Consumption, change in expectations and equity returns

Abstract

Using quarterly data for the period 1959 to 2008, I study the relationship between excess stock returns and the change in expectations of the consumption–wealth ratio and of future long-run consumption growth. Using a vector error correction model (VECM), I estimate revisions in expectations on the consumption–wealth ratio and on the discounted value of future consumption growth; the latter being of high relevance in the asset pricing literature but difficult to identify empirically. My findings show that these variables are strong predictors of future excess stock returns when compared to several common predictor variables. Furthermore, these results seem to be robust in out-of-sample and in-sample analyses, and appear not to be driven by persistence.

Keywords:

• consumption
• expectations
• consumption–wealth ratio

JEL Classification:

• E21
• E32
• G12

Notes

¹  For any particular asset, the Federal Reserve estimates first the participation of other economic agents and then attributes the rest to households. The Federal Reserve bundles households and nonprofit organizations together, so the net worth series is a measure of household and nonprofit organizations total wealth.

²  In addition to predicting future equity returns, Afonso and Sousa (2011) and Sousa (2012a, b, 2013) show that the consumption–wealth ratio and the wealth-income ratio can also predict future government yields.

³  In the DLS cointegrating equation, l = 1. The cointegrating vector is mainly unaffected by the choice of lag length.

⁴  Results are presented for efficient GMM, using a Newey–West weighting matrix.

⁵  Campbell et al. (1997) define the stochastically detrended interest rate as the current 3-month Treasury bill rate minus the past 12-month average.

⁶  For brevity, in this article I only provide results when using excess stock returns, since they are qualitatively similar to those when raw stock returns are used instead. All results can be provided upon request.

⁷  Campbell and Thompson (2008) further demonstrate that the correct way to interpret an R² is to compare it to the squared Sharpe ratio. They state that if an R² from a predictive regression is larger than the squared Sharpe ratio, then an investor can get a larger portfolio return based on the information obtained from such regression.

⁸  I follow Harvey et al. (1998) and correct for finite sample in the Diebold–Mariano statistic.

⁹  Goyal and Welch (2003, 2008) argue that if a model is to be considered a good predictor, then at the very least it should offer more accurate out-of-sample forecasts relative to the most naïve forecast: the prevailing mean of returns during the estimation period. In unreported results, I conduct Goyal and Welch’s (2003, 2008) suggested tests and confirm the superior out-of-sample predictive power of the change in expectations model. Results are available upon request.