One of the more significant developments in econometric modeling over the past decade has been the invention of the forecasting technique known as Bayesian vector autoregression (BVAR). This paper provides a detailed description of the process of specifying a BVAR model of quarterly time series on the U.S. macroeconomy. The postsample forecasting performance of the model is evaluated at an informal level by comparing the model's performance to certain naive forecasting methods, and is evaluated at a formal level by means of efficiency tests. Although the null hypothesis of efficiency is rejected for the model's forecasts, the accuracy of the model exceeds that of naive forecasting methods, and seems comparable to that of commercial forecasting firms for early quarter forecasts.
The new classical view that macroeconomic fluctuations can be modeled as an equilibrium system perturbed by transitory monetary disturbances has been challenged in recent years by another equilibrium view of fluctuations, the so-called real business cycle theory. In this latter framework, shocks to the production function induce both intertemporal substitution of labor supply and permanent shifts in the stochastic trend of output. Monetary shocks, on the other hand, play only a minor role in this view of the cycle. Much of the empirical support for the real business cycle view of fluctuations is based on a re-examination of traditional methods for detrending economic time series. The issues raised by the real business cycle theorists are not new; indeed, they go back at least to the NBER's first business cycle studies. However, the real business cycle theorists attach a radical economic interpretation to what, on the surface, appears to be a purely technical note on the proper method for detrending economic data. This paper reviews the debate over stochastic trends, discusses the economic implications of the real business cycle interpretation of stochastic trend models, and weighs the time series evidence for some of the stronger claims made by real business cycle theorists. We conclude that, while this literature raises real and useful questions about the interpretation of observed fluctuations, the new classical view of the cycle is not ruled out by the data.
This paper reports some empirical evidence on the relation between the expected real interest rate and monetary aggregates in postwar U.S. data. We find some evidence against the hypothesis, implied by the Real Business Cycle model of Litterman and Weiss (1985), that the expected real interest rate follows a univariate autoregressive process, not Granger-caused by monetary aggregates. Our findings, however, are consistent with a more general bivariate model--suggested by what Barro (1987, Chapter 5) refers to as "the basic market-clearing model"--in which the real rate depends on its own lagged values and on lagged output. Taking this bivariate model as our null hypothesis, we find no evidence that money-stock changes have a significant liquidity effect on the expected real interest rate.