Long-term interest rates have been falling globally since the early 1980s and have reached historically low levels. Past forecasts largely missed this secular decline. This paper reviews methodologies for making long-term interest rate projections. We synthesize results from studies that use long historical series and cross-country data to estimate the trend and decompose it into components. We then construct a set of economic indicators that are potentially useful in interest rate forecasting. We add international, forward-looking economic indicators as explanatory variables in a standard macrofinance forecasting model. We find that the model with international variables can outperform the other models by better tracking the falling trajectory of United States interest rates in the post-2008 period, a trend missed by domestic variables. Further, we find that global economic indicators, especially the composite leading indicator for the European Union, are capable of accounting for a large portion of yield variance not only in the U.S. but in other advanced economies as well.
Abstract
Downloads
Key Findings
- Two methodologies are potentially useful for constructing long-range interest rate projections: semi-structural methods of interest rate trend decomposition and standard statistical forecasting models with an extended set of explanatory variables, including forward-looking economic indicators. These methodologies use different data and samples, and they provide complementary pieces of information.
- We perform a decomposition of the long-run nominal interest rate over the period 1981 to 2019 under the restriction of long-run inflation neutrality. Three variables, the earnings-price ratio of the stock market, the weighted average of past and forecasted consumption growth, and year-on-year productivity growth, explain 87% of variation in the 10-year real rate. The relative importance of the various macroeconomic determinants changes over time, with the earnings-price ratio mattering most in the 1981 to 1988 period and consumption growth most significant following recessions.
- We add international, forward-looking economic indicators as explanatory variables in a standard macrofinance forecasting model. We find that the model with international variables can outperform the other models by better tracking the falling trajectory of United States interest rates in the post-2008 period, a trend missed by domestic variables. Further, we find that global economic indicators, especially the composite leading indicator for the European Union, are capable of accounting for a large portion of yield variance not only in the U.S. but in other advanced economies as well.