LAUSR

Abstract We study a series of growth models in which households’ preferences display ‘jealousy’ or ‘external habits’: a negative dependence on average consumption. We argue that accounting for consumption externalities in growth models requires consideration of both their static and dynamic effects. In a static sense, and as highlighted in the literature, individual households do not internalize the effect of their choices on average consumption. In a dynamic sense, and differently from most of the literature, households do not take into account the effect of changes in average consumption on the evolution of the (shadow-) value of their own wealth. Accordingly, the equilibrium paths of fully endogenous growth models involve inefficiently low savings and growth because of external habits. In semi-endogenous models, consumption externalities have no growth effects but rather level effects on the long-run research intensity. In exogenous growth models, externalities only slow down the convergence to balanced growth. We then use Consumer Expenditure Survey data to test the extent of external habits: our identification relies on a two-stage estimator that uses the Tax Reform Act of 1986 and the Omnibus Budget Reconciliation Act of 1993 as positive and negative consumption shocks to US top incomes, respectively. In the first stage, we use a difference-in-difference approach to exploit the exogenous variation in consumption due to federal tax reform. We then use the predicted values for average within-cohort consumption by income deciles as an instrument to estimate the extent of social preferences in an estimator that is directly comparable with the household policy function implied by our benchmark model. We find highly significant, long-run external effects ranging from 17% to 27%, depending on the model specification. We also find that consumption externalities are not statistically significant at the very top of the income distribution: this finding provides a rationale for saving rates increasing with incomes, and suggests a channel through which income inequality reverberates into wealth inequality.