Quantifying Consumer Preferences: Volume 288

In January 2009, the U.S. economy sits in its most precarious position since the Great Depression of the 1930s. The crash of the U.S. economy has reverberated throughout the world and adversely impacted virtually every other economic system on the globe. This sad fact is well known and undisputed by economists and social scientists throughout the world. That consumer behavior contributed heavily to this world, economic upheaval is also undisputed. What is less certain is exactly how consumer behavior ultimately contributed to the world economic collapse and what role behavior by consumers will play in ultimately lifting the world economy out of its current dire economic circumstances. Whatever that role ultimately turns out to be, it is clear that economists and other social scientists must understand better how consumers make decisions and what factors impact those decisions, in trying to fine-tune economic policy in order to deal with the current (and perhaps future) economic crisis.

This chapter is an up-to-date survey of the state-of-the art in consumer demand analysis. We review (and evaluate) advances in a number of related areas, in the spirit of the recent survey paper by Barnett and Serletis (2008). In doing so, we only deal with consumer choice in a static framework, ignoring a number of important issues, such as, the effects of demographic or other variables that affect demand, welfare comparisons across households (equivalence scales), and the many issues concerning aggregation across consumers.

This chapter reviews the specification and application of the Deaton and Muellbauer's (1980) almost ideal demand system (AIDS) and the Christensen et al. (1975) translog (TL) demand system. In so doing we examine various refinements to these models, including ways of incorporating demographic effects, methods by which curvature conditions can be imposed, and issues associated with incorporating structural change and seasonal effects. We also review methods for adjusting for autocorrelation in the models' residuals. A set of empirical examples for the AIDS and the log TL version of the translog based on historical meat price and consumption data for the United States are also presented.

This chapter presents the differential approach to applied demand analysis. The demand systems of this approach are general, having coefficients that are not necessarily constant. We consider the Rotterdam parameterization of differential demand systems and derive the absolute and relative price versions of the Rotterdam model, due to Theil (1965) and Barten (1966). We address estimation issues and point out that, unlike most parametric and semi-nonparametric demand systems, the Rotterdam model is econometrically regular.

This chapter presents the indirect preferences for all full rank Gorman and Lewbel demand systems. Each member in this class of demand models is a generalized quadratic expenditure system (GQES). This representation allows applied researchers to choose a small number of price indices and a function of income to specify any exactly aggregable demand system, without the need to revisit the questions of integrability of the demand equations or the implied form and structure of indirect preferences. This characterization also allows for the calculation of exact welfare measures for consumers, either in the aggregate or for specific classes of individuals, and other valuations of interest to applied researchers.

This chapter presents an exposition of the Generalized Fechner–Thurstone (GFT) direct utility function, the system of demand functions derived from it, other systems of demand functions from which it can be derived, and its purpose and the econometric circumstances that motivated its original development. Its use in econometrics is demonstrated by an application to household consumer survey data which explores the relationship between prices, on the one hand, and expected exogenous preference changers such as household size, schooling of heads of household, and other social factors, on the other.

A concise introduction to the normalized quadratic expenditure or cost function is provided so that the interested reader will have the necessary information to understand and use this functional form. The normalized quadratic is an attractive functional form for use in empirical applications as correct curvature can be imposed in a parsimonious way without losing the desirable property of flexibility. We believe it is unique in this regard. Topics covered include the problem of cardinalizing utility, the modeling of nonhomothetic preferences, the use of spline functions to achieve greater flexibility, and the use of a “semiflexible” approach to make it feasible to estimate systems of equations with a large number of commodities.

Lewbel and Pendakur (2009) developed the idea of implicit Marshallian demands. Implicit Marshallian demand systems allow the incorporation of both unobserved preference heterogeneity and complex Engel curves into consumer demand analysis, circumventing the standard problems associated with combining rationality with either unobserved heterogeneity or high rank in demand (or both). They also developed the exact affine Stone index (EASI) implicit Marshallian demand system wherein much of the demand system is linearised and thus relatively easy to implement and estimate. This chapter offers a less technical introduction to implicit Marshallian demands in general and to the EASI demand system in particular. I show how to implement the EASI demand system, paying special attention to tricks that allow the investigator to further simplify the problem without sacrificing too much in terms of model flexibility. STATA code to implement the simplified models is included throughout the text and in an appendix.

The chapter reviews and extends the theory of exact and superlative index numbers. Exact index numbers are empirical index number formula that are equal to an underlying theoretical index, provided that the consumer has preferences that can be represented by certain functional forms. These exact indexes can be used to measure changes in a consumer's cost of living or welfare. Two cases are considered: the case of homothetic preferences and the case of nonhomothetic preferences. In the homothetic case, exact index numbers are obtained for square root quadratic preferences, quadratic mean of order r preferences, and normalized quadratic preferences. In the nonhomothetic case, exact indexes are obtained for various translog preferences.

The standard approach in measuring demand responses and consumer preferences assumes particular parametric models for the consumer preferences and demand functions, and subsequently fits these models to observed data. In principle, the estimated demand models can then be used (i) to test consistency of the data with the theory of consumer behavior, (ii) to infer consumer preferences, and (iii) to predict the consumer's response to, say, new prices following a policy reform. This chapter focuses on an alternative, nonparametric approach. More specifically, we review methods that tackle the earlier issues by merely starting from a minimal set of so-called revealed preference axioms. In contrast to the standard approach, this revealed preference approach avoids the use of parametric models for preferences or demand. The structure of the chapter is as follows. First, we introduce the basic concepts of the revealed preference approach to model consumer demand. Next, we consider issues like goodness-of-fit, power, and measurement error, which are important in the context of empirical applications. Finally, we review a number of interesting extensions of the revealed preference approach, which deal with characteristics models, habit-formation, and the collective model.

This chapter discusses new developments in nonparametric econometric approaches related to empirical modeling of demand decisions. It shows how diverse recent approaches are, and what new modeling options arise in practice. We review work on nonparametric identification using nonseparable functions, semi- and nonparametric estimation approaches involving inverse problems, and nonparametric testing approaches. We focus on classical consumer demand systems with continuous quantities, and do not consider approaches that involve discrete consumption decisions as are common in empirical industrial organization. Our intention is to give a subjective account on the usefulness of these various methods for applications in the field.

International tourism is a major source of export receipts for many countries worldwide. Although it is not yet one of the most important industries in Taiwan (or the Republic of China), an island in East Asia off the coast of mainland China (or the People's Republic of China), the leading tourism source countries for Taiwan are Japan, followed by USA, Republic of Korea, Malaysia, Singapore, UK, Germany and Australia. These countries reflect short, medium and long haul tourist destinations. Although the People's Republic of China and Hong Kong are large sources of tourism to Taiwan, the political situation is such that tourists from these two sources to Taiwan are reported as domestic tourists. Daily data from 1 January 1990 to 30 June 2007 are obtained from the National Immigration Agency of Taiwan. The heterogeneous autoregressive (HAR) model is used to capture long memory properties in the data. In comparison with the HAR(1) model, the estimated asymmetry coefficients for GJR(1,1) are not statistically significant for the HAR(1,7) and HAR(1,7,28) models, so that their respective GARCH(1,1) counterparts are to be preferred. These empirical results show that the conditional volatility estimates are sensitive to the long memory nature of the conditional mean specifications. Although asymmetry is observed for the HAR(1) model, there is no evidence of leverage. The quasi-maximum likelihood estimators (QMLE) for the GARCH(1,1), GJR(1,1) and EGARCH(1,1) models for international tourist arrivals to Taiwan are statistically adequate and have sensible interpretations. However, asymmetry (though not leverage) was found only for the HAR(1) model and not for the HAR(1,7) and HAR(1,7,28) models.

In this chapter, we describe how random utility maximization (RUM) discrete choice models are used to estimate the demand for commodity attributes in quality-differentiated goods. After presenting a conceptual overview, we focus specifically on the conditional logit model. We examine technical issues related to specification, interpretation, estimation, and policy use. We also discuss identification strategies for estimating the role of price and non-price attributes in preferences when product attributes are incompletely observed. We illustrate these concepts via a stylized application to new car purchases, in which our objective is to measure preferences for fuel economy.

Equivalence scales are deflators (or “scales”) by which the incomes of different household types can be converted to a comparable, needs-adjusted basis. They are measures of intra-household sharing potentials and differences in family members’ needs (i.e., of adults vs. children). One strand of literature uses econometric approaches to derive equivalence scales from household expenditure and time-use data. Another strand uses survey responses of people to quantify equivalence scales directly. Equivalence scales are potentially useful in several areas such as welfare-system design, income taxation, measurement of poverty and inequality, and determining lost earnings damages. This chapter surveys the literature on equivalence scales and presents some applications.

The estimation of regression models subject to linear restrictions is a widely applied technique; however, aside from simple examples, the equivalence between the linear restricted case to the reparameterization and the substitution case is rarely employed. We believe this is due to the lack of a general transformation method for changing from the definition of restrictions in terms of the unrestricted parameters to the equivalent reparameterized model and conversely from the reparameterized model to the equivalent linear restrictions for the unrestricted model. In many cases, the reparameterization method is computationally more efficient especially when estimation involves an iterative method. But the linear restriction case allows a simple method for adding and removal of restrictions.

In this chapter, we derive a general relationship that allows the conversion between the two forms of the restricted models. Examples emphasizing systems of demand equations, polynomial lagged equations, and splines are given in which the transformation from one form to the other are demonstrated as well as the combination of both forms of restrictions. In addition, we demonstrate how an alternative Wald test of the restrictions can be constructed using an augmented version of the reparameterized model.