The Real Interest Rate Decline in Long Historical Perspective
The real interest rate has dropped sharply in the twenty-first century. To what extent is this likely to be temporary, rather than persistent? In Long-Run Trends in Long-Maturity Real Rates 1311–2021 (NBER Working Paper 30475), Kenneth Rogoff, Barbara Rossi, and Paul Schmelzing examine a rich dataset on long-maturity sovereign debt over the past seven centuries and find that long-term global real interest rates have exhibited a persistent downward trend of about 1.6 basis points per year. Although evident in the long time series, this pattern is not found in data covering the last 150 years. Further, over the full data sample, real interest rates appear to be stationary around a declining trend. This contrasts with the finding in more recent data that they exhibit a significant random walk component and contributes to resolving a longstanding puzzle in the literature — namely, that although most economic models assume interest rates to be stationary, the empirical literature typically could not reject a unit root.
The researchers focus on long-maturity interest rates because they are more important than short-maturity rates for key economic variables such as investment, and because long-maturity debt has traditionally been the near-exclusive sovereign debt basis, and can therefore be measured over a longer historical time span. Italian city republics were issuing consolidated long-maturity debt by the 13th century, and a pan-European secondary market soon emerged. Until the late nineteenth century, roughly three-quarters of all sovereign consolidated debt consisted of long-maturity assets.
Seven centuries of data suggest a gentle trend decline in long-term real interest rates and indicate that sharp declines, such as the one after the global financial crisis, are usually due to cyclical factors.
The researchers study a comprehensive sample of long-maturity liquid and voluntary interest rates for various intervals over the 1311–2021 period in eight economies: Italy, the UK, Holland, France, Germany, Spain, the US, and Japan. The average maturity of the bonds being analyzed closely approximates modern 10-year Treasury bonds. The US and Japanese data series begin much later than those for other nations. In order to construct real rates, nominal interest rates are adjusted with a seven-year lagged inflation measure designed to approximate inflation expectations.
The researchers test for structural breaks, changes in the level of real interest rates, around five significant dates in economic and financial history: 1349, which corresponds to the Black Death; 1557, a year which saw a wave of European financial shocks; 1694, which captures the so-called revolution in “credible commitments” arising from institutional reform in Britain; 1914, which corresponds to the end of the multicentury fixed exchange rate system, the outbreak of World War I, and the founding of the Federal Reserve; and 1981, which many analysts point to as an inflection point in advanced economies. When interest rates are weighted by national GDP, the years 1349 and 1557 stand out for breaks in the trend rate of interest rate decline. The evidence does not point strongly to any recent break.
The researchers identify four eras of low real interest rates: prior to the Black Death in 1311–53, after the Great Bullion Famine in 1483–1541, during a credit boom in 1732–1810, and during the foreign exchange transition era of 1937–85. Each of these eras ended abruptly. They also find that, until very recently, there is little evidence of a positive correlation between real rates and either output or population growth, and if anything some support for a negative association, which is inconsistent with popular explanations of the recent real interest rate decline.
—Whitney Zhang
