Do very long-run data show continuity in short-term real interest rates?
Core argument: Multi-century data on short-maturity real interest rates show strong gravitation toward a long-run trend, relatively fast adjustment speeds, and few structural breaks, results the authors find survive out-of-sample tests.
We analyze empirical evidence on long-run samples of short-maturity real interest rates as well as term spreads based on multi-century data. We find strong and consistent evidence of [gravitation towards the long-run trend] in long horizon series, relatively fast adjustment speeds, and a paucity of structural breaks – results that survive out of sample tests. The use of very long-run data offers a fresh perspective for ongoing monetary policy debates surrounding r*, and also provides a crucial missing link to reconstructing the long-run properties of term spreads. On balance and against limited post-COVID data, our evidence suggests caution on the idea of a break in short-term real interest rate behavior and instead points to elements of continuity over very long time periods. Relatedly, we show that term spreads are secularly rising while inflation volatility trends in the exact opposite direction – a finding questioning the emphasis of influential term structure models. The faster decline in short-maturity yields [not a rise in long yields] accounts for rising term premia.



