Go to main content
Formats
Format
BibTeX
MARCXML
TextMARC
MARC
DataCite
DublinCore
EndNote
NLM
RefWorks
RIS

Files

Abstract

In this study, we show how to estimate large consumer demand and household collective models using big microdata in an efficient way. The dissertation is composed of three chapters, where we reduce aggregation bias in consumer demand systems, nest heterogeneous habit in household’s demand models, and evaluate the degree of intrahousehold inequality within 2-3 person households in the United States.

In Chapter 1, we show how inferential errors due to inconsistent aggregation bias can be properly reduced. To manage dimensionality in consumer demand estimation, researchers' common practice is to aggregate elementary products to a higher level prior to econometric estimation. Inconsistent product aggregation, however, introduces bias to econometric estimates and policy-relevant inferences. We propose two alternative strategies for bias reduction. The first strategy uses the relative prices of elementary products as control variables in the aggregate demand. The second uses a residual-based instrumental variable method to achieve independence between the instrument and the residual.

In Chapter 2, we study the heterogeneity of habit strength in households’ demand for regular carbonated sweetened beverages (CSBs) and beer in the United States. A demand model that nests a smooth transition function is used to describe habit-based consumption patterns, revealing heterogeneous strengths of habits among households. We find that more habitual consumers, those with a strong preference for a particular product, are not as sensitive to price or expenditure as the aggregate population.

In Chapter 3, we examine the degree of intrahousehold inequality within 2-3 person households in the United States. Using structural collective model and household scanner data, we estimate the fraction of household resources that are consumed by husbands, wives and children. We find intrahousehold inequality exists in two-person (i.e. husband and wife) households but not in three-person (i.e. parents and one child under 5-year-old) households. The policy implication of this finding is two-fold: 1) this empirical finding suggests the current household income eligibility threshold for the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC Program, a social welfare program targeting at women and children) should be increased for two-person households so that women who are currently not eligible for WIC Program but should have been can be covered; 2) Cost-efficient public policies can be achieved by allowing eligibility thresholds to vary by household demographics (e.g. education level, employment status, etc.).

Details

PDF

Statistics

from
to
Export
Download Full History