Niels Bohr, a Nobel Prize winner in Physics once said, “It is exceedingly difficult to make predictions, particularly about the future”. Undoubtedly, forecasting the economy is difficult at the best of times and during times of greater uncertainty this exercise can be even more challenging. For instance, we face such uncertainty when wondering where the economy is heading, who will win elections, and so on. Such unknowns about the future affect today’s decisions to consume, to invest, to hire or fire workers.
Why interest rates?
Interest rates matter a great deal to savers, to debtors and to the economy as a whole. More particularly, interest rates play a key role in the transmission of monetary policy decisions on the economy. In the past decade, nominal and real interest rates have been remarkably low. Major central banks decreased their policy rates to unprecedented low levels in response to the 2008 financial crisis and then kept them low for a long time. As economies are recovering, markets are speculating on the timing and size of future interest rate increases. For instance, in the United States, speculation about the normalisation of interest rates intensified as of late-2014, increasing interest rate uncertainty well
before the Federal Reserve (Fed) actually raised its policy rate for the first time in December 2015. Since then, the Fed has carried out three further hikes, the latest being in June 2017. In the euro area, markets anticipate that policy rates will remain at their current level for at least several quarters, while in the United Kingdom, speculation about the timing of the first policy rate hike is on the rise.
How does interest rate uncertainty behave?
Istrefi and Mouabbi (forthcoming) construct a new measure that summarises uncertainty about future interest rates of professional forecasters in public and private economic institutions in several countries. The uncertainty relates to the 3-month interest rate, forecasted 3 months ahead by professional forecasters in Consensus Economics’ surveys. With these interest rate forecasts and their respective realised outcomes, we construct the interest-rate uncertainty measure as the sum of two components: (i) disagreement between forecasters and (ii) the time-varying variability of the average forecast errors. Professional forecasters have different views on the level of future economic variables, including interest rates. For example, even if their assessment of future interest rates is based on the same publicly available data, forecasters can interpret the data in different ways, leading to different predictions, and thus disagreement, which is captured in the first component. The second component captures how difficult it is, for all forecasters, to predict future interest rates.
Interest rate uncertainty fluctuates substantially over time (Figures 1 and 2). The United States displays particularly high levels of interest rate uncertainty around the post dot-com bubble and after the September 11 attacks, in the early 2000s. Spikes in uncertainty are also observed for all countries during the Great Recession of 2008. Interest rate uncertainty is at its lowest from end-2008 for the United States, and from mid-2013 for the euro area. This muted uncertainty reflects low levels of disagreement between forecasters about the future path of short-term interest rates as well as a low variability of forecast errors. Indeed, during these periods, nominal policy interest rates reached levels close to zero (and even modestly below in some cases), and central banks communicated forward guidance that policy rates would remain low for a long time.