Working Paper: NBER ID: w18870
Authors: Ulrich Mueller; Mark W. Watson
Abstract: Long-run forecasts of economic variables play an important role in policy, planning, and portfolio decisions. We consider long-horizon forecasts of average growth of a scalar variable, assuming that first differences are second-order stationary. The main contribution is the construction of predictive sets with asymptotic coverage over a wide range of data generating processes, allowing for stochastically trending mean growth, slow mean reversion and other types of long-run dependencies. We illustrate the method by computing predictive sets for 10 to 75 year average growth rates of U.S. real per-capita GDP, consumption, productivity, price level, stock prices and population.
Keywords: uncertainty; long-run forecasts; economic variables; predictive sets
JEL Codes: C22; C53; E17
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Uncertainty about the long-run average growth of economic variables (D89) | Predictive sets that maintain a prespecified probability of containing the true average growth rate (C51) |
| Predictive sets constructed using Bayesian and frequentist methods (C11) | Effectively capture the uncertainty surrounding long-run predictions (D84) |
| Incorporating uncertainty about the parameters that characterize the stochastic process (C51) | Accurate long-run forecasting (C53) |
| Frequency characteristics of the underlying processes (C22) | Predictive uncertainty (D80) |
| Local-to-zero spectrum (C46) | Predictive uncertainty (D80) |
| Approach that accounts for parameter uncertainty (C51) | More reliable long-run predictions (C53) |
| Traditional methods that do not account for parameter uncertainty (C51) | Less reliable long-run predictions (C59) |