Using the Consumption-(Dis)aggregate Wealth Ratio to Derive and Predict Future Expected Returns

Abstract

Lettau and Ludvigson (2001) analyzed the combination of both the financial and macroeconomic markets and found predictive power in the deviation of the common trend of consumption, asset wealth, and labor income on stock returns over short-term horizons\cite{paper1}. In their paper, Lettau and Ludvigson use the return on the S\&P 500 Index as a representation of asset returns. Instead, I estimate the present value of expectations of returns on aggregate wealth and consumption growth using a Vector Autoregression (VAR) which implements different conditional assumptions for the return on human capital, a decomposition of assets wealth with individualized proxies for their returns, the change in consumption growth, and the estimated trend deviation term, $\widehat{cay_t^D}$, as variables. Then, I move on the test the predictive power of the trend deviation term, $cay_t^D$ on different expectations of returns on aggregate wealth. The theoretical framework suggests that as long as expected future returns on human capital, and consumption growth are not too variable or if these variables are highly correlated with expected returns on assets, then $cay_t^D$ will be a good forecaster of expected future assets returns of the market. In this work, the empirical results do not coincide with the theoretical framework potentially due to many factors including unobserved errors in data collection, assumptions in the construction of return proxies, etc. Moreover, I provide some anecdotal evidence suggesting that the return of human capital, when factored into the return of aggregate wealth using either the growth rate of labor income or the return on a risky asset as its proxy, acts as a hedging device, insulating a agent against volatility and risk.