«Internal Inconsistency and Risk Aversion: Implications on Smoking Decisions ANA I. GIL-LACRUZ Dpto. de Economía y Organización de Empresa, ...»
ESTUDIOS ECONOMÍA APLICADA V O L. 29 - 1 2011 P Á G S. 1 - 18
Internal Inconsistency and Risk Aversion:
Implications on Smoking Decisions
ANA I. GIL-LACRUZ
Dpto. de Economía y Organización de Empresa, UNIVERSIDAD DE ZARAGOZA. SPAIN.
Dpto. de Psicología y Sociología, UNIVERSIDAD DE ZARAGOZA. SPAIN. E-mail:
ABSTRACTThe main contribution of this paper is an analysis of the nature of the link between internal coherence and risk aversion. Both variables play an important role in individual decisions concerning risk behaviors. We compare the levels of internal consistency and risk aversion among smokers and non-smokers. To measure the individual internal coherence and risk aversion, we use a survey that includes lottery questions. Our results confirm that smokers are consistent in their decisions and they behave as risk averse. These results should be treated with circumspection as lottery questions are based on monetary expectations that depend on socio-economic conditions and they obviate other dimensions such as social recognition.
Keywords: Risk aversion; expected utility; decision-making; risk behaviors.
Consistencia interna y aversión al riesgo: implicaciones en la decisión de fumar
RESUMENLa principal contribución de este artículo es analizar la naturaleza de la asociación entre la coherencia interna y la aversión al riesgo. Ambas variables juegan un papel principal en las decisiones individuales sobre comportamientos de riesgo, tal que como caso particular, comparamos los niveles consistencia interna y aversión al riesgo entre fumadores y no fumadores. Para medir la coherencia interna y la aversión al riesgo individual, recurrimos a una encuesta que incluye preguntas de loterías. Nuestros resultados confirman que los fumadores son consistentes en sus decisiones, y de hecho, se comportan como adversos al riesgo. Estos resultados hay que tomarlos con cautela, pues las preguntas sobre loterías se basan principalmente en expectativas monetarias que dependen de condiciones socioeconómicas obviando otras dimensiones como el reconocimiento social.
Palabras clave: Aversión al riesgo; utilidad esperada; toma de decisiones; comportamientos de riesgo.
JEL Classification: D81 The authors would like to thank the German Institute of Applied Economics (DIW) for providing the data base and financial support. The authors are also grateful to Jonathan Baron and the two anonymous referees of Estudios de Economía Aplicada for comments that led to an improvement in research quality. The usual disclaimer applies.
_________ Artículo recibido en octubre de 2009 y aceptado en octubre de 2010 Artículo disponible en versión electrónica en la página www.revista-eea.net, ref. ə-29101
1. INTRODUCTION The state of health is an accumulative process that depends on both the health goods and services that citizens receive and their life styles. Risk behaviors, such as smoking, heavy drinking and lack of physical exercise reinforce the incidence of illnesses that, in most cases, could be avoided. The burden of risk habits imposes such a high costs to the individual and society in general that preventive policies are increasingly important in the structure of public budgets.
Anti-drug policies are essentially twofold: supply-side measures (taxes and control of trafficking) and demand-side measures (new endowments of information). Indirect taxes represent one of the most commonly applied instruments used to control alcohol and tobacco demand: higher tax rates increase prices and reduce purchasing power (Lewit Coate, 1982; Pogue Sgontz, 1989; Keeler et al., 1993; Saffer Chaloupka, 1994; Crawford Tanner, 1995;
Chaloupka Wechsler, 1997; Crawford et al., 1999; Escario Molina, 2004). With regards to new endowments of information, informative campaigns are justified if consumers are currently misinformed about the characteristics of the goods they consume (Slovic, 2000; Duarte et al., 2006). For example, if smokers undervalue the dangers of tobacco products, once they are properly informed, they will reconsider their decisions about smoking. The new endowment of information will be effective if it changes the individual’s structure of preferences. The effectiveness of providing more information to prevent drug use is disputed. Viscusi (1990) affirms that drug-consumers make decisions whilst aware of the risks they are taking, even though they might not have a clear idea of the magnitude of these risks.
Risk aversion indexes are fundamental measures to evaluate the degree of aversion with which citizens make decisions. Risk aversion indexes are based on individual optimal risk levels on a scale of preferences. Experiments such as those that require the answers to lottery games, offer social researchers a powerful tool for calculating individual levels of risk aversion. In lottery questions, participants usually reveal their preferences for a fixed amount of money or a lottery ticket. The expected utility framework has been implemented to analyze risk attitudes and behaviors with data drawn from lottery games (Blondel et al., 2007; Dave Saffer, 2007; Sasaki et al., 2006; Schunk Winter, 2007; Wärneryd, 1996).
The main advantage of using lottery questions is that they are characterized by informative transparency and uncertainty. A lottery game is defined by outcomes and probabilities so participants foresee the occurrence of an outcome and make the corresponding decisions. If they were wrongly informed about Estudios de Economía Aplicada, 2011: 1-18 Vol. 29-1
INTERNAL INCONSISTENCY AND RISK AVERSIÓN: IMPLICATIONS ON SMOKING DECISIONSoutcomes and/or the probabilities, they would not be able to make correct decisions.
The authors believe that lottery questions are suitable for general indexes of
risk aversion because, for example, smoking also implies the two dimensions:
smoking generates illnesses to a degree of probability - a World Health Organization anti-tobacco campaign states that smokers have a 12-times higher probability of suffering from laryngeal cancer than non-smokers (WHO, 1998).
Citizens know that smoking is dangerous but they might be wrongly informed about the dangers of smoking and/or the probability of occurrence.
What might seem a relatively easy mathematical exercise becomes more complex in reality because individuals are not always consistent with their scale of preferences. If people fail in ordering their preferences, the endowment of information (warning labels, anti-drug commercials or informative pamphlets etc.) might cause the expected effects, but probably not in the desired dimension. Given that governments allocate important economic resources in health policies aimed at promoting healthy habits (especially informative strategies), the implications of this research are important for policy makers. If smokers make mistakes when ordering preferences, demand-side measures might have limited impact on reducing tobacco consumption it might therefore be more useful to implement supply-side measures, such as higher taxes or tougher trafficking controls.
The main contribution of this paper is to explore the nature of the link between individual internal inconsistency and the degree of risk aversion. Both variables play an important role in individual decisions concerning risk behaviors. We compare the levels of internal inconsistency and risk aversion among smokers and non-smokers. We use lottery questions to measure the individual internal coherence and risk aversion. We use data from the German Personality and Daily Life Survey (2004) because it gives important information on internal coherence and risk aversion indexes. Our results confirm that smokers are consistent in their decisions, and they even behave as risk averse. It is clear that they perceive themselves as risk averse.
The rest of the paper is structured as it follows: Section 2 outlines the theoretical framework; Section 3 deals with the data base; Section 4 summaries the main results and Section 5 concludes the work with the main findings of our research and the corresponding policy implications.
2. THEORETICAL FRAMEWORKAnyone who faces the choice of different options in a framework of uncertainty will consider the consequences of the choices and their corresponding probabilities. The expected utility theory offers a simple measure based on a set of information about a lottery L which is fully represented by
The expected utility hypothesis can be derived from three axioms: ordering, continuity and independence. The ordering axiom implies that individuals are able to determine their preferences from among different options, it requires completeness and transitivity. The continuity axiom determines that if there are three outcomes which are ordered by the level of preference, there is always a probability, a compound option with the most preferable and less preferable multiplied by this probability is indifferent to the second best option multiplied by one minus this probability. The independence axiom makes it possible to keep the structure of the preferences when relaxing the value of the probability (Lancsar Louviere, 2006; Starmer, 2000).
We assume, for reasons of simplicity, that the choices people face involve three different outcomes, y RH, y RL and yS ( yRH yS yRL ) thus u y RL 0, u yRH 1 and u yS x with 0 x 1. The expected utility criteria for choosing between two options, one that is of higher risk R yRH, pRH ; yRL, pRL ; 0.1 pH pL and one that is safer S yS, pS is
to select that option that is associated with a higher expected utility:
S R xpS pH (2) Expected utility theory helps economists to understand how people make decisions, it does, however, have some drawbacks, for example, in the context of probability weighting or loss aversion (Starmer, 2000). The fact that people exhibit inconsistencies when confronted with multiple options is not new (Hilton, 1989). Up to now the main contributions to the debate on expected utility and risk aversion lies in testing the hypotheses of expected utility theory with respect to income changes (McKee, 1989; Palacios-Huerta Serrano, 2006).
Researchers need to address the issue of inconsistency with expected utility to better predict individual behavior. To test theoretical hypotheses, researchers control experiments by altering the attributes of the game. In this paper we focus on individual inconsistencies with expected utility, assuming that money behaves as normal good for everybody, independently initial economic resources. To give an example of a failure in internal consistency, let’s assume that an individual has revealed preferences for the fixed amount of money, Estudios de Economía Aplicada, 2011: 1-18 Vol. 29-1