Excess Sensitivity and Asymmetries in Consumption: An Empirical Investigation

Preprint
Publisher Website
Google Scholar
    • preprint
      • Published in RePEc
Abstract
Most empirical studies on liquidity constraints classify a consumer as being constrained on the basis of a single indicator such as the asset to income ratio. In this analysis, we model the probability that a consumer faces liquidity constraints as a function of multiple social and economic factors. This probability function is estimated simultaneously with the degree of excess sensitivity of consumption to income in a switching regressions framework. The switching regressions apply optimal weights to the densities for the Euler equations on the two states and are less susceptible to sample misclassification. We are also able to use exclusion restrictions on the Euler equations for the constrained and the unconstrained individuals to discriminate between excess sensitivity due to liquidity constraints, from that due to myopic behaviour and a certain type of time non-separable preferences. Our results based on data from the CEX confirm that liquidity constrained consumers are excessively sensitive to variables already known to economic agents. However, there is evidence that the unconstrained consumers also exhibit behaviour that is consistent with the theoretical predictions. Further analysis suggests that such behaviour could be explained by time non-separable preferences. La plupart des études empiriques sur les contraintes de liquidité déterminent si un consommateur est contraint en fonction d'un indicateur unique comme le ratio des actifs sur le revenu. Dans la présente analyse, nous modélisons la probabilité qu'un consommateur subisse des contraintes de liquidité comme une fonction de plusieurs facteurs économiques et sociaux. Cette fonction de probabilité est estimée simultanément avec le degré de sensibilité excessive de la consommation au revenu dans un cadre de régressions à changement de régime. Les régressions à changement de régime appliquent des poids optimaux aux densités des équations d'Euler dans les deux états et sont moins susceptibles d'erreurs de
All Related Versions

This publication has 0 references indexed in Scilit: