5. A Regime-Based View of the Economy

"The St. Louis Fed has a long tradition of challenging the status quo. In 2016, Jim Bullard and the St. Louis Fed’s Research division pivoted to a new approach for evaluating the U.S. macroeconomy, which also had implications for how it views optimal monetary policy. Rather than assuming the economy will converge to one long-run outcome—the conventional approach—the St. Louis Fed now assumes the economy can switch between different states, or regimes, and the regime will influence the outlook for the macroeconomy and monetary policy."

—Christopher Waller, Executive Vice President and Director of Research

The New Narrative

Forecasting the macroeconomy is important for setting monetary policy, but when the forecasts are often wrong, something should change. In mid-2016, the St. Louis Fed changed to a regime-based way of thinking about how the economy works.

James Bullard shared some reflections on his first 10 years as Bank president during recent conversations with staff. The following are excerpts from those discussions.

Coming out of a recession, a typical forecast would suggest that the economy will grow faster for a while than it otherwise would, that job growth will be higher than normal for a while and that inflation might start to pick up and possibly go above the Fed's 2 percent target. Then, these variables would settle back to their steady-state rates of growth. That is, they would return to their average historical values. This approach to forecasting assumes that the economy will ultimately converge to a single, long-run outcome. It was the common approach used by many FOMC participants, including me.

Given that viewpoint, and since the Fed’s dual mandate (of stable prices and maximum sustainable employment) was close to being achieved in 2014, I had been an early proponent of moving forward with the normalization process, which included the policy rate's returning to its steady-state value.

But, by mid-2016, the Research team here at the Bank and I had become increasingly frustrated because our forecasts of the macroeconomy under this approach turned out to be wrong for four or five years in a row. Similarly, the "dot plots" in the FOMC's quarterly Summary of Economic Projections (SEP)—including the St. Louis Fed's dots, or projections for the policy rate—repeatedly projected many more increases in the policy rate over the forecast horizon than actually occurred. The conventional approach wasn’t useful.

Therefore, at the St. Louis Fed, we changed our approach to near-term forecasts of the macroeconomy and monetary policy in June 2016. The new approach required us to think differently about the possible long-run outcomes of the macroeconomy. Instead of having only one such long-run outcome, as was the thinking behind our previous narrative, the macroeconomy could switch between regimes (or steady states) and, therefore, could have a set of possible long-run outcomes.

The basic idea behind the new narrative was that there are three fundamental factors that can determine the nature of the regimes: productivity growth (which could be high or low), the real interest rate on short-term government debt (which could be high or low) and the state of the business cycle (expansion or recession).

The current regime appears to be characterized by low growth, low interest rates and also low inflation, which could be a relatively long-term outcome for the U.S. economy. The regime idea suggests that a situation like this could persist for many years and that we should not expect the same patterns from the previous decades to return, at least not in the near term.

This idea is particularly apt for the current environment. Safe, short-term real interest rates in the U.S. are extremely low and have been trending downward overall since the 1980s. Furthermore, the low safe real interest rates are a global phenomenon. For more recent trends in real-interest-rate regimes, see the presentation I delivered in Washington, D.C., on this topic.¹

As of February 2018, the U.S. appeared to be in a regime of low productivity growth and a high desire for safe assets. The latter is indicated by the relatively large negative value for ξ.
SOURCES: Kahn, James A.; and Rich, Robert W. Tracking Productivity in Real Time. Federal Reserve Bank of New York Current Issues in Economics and Finance, November 2006, Vol. 12, No. 8; Federal Reserve Bank of New York; Bureau of Labor Statistics; Board of Governors of the Federal Reserve System; Federal Reserve Bank of Dallas; and Bullard's calculations.

For purposes of monetary policy, which is regime-dependent, the planning horizon is two to three years.² Given that long-run trends affecting the economy are unlikely to turn around in two to three years, we assume in our new narrative that the current regime will continue over that horizon.

The St. Louis Fed's projections for monetary policy are, therefore, calibrated for the low regime. Hence, our projected policy rate path is relatively flat over the forecast horizon, which stands in contrast with the FOMC's median path. If a regime switch were to occur, our forecasts would then be calibrated for that new regime. Upside risks to our forecasts (e.g., higher inflation, an increase in the real rate or higher productivity growth) would lead us to steepen our path for the policy rate.

Endnotes

1. See my presentation from Feb. 26, 2018, R-Star Wars: The Phantom Menace. [ back to text ]
2. Under this regime-based approach, the St. Louis Fed stopped providing long-run projections in the FOMC’s SEP because there is not a single, long-run steady state for the economy. [ back to text ]

Additional Resources