Search

Children, prime-aged adults, and the elderly vary in their consumption patterns, based on shifts in their needs, preferences and resources. As the composition of the United States population is changing, an understanding of the factors which influence consumption is important in determining what this means for consumer demand. The first chapter makes use the Consumer Expenditure Survey (CEX) extracts of Harris and Sabelhaus (2000) and Consumer Price Index (CPI) data, to investigate these consumption shifts which occur throughout an individual's lifetime. The Almost Ideal Demand (AID) system of Deaton and Muellbauer (1980a) provides the basic empirical specification from which this analysis proceeds. We find significant influences of demographic variables on consumption, and in particular that household demand for a particular commodity group is significantly influenced by the age composition of its members. The analysis contributes to existing literature through the use of multiple years of household level data and a theoretically plausible demand system for the case of the United States. The analysis is further expanded to illustrate the effects of households with non-purchasing behavior. A methodology proposed by Tauchmann (2005) is applied to correct for household selection into zero expenditures. Selection correction terms for all equations are highly significant, with estimates for commodity groups displaying large proportions of zero expenditures demonstrating the largest changes. These results indicate the importance of correcting for the selection process in demand analyses. Individuals pursuing higher education are faced with many decisions. In Chapter 2, the returns to college attendance are examined, with focus on the influence of location. The decision of whether to pursue a post-secondary education is a complex process with many components. Students must decide between competing colleges, weighing and comparing their different attributes. One characteristic which is factored into this decision is the geographic location of the institution. By using a merged data set from the Integrated Postsecondary Education Data System (IPEDS) and National Longitudinal Survey of Youth, 1979 Geocode (NLSY79 Geocode), the study investigate the returns to crossing state boundaries for college attendance. Results suggests there is a higher return to post-secondary education received out-of-state (OOS) for both males and females as compared with returns to attending an in-state institution under a Random Effects model. This suggests students may reap rewards from migration for the purpose of education. The estimates from a Fixed Effects model are similar. Finally, we explore specific consumption behavior in the case of alcohol. This commodity is chosen due to its unique considerations, as consumption of alcohol is strongly correlated with household and individual characteristics and potentially imposes societal burdens exceeding consumption. In particular, we examine the influence an individual's occupation has on the probability of alcoholism. Selected data from the National Longitudinal Survey of Youth, 1979 Cohort (NLSY79) is used, as it provides specific information regarding these behaviors. Results indicate there is no substantially meaningful relationship between occupation and the probability of alcoholism, although blue-collar workers tend to drink in higher quantities. Individual consumption behavior has multiple sources of influence, and can be modeled as a series of choices, from the decision of whether to consume, to where to consume, and levels of consumption. The purchasing and expenditure decisions for each commodity or bundle of goods are motivated by a variety of factors. Results of the analyses presented herein indicate a need for further research and emphasize the importance of demographics in investigating household decision-making behavior.