Footnote:
Nach Informationen von SSRN wurde die ursprüngliche Fassung des Dokuments November 1998 erstellt
Description:
This is an examination of VAR modeling used to generate expectations for forward-looking variables in the Federal Reserve's macro-economic model FRB/US. Aside from analyzing economic as well aseconometric properties of VAR models currently in use, this paper evaluates certain improvements to core VAR model. One important issue is the dynamics of such models: how do variables behave over time and how do they relate to each other, especially when disturbed by exogenous events? For this purpose, this paper uses three analytical tools, two econometric and one economic, but all resulting in easily interpreted graphics. The two econometric methods, basically "fact-matching exercises, are employed to compare the autocorrelation functions implied by each model with those estimated for the data, using both time- and frequency-domain techniques. The economic method employs impulse response functions already familiar in the VAR modeling literature.These techniques are used to examine the short and long-run dynamics of output, inflation, and the funds rate in three versions of the VAR models. In contrast to the methodology of Cogley and Nason (1996), this paper uses stylized facts not based on the output of a VAR model but on the sample estimates of autocorrelations and power spectra of the time series of output, inflation, and the federal funds rate, themselves. Further, because the most interesting implications of macro models are often not their univariate time series properties, but their relationships with each other over time, this paper examines dynamic cross-correlations in both time and frequency domains. One interesting topic is the "price puzzle," an awkward issue in monetary policy modeling because both data and VAR models exhibit an apparent positive initial correlation between the federal funds rate and the inflation rate. Another topic is the inflation/output tradeoff characteristic of each model. Both of these have obvious relevance for monetary policy