Title: Consumption-Based Modeling of Long-Horizon Returns

Co-author: David Marshall, Federal Reserve Bank of Chicago

Status: Mimeo

Date of Last Revision: April, 1999

Abstract: Numerous studies have documented the failure of co-movements of returns and consumption growth to explain the equity premium puzzle. This failure has sometimes been attributed to frictions, transaction costs or durability. If such frictions are important, they should primarily affect the higher frequency comovements in consumption growth and asset returns. The long-swings, or lower-frequency comovements should be less affected. Consequently if transaction costs are important, tests of the consumption based asset pricing model which concentrate on lower-frequency components may be more successful.

We investigate this hypothesis using a variety of diagnostic tests. We first use coherence analysis and bandpass filtering analysis to show that, while there is a complete lack of correlation between asset returns and consumption growth at frequencies higher than about 0.7 (/year) (swings longer than 1.4 years), the coherence/correlation between the two series at lower frequencies is above 60%. We perform \citeN{hansen/jagannathan:91} bounds tests, chi-squared tests of moment restrictions, and \citeN{hansen/jagannathan:94} specification tests of two consumption-based models of the asset-pricing kernel: the \citeN{abel:90} ``Catching up with the Joneses" preferences, and \citeN{constantinides:90} habit-formation preferences. While neither model performs well at the quarterly horizon, both models perform well at longer horizons.

View/Print This Paper (Acrobat/pdf format -- 557 K)

Return to main menu.