By Carlos Viana de Carvalho, Andrea Ferrero, and Fernanda Necchio
The demographic transition can affect the equilibrium real interest rate through three channels. An increase in longevity – or expectations thereof – puts downward pressure on the real interest rate, as agents build up their savings in anticipation of a longer retirement period. A reduction in the population growth rate has two counteracting effects. On the one hand, capital per-worker rises, thus inducing lower real interest rates through a reduction in the marginal product of capital. On the other hand, the decline in population growth eventually leads to a higher dependency ratio (the fraction of retirees to workers). Because retirees save less than workers, this compositional effect lowers the aggregate savings rate and pushes real rates up. We calibrate a tractable life-cycle model to capture salient features of the demographic transition in developed economies, and find that its overall effect is a reduction of the equilibrium interest rate by at least one and a half percentage points between 1990 and 2014. Demographic trends have important implications for the conduct of monetary policy, especially in light of the zero lower bound on nominal interest rates. Other policies can offset the negative effects of the demographic transition on real rates with different degrees of success.
Interesting paper in the light of the recent discussion about the decrease in the real interest rate across developed economies. There is no doubt that demographics have a role to play. One has to be careful, though, as there are many interest rates out there, and they may be also composed of various risk premia and influenced by broader portfolio considerations than aggregate savings. Still, the paper shows that the demographic factors can account for a very substantial decrease in interest rates.
Thanks for picking up our paper. You’re absolutely right that there are many interest rates out there and many other factors influence them. We like to think of demographics as a common factor that works at low frequency. We are doing empirical work to check if that prediction can be teased out of the data.