Abstract

Using household panel data from the PSID, a semiparametric panel data model is estimated and examined for the presence of habit formation. Previous studies in this area were based on linear models, and the semiparametric model is an improvement over the latter in that it does not specify a particular functional form on the data generating process. Estimates of the model show evidence of habit formation for a substantial percentage of the observations. An examination of households by their demographic characteristics as well as by job status yields additional information and supporting evidence on the existence of habit formation.

Highlights

  • The results show no evidence of habit formation in the data.Dynan estimates the model with different instrument sets and an additional variable, but the results are qualitatively the same, with the data showing no evidence of habit formation

  • In the original parametric panel data model, there was no evidence of habit formation in household food consumption

  • Dynan simplifies of the Euler equation from the first order condition of the household’s problem into a log linear form, but this simplification may result in a loss of information that results in a weaker result for the evidence of habit formation

Read more

Summary

Introduction

Habits can enter the utility function either goods allows Frank to show that interpersonal comparisons additively in the following form, do matter when consumers make their choices about consumption bundles In such a framework, individuals will make consumption decisions that are suboptimal such that α ~ci,t = ci,t − ci,t −1. Carroll (2000) examines the consumer’s optimization problem with multiplicative habits, based on the model of Abel (1990), where a consumer’s utility depends on a ratio of current and past consumption Abel used this multiplicative form of habits in the household’s utility function and examined how asset pricing can be affected by habits and catching up with the Joneses preferences. If a household is more “habit bound”, it will be slow to change levels of consumption immediately

The Data
Estimation
Additional Estimation Methods
Coefficient Estimation Results
Households by Age Ranges
Households by Gender
Households by Race
Conclusion
Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call