent years, Bayesian vector autoregression (BVAR) forecasting models have demonstrated considerable success in forecasting macroeconomic and regional economic variables. In spite of this success, these promising forecasting models have yet to be widely used in business forecasting. This is due, in part, to the rather formidable practical problem of specifying an appropriate BVAR forecasting model. The purpose of this paper is to simplify the model selection process by offering a systematic BVAR forecasting model selection procedure that is readily implemented using a popular software package. A practical five-step procedure is presented and then illustrated using a business forecasting application.

Statistical inference in multiperiod event studies

We examine several event-study test statistics that can be used to detect abnormal performance during amultiperiod event window. We demonstrate that one of the most commonly used test statistics does not, under the assumptions made, have the distribution claimed (standard normal), and thus tests using it will be biased. The magnitude of that bias is shown to increase with the length of the event window and can generally be expected to lead to excessive rejection of the null hypothesis. We also compare the relative power of alternative test statistics that are normally distributed and are straightforward to apply.

Bootstrapping structural VARs: Avoiding a potential bias in confidence intervals for impulse response functions

Constructing bootstrap confidence intervals for impulse response functions (IRFs) from structural vector autoregression (SVAR) models has become standard practice in empirical macroeconomic research. The accuracy of such confidence intervals can deteriorate severely, however, if the bootstrap IRFs are biased. In this paper, we document an apparently common source of bias in the estimation of the VAR error covariance matrix. The bias is easily corrected with a straightforward scale adjustment. This bias is often unrecognized because it only affects the bootstrap estimates of the error variance, not the original OLS estimates. Nevertheless, as we illustrate here, analytically, with sampling experiments, and in an example from the literature, the bootstrap error variance bias can have significant distorting effects on bootstrap IRF confidence intervals even if the original IRF estimate relies on unbiased parameter estimates.

Real Wages and Monetary Policy: A DSGE Approach

Economists have long investigated the cyclical behavior of real wages in order to draw inferences regarding the relative stickiness of prices and wages. Recent studies have adopted techniques intended to identify monetary shocks and examined the response of real wages and output or employment to such shocks. A finding that real wages are procyclical in response to a positive monetary policy shock, for example, is taken as evidence that prices are stickier than wages. In this paper, we show that factors other than wage and price stickiness affect the response of real wages to a monetary policy shock. Accordingly, examining the response of real wages is not enough to sort out the relative stickiness of prices and wages. We use two prominent DSGE models to help us address this issue. These models incorporate both sticky wages and prices but in different ways. The first model (Huang, Liu, and Phaneuf, American Economic Review, 2004) is relatively simple and is not intended for policy analysis. Its relative simplicity allows us to approach the issues both analytically and through simulations. The second model (Smets and Wouters, American Economic Review, 2007) is a relatively complex model of the U.S. economy with many frictions and intended to be useful for policy analysis. Because of its complexity, we must rely principally on simulation exercises. Using these models we offer robust evidence that the real wage response to monetary policy is affected in important ways by properties of the economy other than stickiness of wages and prices, such as the importance of intermediate goods in the production process and the size of key elasticities. Consequently, we cannot appropriately infer the relative stickiness of wages and prices from examining only the response of real wages to a monetary policy shock.

State-Level Evidence on the Cyclicality of Real Wages

There is considerable evidence that real wages have become more procyclical over time in the U.S. A novel explanation for this phenomenon has been recently offered by Huang, Liu, and Phaneuf (2004), HLP. They develop a model to show that, as the input-output structure of an economy becomes more sophisticated, the response of real wages to an aggregate demand shock becomes more procyclical (less countercyclical) in an environment with both sticky wages and prices. We test the basic prediction of their model using state-level data in the U.S. We exploit the fact that the economies of individual states in the U.S. exhibit a range of input-output structures in the production process. This variation allows us to examine how the cyclical response of real wages to aggregate monetary policy varies with production structure in order to test the HLP hypothesis. We find strong support for the hypothesis. Our direct empirical test complements the evidence provided by HLP in their original paper.