The uncertain nature of the natural rate of interest

Over the past couple of months something strange swept across parts of Europe—yields on government-issued bonds turned negative as the European Central Bank prepares to buy bonds as part of its forthcoming quantitative easing program. This short-term phenomenon raises the question about the long-term trends in interest rates. Are the extremely low rates in not only the Eurozone but also the United States the result of a long-run trend known as secular stagnation? Or are these negative bond yields just the result of a particularly nasty business cycle?

A new paper argues that the concerns about permanently lower interest rates are overhyped. The authors, James D. Hamilton of the University of California at San Diego, Ethan S. Harris of Bank of America Merrill Lynch, Jan Hatzius of Goldman Sachs, and Kenneth D. West of the University of Wisconsin, take aim in particular at the idea that the main determinant of the natural rate of interest, or the rate when the economy is in equilibrium, is the growth rate of the economy. If this relationship doesn’t hold up, then slower economic growth in the future won’t necessarily result in lower interest rates.

Taking a look at the history of interest rates over the past 200 years, the authors argue that other factors are critical to determining the natural rate. They list quite a few factors, among them demographics, debt defaults by national governments, and the level of financial regulation. The last factor is particularly interesting. The four economists show that periods of tighter financial regulation coincide with lower interest rates. Given the deregulation of the finance sector during the 1970s and 1980s, they would expect the natural rate to be higher as other sources of financing are available in the economy. One case in point: Companies can access the bond market instead of getting a loan from a bank.

But the broader point the paper makes is that determining the natural rate of interest is quite difficult. According to the data, the rate has changed quite a bit over the years. After reading their paper, it doesn’t take much to think that accurately estimating the rate is a fool’s errand—mainly because it is so difficult to know the actual long-term natural rate of interest.

Matthew Klein at FT Alphaville takes this seeming inability to be further proof of nihilism when it comes to monetary policy. If such an important parameter can’t be estimated then he says monetary policy makers are “forever doomed to be fumbling about in the dark rather than smoothing out the vicissitudes of the cycle.” But it seems unwarranted to abandon efforts to smooth out the business cycle just because a long-run parameter is uncertain.

Indeed, Hamilton, Harris, Hatzius, and West incorporate this uncertainty into the Federal Reserve staff’s preferred macroeconomic model and find that it leads to the conclusion that monetary policy should have more “inertia.” In other words, uncertainty means that the Fed should wait a while before shifting course. Given the current situation, this would mean waiting a while to raise interest rates and then rapidly raising rates when the time comes.

Intriguingly, this recommendation also arises from models that assume we know the natural rate of interest. Greg Ip used a model created by Federal Reserve of San Francisco economists to see how interest rates should be increased if we believe that demand is the major problem at the moment and that the long-run rate has not changed at all. The policy recommendation: keeping interest rates low for a long time and then rapidly raising them.

Economic models are simplifications of reality and are often quite dependent upon assumptions, guesses, and estimations of important parameters in the far more complicated real economy. At times, these inputs can be so wildly off base as to make the entire endeavor worthless. As this new paper shows, it can be incredibly difficult to figure out even an adequate assumption. Yet economists and policymakers will continue to need models to interpret a complex world, so we’ll have to continue onwards comfortable with a certain amount of fog shrouding the path ahead.