Using Factor Analysis to Identify Consumer Preferences and Understand Economic Valuation

De rationes disputandum est:

the profile of a “warm-glower”,

a note on consumer’s behavior and public policy implications

[*]

Abstract

In this paper, we want to investigate the issue why an economic agent chooses (or not) to privately contribute for a public good. “De gustibus disputandum non est”. However, “de rationes disputandum est”. We empirically identify, measure and discuss that peculiar psychological motivation (based on warm-glow) that spurs individuals to privately contribute for a public good. Using Nunes and Schokkaert (2002) computed factor scores, we estimate warm-glow and empirically capture the relationship between contributors’ warm glow motivation and a set of socioeconomic and attitudinal/behavioural variables. Econometric analysis allows us to explain and profile different motivational profiles, based on elicited WTP based on CV questionnaire. In particular, we empirically capture the profile of those respondents, whose contributions to the public good is motivated and inspired by warm glow. Finally, the paper discusses some implications of the presence of warm glow motivation for public policy. The latter point is very important because is related to the crucial theme about implementation of CV results.

Keywords:

willingness to pay; latent factor; ego driven warm glow; social oriented warm glow; public policy design.

JEL Classification: D11, D12, Q26

1

1. Introduction

“De gustibus disputandum non est” claim the ancient Latin proverb and the seminal economic paper by Becker and Stigler (1977)[1]. This means that every rational individual can rank preferences the way she prefers. No ex-ante, axiomatic ranking can be imposed to individuals’choices, taste formation and preferences’ structure. However, if one cannot discuss about individuals preferences’ structure, one can investigate about the inner motivations that drive individuals to choose. Therefore, paraphrasing the Latin motto, we can add that “de rationes disputandum est”. It is possible to discuss about inner motivations that inspire economic agents’ choices, actions, taste and preferences. Investigating economic agents’ motivations to contribute for the private provision of a public good is the topic of the paper. We do not address the issue whether an economic agent wants to contribute (or not) for a public good; whether that agent expresses (or not) a preference (in monetary terms) for the good. We want to investigate the issue why an economic agent chooses (or not) to privately contribute for a public good. “De gustibus disputandum non est”. However, “de rationes disputandum est”.

In a seminal paper, Andreoni (1989, p. 464) states that “When people make donations to privately provided public goods, such as charity, there may be many factors, influencing their decisions other than altruism. As Olson (1965) noted, “people are sometimes motivated by a desire to win prestige, respect, friendship, and other social and psychological objectives” (p.60), or as Becker (1974) observed, “apparent “charitable” behaviour can also be motivated by a desire to avoid scorn of others or to receive social acclaim (p. 1083). Clearly social pressure, guilt, sympathy, or simply a desire for a “warm glow” may play important roles in the decision of agents. While such warm-glow giving has been acknowledged in the literature, the most common approach has been to assume that preferences depend only on private consumption and the total supply of the public good and not on the individual donations per se”.

In this paper, we want to empirically identify, measure and discuss that peculiar psychological motivation (based on warm-glow) that spurs individuals to privately contribute for a public good. In our framework, “warm glow” is defined as the feeling of moral satisfaction that economic agents experience when contributing to the private provision of public goods. Warm glow is the enjoyment derived by contributors to public goods, when they consider the benefits realized by the beneficiaries, since, in Andreoni’s framework, giving produces a pleasurable feeling.

In our opinion, understanding and targeting the warm-glow effect allows the researcher to be more rigorous when attempting to analyze economic agents’ behaviour and choices. Understanding, targeting and measuring the inner motivations that spur actions and choices might allow to better design and implement public policies.

The paper starts from the work and the dataset gathered by Nunes and Schokkaert (2003). The authors report the results from a contingent valuation (CV) study designed to investigate the influence of warm glow in willingness to pay (WTP) responses for the provision of a public good (a National Park in Portugal). The authors measure inter-individual differences in warm glow motivations through a factor analysis, performed on a list of attitudinal items. Socioeconomic variables and motivational factor scores are significant in the explanation of the individual WTP measures. The authors compute “cold” WTP measures by taking out the effect of the warm glow motivation.

In this paper, we extract Nunes and Schokkaert computed factor scores, in order to estimate warm-glow and empirically capture the relationship between contributors’ warm glow motivation and a set of socioeconomic and attitudinal/behavioural variables. Econometric analysis allows us to explain and profile different motivational profiles, based on elicited WTP based on CV questionnaire. In particular, we empirically capture the profile of those respondents, whose contributions to the public good is motivated and inspired by warm glow. We name “warm-glowers” that type of contributors.

There is a large body of experimental literature trying to identify/profile types of contributors (and related motivations for contributing). When studying the issue of the private provision of public goods (PG), standard economic theory (based on the model of a selfish and perfectly rational representative agent) is able to profile and explain the free-rider behaviour. However, standard economic theory is unable to account for many phenomena of giving, like, for instance, the existence of co-operators, over-contributors, and reciprocators, observed in experiments. Quoting Burlando and Guala (2005): “The most obvious departure from the standard approach consists in relaxing the selfishness and the full rationality assumptions”. This can be done by introducing altruism and error parameters (Palfrey and Prisbey 1996, 1997). Empirical estimates of these augmented models however have discovered a considerable heterogeneity among the subjects participating in PG experiments. The implications and relevance of heterogeneity has generated a lively discussion of PG experiments pointing in this direction include, among others, Andreoni and Miller (2002), Fehr and Gachter (2000), Fischbacher et al. (2001).

We can highlight two relevant streams of literature for warm-glow motivation:

(a) papers trying to perform experiments in order to identify and measure the extent of warm glow (among the others, Palfrey and Prisbey, 1996, 1997; Andreoni and Miller, 2002);

(b) papers trying to perform experiments in order define and identify contributor types and estimate their proportion in a subjects’ population (among the others, Fischbacher et al. ,2001; Keser and van Winden (2000).

This work follows the above literature only because it adopts the economic definition of warm glow, as defined by the seminal paper by Andreoni (1989).

This paper totally differs from this stream of research for several reasons. To our knowledge, it is the first attempt to “profile” respondents with particular motivational characteristics (warm glowers) based on elicited WTP, based on CV questionnaires. The consumer motivational profile (warm-glow) is given a quantitative dimension by making the use of factor analysis. Moreover, we do not only “separate”, identify and attach measure a particular motivational category (warm-glowers), but we look further into the “motivational box”, trying to empirically capture the relationship between individual’s socio-economic characteristics and individual’s warm glow motivational latent construct. Finally, the paper discusses some implications of the presence of warm glow motivation for public policy. The latter point is very important because is related to the crucial theme about implementation of CV results. We discuss possible answers to several questions: When a typical warm glow profile is identified, and once we have disentangled the warm glow valuation component from the reported WTP responses, what valuation results should be used in cost-benefit exercises? The reported WTP values? Or alternatively, the dry WTP results, which are corrected for warm glow? Or, in a broader context, how should a policy maker behave? Should reported WTP values be respected, or not? And what are the respective implications in terms of public policy design? Why should reported WTP be corrected for warm glow? Isn’it a factor that increases utility and from a welfare perspective what counts is aggregate utility? So every increase in utility should be counted when considering a change according to a standard cost-benefit analysis. The same is true for altruism. The usual argument against counting altruistic preferences in benefit aggregation is double counting. But isn’t it the case that in a more altruistic society, welfare is larger all things equal? Why therefore should there be a need to correct the naturally expressed preferences whenever they are accurately measured?

The remainder of the paper is organized as follows: section 2 presents an overview on Andreoni’s theoretical model on impure altruism. Section 3 reports the methodology that underpins the identification of the latent warm glow motivation. Section 4 reports the empirical results and section 5 contains a discussion regarding the assessment of the profile of the warm glower. Discussion about public policy implications, when considering warm glowers in the provision of the public good is provided in section 6. Section 7 concludes.

2. Theoretical framework: Andreoni’s impure altruism model

Andreoni (1989; 1990) developed a general model of private contributions to provide public goods that includes impure altruistic considerations derived from giving. This model assumes that individuals contribute to the provision of a specific public good for two reasons. First, because they simply want more of it. Second, because they derive some private benefits from giving to the good. The latter reason signals the presence of impure altruistic considerations and it is often interpreted in terms of warm glow or satisfaction derived from the act of giving per se. In this perspective, the contribution to the good enters an individual’s utility function twice: once as a private good and once as a public good.

In this section, we present an overview of Andreoni’s model to provide a theoretical framework for our work. Following Andreoni, individuals may have different preference mappings with respect to private provision of public goods and therefore a number of utility functions can be derived to take account of possible different underlying preferences. Suppose that Ui is the utility function of individual i (i = 1,…,n); xi is the private goods consumption of individual i, G is the level of provision of (services from) the public good and gi is the individual’s donation to the provision of the public good. The sum of the individual contributions equals the total supply of the public good:

(1)

From these assumptions, we can distinguish three different types of utility functions, corresponding to three different possible underlying preferences and related psychological profiles:

(2) Ui = Ui (xi, G)

(3) Ui = Ui (xi, gi)

(4) Ui = Ui (xi, G, gi)

Equation (2) describes the utility function of a pure altruistic individual, where there are no warm glow benefits from contributing to the public good G. For such individual, the private donation gi only enters the utility function as part of the level of public benefits G. This utility function profiles a purely altruist contributor, whose motivation, only depends on the desire to provide the good for the sake of the good existence. The contributor cares about the well-being of the recipients.

Equation (3) describes the utility function of a pure egoistic individual, where private benefits are the only benefits received by the individual i. Such individual does not consider the total level of public good G, but only his personal contribution (and benefit) for the provision of G. In this case, the contributor’s motivation might only depend on the possibility (or not) to make the private use of the provided public good. The contributor does not care about the well-being of the recipients.

Finally, equation (4) describes the utility function of an individual, where both the level of G and the amount of gi become explicit arguments in individual i’s utility function. The contributor cares about the well-being of the recipients but also derives pleasure form giving per se. Andreoni (1989) calls impure altruism this combination of pure altruism and warm glow This latter motivational profile can be described and explained in terms of different elements, such social pressure, feelings of sympathy and moral satisfaction, or the simply the desire for a warm glow.

3. Identifying the warm glow contributor profile

The present study focuses on the impure altruistic model formulation. The purpose of our research work is twofold. First, we aim at “looking into warm glow”, in order to identify the elements affecting such psychological attitude towards private contributing for a public good. Second, we want to get some empirical information about the profile of people having a utility function as described by Equation (4).

In this section we provide an econometric analysis of individuals’ warm glow motivational profile. In order to do so, we regress warm glow motivation against a series of variables containing information about the respondents’ characteristics.

In formal terms we propose to estimate the following equation,

(5) =+ +for all i-respondents

where , the dependent variable, is the vector with all the individual observations for the level of warm glow motivation, is a vector with the observations of all individuals for the socio-economic characteristic k, is a vector with disturbance terms and a’s are coefficients to be estimated.