Capital, Interest, and Aggregate Intertemporal Substitution

Financial economics research has suggested that expected returns are not the same across assets, and that their movements over time are not simply described or explained. I argue that this suggestion has implications for the study of substitution over time --namely that 'the' interest rate in aggregate theory is not the promised yield on a Treasury Bill or Bond, but should be measured as the expected return on a representative piece of capital. Furthermore, tax policy changes can be used to distinguish intertemporal preferences (a.k.a., the supply of savings) from the demand for capital, and to distinguish Fisherian from non-Fisherian interpretations of consumption-interest rate comovements. I so measure the interest rate using U.S. national accounts data, and find consumption growth to be both interest elastic and forecastable with tax and interest rate variables. In other words, standard proxies for 'the intertemporal marginal rate of substitution' and for 'the marginal product of capital' move together, once we allow for taxes, even though both fail to be significantly correlated with any particular financial asset's return