Risk perception by consumers
Abstract: This article reviews the literature that has been written in the field of risk perception by consumers. This is a widely known subject and a lot has been written about it. The example of mortality risk and even more specific, the smoking risk perception will be highlighted and treated as a practical example. The government, as a regulating force, can affect the risk perception by consumers through informational programs or through controlling the risk. Although it is a hard task, the government will strive for efficiency by closing the gap between actual risk and risk perception.
Key words: Risk perception, mortality risk, government action.
Economics of Regulation and Competition
Introduction:
The literature review will focus on the risk perception of consumers. That is, in the field of risk perception, there exists some irrationality and biases on which the review will focus. For example, consumers tend to overestimate low-probability (but high-risk) events (Benjamin & Dougan, 1997; Viscusi, Harrington Jr. & Vernon, 2005). This is just one of the biases consumers tend to make while perceiving risk. Because risk is an uncertain and hard to quantify factor, it’s difficult for consumers to respond adequately to situations in which risk might occur. Besides, risk can have different values to different consumers. It is the perception of risk that is the main topic of this literature review.
Risk in general:
In economic theory, risk is often seen in three different dimensions: the final outcomes of different states of nature, subjective beliefs about the probabilities that each outcome will occur and the willingness to accept a certain level of risk (Gerking & Harrison, 2006). Immediately, one can see that the perception of risk could differ from the actual risk, since all individuals have subjective beliefs about the risk probability. Also, the willingness to accept a certain level of risk differs among individuals. There are so-called risk lovers as well as people who act in a more risk averse way. This difference than is the main basis for the peoples response towards risk and the action they take regarding certain risky events. Therefore, one can argue that the risk perception by consumers differs from the actual risk incurred by these consumers.
Other research contributed to this topic by adding some more reasons why risk perception among consumers might differ. For example, cultural similarity and personal relevance were mentioned as having an effect on risk perception (Carvalho et al., 2008). They demonstrated that when the risk of an event occurring is already high, people perceive their risk to be even higher if the threat originates from a culturally similar country. The overestimation of the risk is therefore larger, when the threat is coming from a culturally similar country or area. Personal relevance on the other hand tends to decrease the perceived risk and the accompanying behavioural intentions, thereby opposing the earlier mentioned effect of cultural similarity.
Since risk and uncertainty go hand in hand, risk perception also depends on the level and on the source of uncertainty. For example Jia, Dyer and Butler (1999) presume that there is no risk when there is no uncertainty involved. Uncertainty is a direct consequence of asymmetric information. Uncertainty comes from the human inability to access and assess all relevant information, when making a decision in life (Baz et al., 1999). The way people cope with this informational uncertainty, is the way people perceive this risky uncertainty and respond to it. Since different individuals have different access to certain information, the uncertainty perception and therefore the risk perception differs across individuals. Baz et al. (1999) also researched whether the time horizon of a decision has an influence on risk. They conclude that this time horizon in fact does influence decision situations under risk.
The relationship between the processing of information by consumers on the one hand and the behaviour it generates on the other hand has been the topic of research in multiple articles. Not only Baz et al. (1999) as mentioned above, but also for example Magat, Viscusi and Huber (1987) draw conclusions regarding this relationship. Individuals who will differ in risk perception will also differ in behaviour. Because these individuals have different opinions about risk in certain events, their action for these event will be different too. Also Weber and Bottom (1989) related risk perception, depending on risk preference which is the willingness to accept risk, to choice behaviour (Weber & Milliman, 1997).
Desjardins, Dionne and Fluet (2007) is another example of a research who establish this relationship, in a special case. They study the relation between predicted risk perception and driving behaviour in the case of impaired driving, meaning driving under influence. They conclude by showing that such a relationship exists.
Mortality risk:
Individuals tend to overestimate risk associated with low-probability events. In contrast, they also tend to underestimate risks in events which have a higher probability (Viscusi et al., 2005), where the probability depends on the frequency a certain event has occurred. For example, in the book by Viscusi et al. (2005), a figure is published where they show the difference between perceived risk and actual risk, regarding mortality and causes of deaths. In this diagram, it becomes clear that certain low-probability causes of deaths, such as tornadoes or floods, are overestimated by the individuals, when it comes to the number of deaths with these certain causes. People thus think that for example tornadoes take more lives than they factually do. A tornado has a relatively low probability of actually causing people to die, but this probability is being overestimated by individuals. The vice versa is also true. For example, diseases such as cancer, take relatively many lives compared to the other causes of death. However, the perception by individuals on the height of casualties, caused by for example cancer, is lower than the actual amount of casualties is. This cause of death, which has relatively high probability of taking lives, is underestimated by individuals. Again it shows, that individuals in general overestimate risk associated with low-probability events and underestimate the risk associated with high-probability events.
On the x-axis is the Actual Frequency of Death graphically shown in a logarithmic form. On the y-axis, there is theEstimated Frequency of Death All the dots are standing for causes of deaths, just as in the book by Viscusi et al. (2005), with an actual value and an estimated value. The actual value therefore represents the real and accurate risk of dying by a certain cause of death. The estimated value than represents the risk individuals think these causes of deaths have, or the risk perception by these individuals. If the actual risk equals the estimated risk, the scatter plot has to have a trend line which is here represented by the 45� line. If the dots in the diagram would show that unity relationship, than this would suggest consumers perceive risk accurately. Since a regression line to the scatter plot of dots above would definitely be flatter than the 45� line, the risk perception differs from the actual risk. It might even be better to show it with the use of an exponential relationship instead of a linear one, but for the interpretation and explanation of the figure this will not matter. In fact, it could be argued that in case of low-probability causes of deaths, consumers overestimate the risk, while the reverse is true for high-probability causes of deaths; these are in general underestimated by the consumers. This phenomenon has also been mentioned in the introduction of this literature review and is hereby graphically proven.
Although, this is the most dominant view in current existing literature on this specific subject, it has been shown that this phenomenon will not hold for every individual or group of individuals in every case (Benjamin, Dougan & Buschena, 2001). This article continued on the previous work by Benjamin and Dougan (1997), which is mentioned above. In this specific research, they hypothesized that people acquire and use information rationally. They also hypothesized that the dominant view on this topic of mortality risk will hold. Meaning that people have a strong tendency to overestimate the number of deaths from infrequent causes and underestimate those from more common causes. At this point, they divide the research into two parts: one where the survey group has to estimate population-wide deaths and the other where the survey group has to estimate the number of deaths among their own age groups. In the former case, the hypothesis holds and the dominant view is supported by this research. In the latter case though, the survey group perceived the risk in a more accurate and an almost completely unbiased way. The explanation for this deviating results can be found in that the scarcity of information induces people to learn considerably more about the factors that directly and significantly affect their lives than about less immediate issues. When the factors thus directly affect your live, you will acquire and use the scarce information in the most rational way. This supports the earlier discussed conclusion by Carvalho et al. (2008) that personal relevance decreases the risk perception to a more true and accurate level.
Another note regarding the dominant view on this topic, as can be seen above in the figure, comes from Viscusi, Hakes and Carlin (1997). Viscusi et al. (1997) researched the measure of mortality risks and show that part of the bias between risk perception and actual risk can be attributed to the measure of risk perception and therefore the bias can simply be reduced by using a better and more accurate measure of risk perception. They also showed that the risk perception estimates for common causes of deaths are more accurate than the estimates for rare causes of deaths. This comes from the fact that there is more, better and more transparent information available for common causes of death than for rare causes.
Gender differences is a topic that also has been thoroughly examined in the literature, with respect to the difference in risk perception. In general, men perceive risk, in this case mortality risk, as lower than women and they are also more likely to underestimate the risks (Lundborg & Andersson, 2007).
Smoking:
Mortality risk is a widely researched topic in current literature. Even more specific, a field where the attention of scientist has been drawn towards this topic, is the field of smoking. The mortality risk one incurs when smoking a cigarette, has been the topic of debate for somewhat years. Gender differences, which exists as is mentioned above, in the field of smoking does also exist. Significant gender differences in the perception of smoking mortality risk have been found (Lundborg & Andersson, 2008). Both the mortality risk involved with smoking, as the addictiveness of smoking has been topic of research in this paper. The results show that females perceive the mortality risk of smoking as greater than males. On the other hand, females also perceive the addictiveness of smoking as less than males. The fact that males perceive the mortality risk as less than females, is a conclusion that has already been drawn above and that is hereby supported again, in the field of smoking.
Risk perception, with respect to smoking and its bad consequences, seem to be high for every individual who smokes on a regular basis. These individuals seem to be aware of the risk involved with smoking, especially the risk of obtaining a fatal (lung) disease. This evidence has been there for the United States, but has also proven to be the case for another major developed country, namely Spain (Anto�anzas et al., 2000). They examined whether the specific risk information efforts and hazard warnings, as it is known in the US, had a particular effect unknown in Spain and thereby explaining some possible difference in risk perception among these countries. However, they found that strong risk awareness, when it comes to smoking, can also be found among Spanish individuals, despite a different smoking information regime. They therefore conclude that the strong smoking risk awareness among individuals can be generalized towards most developed countries.
The relation between smoking risk perception on the one hand and smoking behaviour, or better defined as smoking probability on the other hand, is also a topic of some research. This relation has been shown in more literature, as can be read above. In the case of smoking, results indicate that the risk perception of smoking (cigarettes) is overestimated and that this risk perception has a significantly negative effect on smoking probability (Hsieh & Liu, 1995). The higher an individual will estimate the mortality risk of smoking, the more unlikely it is this individual will actually begin to smoke. Since the risk perception is also overestimated in general, the smoking behaviour of people will be leaning towards non-smoking. This individual risk perception will rely on three different sources: prior belief, individual experience and, importantly, public information.
Government action / policy making:
The fact that risk perception by individuals relies, amongst some other factors, on public information, means that the risk perception can be altered and more or less controlled. Since the government can control the stream of public information and decide on what kind of information they want to make publicly available, they are the organization that can try to control for biases in the risk perception of the individuals. In an efficient market, such a bias will not exist. The perceived risk will match the actual risk and the individual consumers will be fully aware of the risk and have access to all necessary information they need to make good choices and to construct their behaviour in the most appropriate way. Although a perfect efficient market is a hard target for the government, it is something they should strive for. They can alter the risk perception either through informational programs or through controlling the risk (Viscusi, 1995). The effect on risk perception through providing public information is ambiguous. Depending on the character of the information that is generated and on the way in which this information affects the structure of perceived probabilities, the government can alter the risk perception (Viscusi, 1995). Changes in the risk level on the other hand, seem to have a much more unambiguous effect. The government has some working mechanisms to alter the individual risk perception and therefore, the public opinion on perceiving risk. It also has the power to execute this theoretical ideas in practice.
Although it looks rather straightforward in theory, in practice it is harder to understand and to adequately respond to the risk perception by consumers. For example in the case of the mortality risk, the government is aware of the fact that people tend to overestimate low-probability events and underestimate high-probability events. But how then, if at all, should the government respond to this misperceptions; that is an interesting question (Viscusi et al., 2005). One of the characteristics of a democracy is that the government action should reflect the interest of its people. Ideally, the government should be better informed and therefore steadier in perceiving the risk. It can, by correct policy-making, then try to change the risk perception towards a more accurate opinion on risk. The government can do this, because they can acquire more scientific information than an individual and so they are better informed about the actual risk of a certain event (Viscusi et al., 2005).
Conclusion:
This literature review focused on the risk perception by consumers. Since risk is a hard to quantify phenomenon and uncertainty is highly related with risk, there has been written a lot about this specific subject. The subjective beliefs one can have on certain events and the risk that these events might occur, leads to differences in risk perception among different individuals. Since every individual acts on its own, the different risk perceptions will lead to different actions and different behaviour. For example in the case of mortality risk, there is a clear sign of overestimating risks associated with low-probability events and underestimating risk associated with high-probability events. This is the general perception of the public, but among these individuals, the risk perception can also differ. For example in specific fields, like the smoking risk perception, there is a difference in risk perception, based on gender. Males perceive risk in general, in an other way than females do. The task the government has in this subject, is to adequately respond to the misperceptions in society. Since every individual can have a different risk perception, it is a very hard job for the government to treat every individual separately, to make sure the risk perception by these individuals match the actual risk and thereby creating an efficient market. The government can change the general risk perception by consumers, by constructing efficient policy. This is something the government should be keen on and should strive for.
Conclusion:
Anto�anzas, F., Viscusi, W.K., Rovira, J., Bra�a, F.J., Portillo, F. & Carvalho, I. (2000). Smoking Risks in Spain: Part I � Perception of Risks to the Smoker.Journal of Risk and Uncertainty, 21, 161-186.
Baz, J., Briys, E., Bronnenberg, B.J., Cohen, M., Kast, R., Viala, P. et al. (1999). Risk Perception in the Short Run and in the Long Run. Marketing letters, 10, 267-283.
Benjamin, D.K. & Dougan, W.R. (1997). Individuals’ Estimates of the Risks of Death: Part I�A Reassessment of the Previous Evidence. Journal of Risk and Uncertainty, 15, 115-133.
Benjamin, D.K., Dougan, W.R. & Buschena, D. (2001). Individuals’ Estimates of the Risks of Death: Part II�New Evidence. Journal of Risk and Uncertainty 22, 35-57.
Carvalho, S.W., Block, L.G., Sivaramakrishnan, S., Manchanda, R.V. & Mitakakis, C. (2008). Risk perception and risk avoidance: The role of cultural identity and personal relevance. International Journal of Research in Marketing, 25, 319-326.
Desjardins, D., Dionne, G. & Fluet, C. (2007) Predicted risk perception and risk-taking behavior: The case of impaired driving. Journal of Risk and Uncertainty, 35, 237-264.
Gerking, S. & Harrison, G.W. (2006). Risk Perception, Valuation and Policy: Introduction. Environmental and Resource Economics, 33, 267-271.
Hsieh, C. & Liu, J. (1995). Risk perception and smoking behavior: Empirical evidence from Taiwan. Journal of Risk and Uncertainty, 11, 139-157.
Jia, J., Dyer, J.S. & Butler, J.C. (1999). Measures of Perceived Risk. Management science, 45, 519-532.
Lundborg, P. & Andersson, H. (2007). Perception of own death risk. Journal of Risk and Uncertainty, 34, 67-84.
Lundborg, P. & Andersson, H. (2008). Gender, risk perceptions, and smoking behavior. Journal of Health Economics, 27, 1299-1311.
Magat, W.A., Viscusi, W.K. & Huber, J. (1987). Risk�Dollar Tradeoffs, Risk Perceptions, and Consumer Behavior. In W. Kip Viscusi and Wesley A. Magat Magat, W.A., Viscusi, W.K. & Huber, J. (1987). Risk�Dollar Tradeoffs, Risk Perceptions, and Consumer Behavior. In W. Kip Viscusi and Wesley A. Magat (eds). Cambridge: Harvard University Press.
Viscusi, W.K. (1995). Government Action, Biases in Risk Perception, and Insurance Decisions. The Geneva Papers on Risk and Insurance Theory, 20, 93-110.
Viscusi, W.K., Hakes, J.K. & Carlin, A. (1997). Measures of Mortality Risks. Journal of Risk and Uncertainty, 14, 213-233
Viscusi, W.K., Harrington Jr., J.E. & Vernon, J.M. (2005). Economics of Regulation and Antitrust. Cambridge, Massachusetts, USA. The MIT Press.
Weber, E.U. & Milliman, R.A. (1997). Perceived Risk Attitudes: Relating Risk Perception to Risky Choice. Management Science, 43, 123-144.