Economics Update (April-June 1998)

Atlanta Fed Economist Develops
Dynamic Forecasting Model

T he Federal Open Market Committee employs the tools of monetary policy to effectively promote the goals of maximum employment, stable prices and long-term interest rates.

Because there is a considerable lag, however, between when a policy action is taken and when it becomes measurable in the economy, it is crucial that policymakers accurately anticipate the future economy they are acting on. Unfortunately, obtaining an accurate estimation of such projections is a daunting task.

One forecasting approach, economic rules of thumb, while sometimes theoretically grounded, has not proved to provide reliable short-term predictors. Another approach, large-scale econometric models, while ambitious in scope, is often hard to evaluate because of a preponderance of data, variables, implicit assumptions and judgmental adjustments.

But a new model developed by Atlanta Fed economist Tao Zha promises to be more reliable and easier to use than these more traditional approaches. In the Atlanta Fed's Economic Review (First Quarter 1998), Zha presents his model, which represents a new class of forecasting tools called dynamic multivariate models. Because models of this type incorporate and predict only a few economic essentials, they can be independently evaluated and improved. In addition, the empirical uncertainty of forecasts can be assessed through error bands.

Zha's model employs six key macroeconomic variables, but his article focuses on the predictability of the inflation rate. The model's forecasts have proved to be reliable in periods when inflation is difficult to forecast.

Just as important, however, is how the model reacts when its forecasts are less than satisfactory. To address this concern, the model self-assesses the probability that its forecast is in error. Additionally, it quickly incorporates new information into subsequent forecasts and updates its earlier prediction. This dynamic quality makes Zha's model a promising forecasting tool for monetary policymakers.

Return to Index