Life -cycle consumption examined.

Abstract

This set of papers uses new data to construct an empirical measure of the average propensity to consume out of full wealth among older households. Consistent with original neoclassical life-cycle theory, full wealth is defined as the present value of all current and future lifetime resources. This definition implies a simple theoretical expression for the average propensity to consume, with consumption proportional to full wealth. Data now available in the Health and Retirement Study (HRS) allow construction of a credible household-level estimate of full consumption and full wealth. A primary advantage of the average propensity to consume out of full wealth is its theoretical simplicity. First, full wealth is more equally distributed than net worth. Second, the propensity to consume out of full wealth, when compared to the propensity to consume out of income or net worth, has less dispersion cross-sectionally, more consistency over time, and more invariance to circumstances or factors such as pension entitlement, income profile, and past income shocks. Finally, the consumption propensity out of full wealth responds less to unexpected changes in resources, specifically unexpected retirement, than the other consumption measures. Comparing the consumption propensity out of full wealth observed in the HRS to the prediction of a neoclassical model of consumption with uncertainty in mortality and asset returns suggests that heterogeneity in preferences, specifically time preference or the willingness to substitute inter-temporally, is necessary to explain the extent of observed cross-household heterogeneity within the model. Estimates of the model with returns to scale in household size, a fixed cost of working, and household-specific variation in expected asset returns only account for a small portion of the variation in the observed consumption rate. After controlling for subjective mortality, expected bequests, and demographics using survey data, the residual is correlated with the level of wealth and with variables that may indicate an inherent characteristic of patience. Exploring alternative explanations for the heterogeneity, cognitive and planning abilities can help explain the residual heterogeneity in the propensity to consume. In particular, measures of basic cognition, horizon of financial planning, basic financial literacy, and goal-oriented behavior is associated negatively with the unexplained variation in consumption rates. In other words, higher skilled households consume less after controlling for factors in the neoclassical model.