The Perception of Own Death Risk from Wildlife-Vehicle Collisions: the case of Newfoundland s Moose

Transcription

1 The Perception of Own Death Risk from Wildlife-Vehicle Collisions: the case of Newfoundland s Moose Roberto Martínez-Espiñeira Department of Economics, Memorial University of Newfoundland, Canada and Henrik Andersson Toulouse School of Economics, University of Toulouse Capitole, Toulouse, France February 15, 2017 Keywords: Bayesian learning; Moose vehicle collisions; Mortality risk; Road-traffic Extended abstract for ITEA Introduction If individuals have accurate perceptions about risks, i.e. knowledge about the true levels of risks they face, they will be able to make well-informed decisions and expose themselves to an optimal risk level. Numerous studies have examined individuals perception of mortality risk (Slovic, 2000), and the empirical evidence suggests, though, that individuals misperceive mortality risks. This bias does not only influence individuals ability to make well-informed decisions, it may also result in policy makers not allocating resources in an optimal manner, since they may base their decisions on objective risk measures or experts assessments. Biased risk perception may also have an impact on preference elicitation. For instance, if individuals overassess mortality risks, monetary estimates of reducing mortality risks may be positively biased (Gayer et al., 2000; Bleichrodt and Eeckhoudt, 2006). Hence, knowledge about individual risk perception is important not only from a research perspective, but also from a policy perspective. Most studies of risk beliefs examine how individuals perceive the risk of different hazardous activities for a given population and the evidence from them strongly suggests that individuals over- and underassess the probability of low and high risk events, respectively (Andersson, 2011). 1 The Bayesian learning framework has been used to analyze this robust finding of over- and underassessment of low and high level risks. The framework provides an analytical tool to examine how individuals form their risk 1 References and the description of methods are kept at a minimum due to space constraints. For a fuller description of the findings in the literature and the Bayesian updating process, see, e.g., Andersson (2011). 1

2 perceptions and it has shown that the over- and underassessment found is in line with a rational learning process (Viscusi, 1985; Hakes and Viscusi, 1997). Since risk information is both costly and limited, individuals will only partly acquire the information on risk levels and the risk bias will persist. Hakes and Viscusi (2004), Andersson and Lundborg (2007), and Andersson (2011) extended the analysis based on the Bayesian learning framework by using individual-level data. This enabled them to examine how individual risk perception and bias are affected by socio-economic and demographic factors. Based on their analysis they found that several socio-economic and demographic attributes, for instance education, income, health status and gender, indeed influenced individuals risk perception. Moreover, Andersson and Lundborg (2007) and Andersson (2011) who examined individuals perception of their own risk compared to a population average risk in previous studies, found that a similar pattern for road-traffic risk as in the literature, i.e. an over- and underassessment for low- and high-risk groups, respectively. Based on evidence that individuals are more optimistic about risks that they can control, it would had been likely that individuals would underassess road-traffic risk. The findings in Andersson and Lundborg (2007) and Andersson (2011) were therefore an important contribution. In this study we further examines individuals perception of their own mortality risk by employing the Bayesian learning framework on a Canadian data set. This data set contains individual-level data on the perception of own moose vehicle collision (MVC) road-traffic mortality risk and socio-economic and demographic characteristics. Our aim is to analyze individuals perception of their own road-traffic mortality risk and our specific objectives are to examine if: (i) perceived risks differ from objective risks, (ii) the probability of underestimation varies in terms of demographic characteristics, (iii) there is any correlation between the magnitude of bias and individual characteristics, and (iv) the risk perception formation of own risk follows the pattern found in Lichtenstein et al. (1978). We follow the standard in the literature and use the probability of death as the measure of the objective risk. 2 Method Reasons for the difference in lay people s subjective risk beliefs and objective risk measure are individuals known difficulties of judging small probabilities (Kahneman and Tversky, 1979; Kahneman et al., 1982), and that lay people are more influenced in their perceptions by media coverage, own experience, etc., of the hazards (Slovic, 1987). This study examines individuals perception of their own traffic-mortality risk and the empirical evidence suggests a dependence on age and gender, with females and older respondents having a higher perception of risk (Rafaely et al., 2008; Rosenblom et al., 2008). Moreover, the empirical evidence also suggests that individuals will underassess this risk for two reasons: (i) exposure to the risk is to some degree voluntary, which means that it is perceived as less troublesome compared to 2

3 Figure 1: Nature of updating process. Source: Viscusi (1992) Perceived risk Perceived risk = Actual risk Posterior assessment Prior assessment Actual risk involuntarily incurred risks (Sunstein, 2002), and (ii) optimism bias is known to be greater for risks to oneself and risks that are perceived controllable by one s own actions (Weinstein, 1989). Further, we also expect the optimism bias to be larger for men than for women (DeJoy, 1992), and larger for younger than for older male drivers (Matthews and Moran, 1986; Glendon et al., 1996). Despite this evidence suggesting optimism bias for traffic risk, Andersson and Lundborg (2007) and Andersson (2011) did not find any support for a systematic optimism bias, but results in line with a Bayesian updating process, and that the sign and size of the bias depended on gender and age. The basic concept of the Bayesian updating process is illustrated in Figure 1. Individuals prior risk beliefs are represented by the horizontal line. When they obtain new information, for instance through campaigns or their own experience, they update their beliefs. If the new information resulted in perfectly informed individuals, the perceived risk would be represented by the 45 degree line. However, empirical evidence suggests that learning is only partial, and, hence, individuals overassess and underassess low and high probability events, respectively, as represented by the unbroken line (Hakes and Viscusi, 2004; Lichtenstein et al., 1978; Armantier, 2006). 3 Data and selected preliminary results The data used for the analysis were obtained as part of a contingent valuation phone survey, whose primary aim was to estimate the benefits of MVC risk mitigation policies in Newfoundland. 2 The 2 Further details based on the first wave of fieldwork about the survey effort can be found in the larger report by Martínez-Espiñeira and Lyssenko (2014). 3

4 questionnaire was pilot-tested on 150 individuals and, after some minor adjustments, administered in two waves. During the first wave of interviewing, 7,599 households were reached. Out of the 1,449 respondents who agreed to start answering the survey, those under 19 years of age and those who lived in Newfoundland for less than six months were eliminated as non-eligible after a couple of brief screening questions or thanked because they would have overfilled the quota aimed for the share of residents in the Avalon Peninsula. Several more terminated the interview after these preliminary questions. The number of respondents who completed the questionnaire was in the end n = 150 in the pretest and n = 1,207 in the final survey, totalling N= 1,357, with a response rate of around 19% in both cases. For the second fieldwork wave, 957 households were reached and after determining that 11 were ineligible, cooperation was obtained from 265 households, yielding a response rate of 28%. 3. In the end 207 complete observations were obtained from the second wave of interviewing. Table 1 shows results from an OLS regression (OLSRM) where the dependent variable is the logarithm of mortality risk perception, and a conditional mixed process modeling (M8cmp) where a joint model on risk perception and accident experience of hitting a moose (hitmoose) is run. The latter has been run using the Stata cmp command (Roodman, 2011). The advantage of the conditional mixed process model is that it accounts for any endogeneity introduced by multiple choices made by the respondents. Regarding the results we here focus on RM and logtraf, but just to mention one of the other findings, we find males to state a significantly lower level of self-perceived risk. This falls in line with a commonly found effect found in the literature about gender and risk perception. But, back to the two variables of particular interest, RM and logtraf. The former variable RM is the objective MVC mortality risk presented to respondents in the survey. It was varied around the true statistical 10-year risk (a 10- year period was used in line with other studies to make the probability higher with the aim to increase understanding of the risk levels among respondents) and randomly assigned to different respondents. The objective of varying the levels was to examine the effect of providing information in line with a Baysian learning process. The variable logtraf defines the logarithm of the true overall road-traffic risk based on age and gender. It was included to examine whether respondents overall road-traffic risk influences their perception of dying in a MVC. As expected, individuals are positively affected in their responses about their own risk by the objective level of risk suggested by the interviewers (RM) as the average death risk due to a MVC in the province. Moreover, when taking into account the endogeneity in the hitmoose variable we also find a positive statistical significant effect from objective road-traffic risk on risk perception. 3 Only frequent drivers were eligible for the second wave s survey instrument, so interviews with a further 39 individuals had to be terminated 4

5 Table 1: Regression analysis of perceived MVC mortality risk 4 Discussion The (preliminary) findings from this study suggest a positive correlation between respondents objective risk, as defined by the objective risk for their age and gender group, and their perception of the risk of a fatal MVC. Moreover, the analysis also suggest that information on objective risk levels in the survey has an influence on respondents risk perception. On the one hand this indicates that individuals are not well-informed about the risk they face when it comes to MVC, but on the other hand it also suggests that information campaigns can be effective in updating people s risk perception to be more accurate. 5

6 References Andersson H. Perception of own death risk: An Assessment of Road-Traffic Mortality Risk. Risk Analysis. 2011;31(7): Andersson H, Lundborg P. Perception of own death risk: An analysis of road-traffic and overall mortality risks. Journal of Risk and Uncertainty. 2007;34(1): Armantier O. Estimates of own lethal risks and anchoring effects. Journal of Risk and Uncertainty. 2006;32(1): Bleichrodt H, Eeckhoudt L. Willingness to pay for reductions in health risks when probabilities are distorted. Health Economics Letters. 2006;15(2): DeJoy DM. An Examination of Gender Differences in Traffic Accident Risk Perception. Accident Analysis and Prevention. 1992;24(3): Gayer T, Hamilton JT, Viscusi WK. Private Values of Risk Tradeoffs at Superfund Sites: Housing Market Evidence on Learning About Risk. Review of Economics and Statistics. 2000;82(3): Glendon AI, Dorn L, Davies DR, Matthews G, Taylor RG. Age and Gender Differences in Perceived Accident Likelihood and Driver Competences. Risk Analysis. 1996;16(6): Hakes JK, Viscusi WK. Mortality Risk Perceptions: A Bayesian Reassessment. Journal of Risk and Uncertainty. 1997;15(2): Hakes JK, Viscusi WK. Dead Reckoning: Demographic Determinants of the Accuracy of Mortality Risk Perception. Risk Analysis. 2004;24(3): Kahneman D, Tversky A. Prospect Theory: An Analysis of Decision Under Risk. Econometrica. 1979;47(2): Kahneman D, Slovic P, Tversky A. Judgement under Uncertainty: Heuristics and Biases. Kahneman D, Slovic P, Tversky A, editors. New York, NY, USA: Cambridge University Press; Lichtenstein S, Slovic P, Fischhoff B, Layman M, Combs B. Judged Frequency of Lethal Events. Journal of Experimental Psychology: Human Learning and Memory. 1978;4(6): Lundborg P, Andersson H. Gender, risk perceptions, and smoking behavior. Journal of Health Economics. 2008;27(5):

7 Martínez-Espiñeira R, Lyssenko N Valuation of a moose-vehicle accident mitigation policy in Newfoundland. Centre for Applied Research (CARE), Department of Economics, Memorial University of Newfoundland, 2014, Final Report. 414 p. Available at rmartinezesp/mvcprojectjune15th2014.pdf. Matthews ML, Moran AR. Age Differences in Male Drivers Perception of Accident Risk: The Role of Perceived Driving Ability. Accident Analysis and Prevention. 1986;18(4): Rafaely V, Meyer J, Zilberman-Sandler I, Viener S. Perception of traffic risks for older and younger adults. Accident Analysis and Prevention. 2008;40: Roodman DH. Fitting fully observed recursive mixed-process modles with cmp. Stata Journal. 2011;11(2): Rosenblom T, Shahar A, Elharar A, Danino O. Risk perception of driving as a function of advanced training aimed at recognizing and handling risks in demanding driving situations. Accident Analysis and Prevention. 2008;40: Slovic P. Perception of Risk. Science. 1987;236: Slovic P. The Perception of Risk. London, UK: Earthscan; Sunstein CR. Risk and Reason. Cambridge, UK: Cambridge University Press; Viscusi WK. Are Individuals Bayesian Decision Makers? American Economic Review. 1985;75(2): Weinstein ND. Optimistic Biases About Personal Risks. Science. 1989;246: