On the Persistence of Negative Real Rates: A Habit Formation Approach

Abstract

Negative real short-term rates occur with much higher frequency after the turn of the millennium than in the 40 years prior. This phenomenon corresponds with a significant drop in growth after 2000. In a standard representative agent consumption-based asset pricing framework with external habit, and persistent and heteroscedastic growth, we structurally link declines in the permanent component of growth to real short rates operating through a plausible utility curvature that is near two. Declines in the conditional variance of growth during this period play a very limited role in explaining real rates. We find imposing the assumption of habit formation models that surplus consumption is procyclical, is key to linking growth to real rate variation. Our framework does not appeal to an explicit monetary policy or recursive utility to explain real rate behavior. Using one-month rates from 1959 to 2020, predictions from our adapted habit formation framework explain over 40% real rate variance. After 1980, predicted rates explain 76% of real rate variation when compared to Michigan survey of expectations derived rate. Model real rate predictions correctly predict almost 80% of months when realized rates are negative—consistent with a permanent slowdown in growth after 2000. Our nominal short-rate predictions are near perfect.

Keywords: Negative real rates, Habit formation, State space, Consumption dynamics, Long run growth

JEL Classification: G1, G12, E21, E43, E47, C32