Consumption, Income, and Interest Rates: Reinterpreting the Time Series Evidence

Abstract

This paper prcçoses that the tiie-serieS data on ocnsuiTptiOfl, income, ar interest rates are best viewed as generated not by a sfrle representative conswtr bit by two gr4 of cons'.utrS.Half the consumers are forwaxd-loOkirq ax consume their permanent irxt, bit are extremely reluctant to substitute consumption thtertanporafly.Half the consutrS foll the "rule of thumb" of consuming their current incone.The paper documents three empirical rularities that, it axues, are best explained by this medal.First, expected daxes in iie are associated with expected dianges in cx,nsumptiOn.Secon, expected real interest rates are not associated with expected dwges in consumption.Third, perieds in which ocnsulTption is high relative to irsui are typically followed by high growth in ine.The paper concledes by briefly discussing the iiplicatiOns of these firxings for eocnanic policy ar economic research.