Irrational Economist_ Making Decisions in a Dangerous World - Erwann Michel-Kerjan [56]
As noted earlier, another source of uncertainty is ambiguity about the particular time of an event: We do not know exactly when during the next twenty years the earthquake will occur. Seismic studies may provide a probabilistic estimate about when it is likely to happen, but there is still doubt about the exact timing—it can happen anytime within the next twenty years or even later. How would our homeowner react to this imprecise information about the timing of the catastrophic event? Behavioral evidence suggests that she prefers to know exactly when the event will occur and is even willing to pay for acquiring this information.
ATTITUDES TOWARD AMBIGUITY
Ambiguity can be defined as “the subjective experience of missing information” whereby the incomplete information prevents us from making a safe prediction. Mainstream economic analysis assumes that probabilities either are precisely known (as in roulette) or can be inferred from objective judgments (as in weather forecasts). From a behavioral standpoint, this is not necessarily true. Risk (known probabilities) and uncertainty (unknown or incompletely known probabilities) may lead to different behaviors. As Ellsberg demonstrated, individuals are generally averse to events with ambiguous probabilities (i.e., those for which the probability distribution is not known precisely). For example, given a choice between a gamble where I can get $100 with a 50 percent chance of success and one where I can get $100 but with unknown odds, I will most likely choose the first one. In short, ambiguity aversion is quite common for unknown probabilities.
Most of the research on ambiguity focuses on vague probabilities. However, in some situations, uncertainty can also exist with respect to outcomes (i.e., whether the outcome is precisely known or not). Table 12.1 presents some stylized examples of this taxonomy of precise/imprecise probabilities and outcomes. For instance, on a roulette table, the chances of winning or losing and the amount at stake are well specified; hence this example falls into the category of precise probabilities/precise outcomes. The outcome of an airplane crash is almost always certain, whereas the probability of having such an accident is not—a case of ambiguous probabilities/precise outcomes. We could estimate (more or less) our chances of getting a tax audit depending on our income bracket, but it is difficult to foresee any possible penalties—a case of precise probabilities /ambiguous outcomes. And, finally, if we do not have a clear prediction about the probability and the magnitude of damage from an earthquake, we are looking at a case of ambiguous probabilities/ambiguous outcomes.
TABLE 12.1 Examples for Different Sources of Ambiguity
Behavioral studies show that many people are averse to ambiguity for both probabilities and outcomes. Thus we don’t like to leave things to total chance (ambiguous probabilities). We also do not like the vagueness associated with possible consequences (ambiguous outcomes). For instance, we prefer to know the exact possibility of a financial crash or an earthquake and also whether a market crash or an earthquake will destroy none, half, or all of our endowments. Some studies have shown that we are even willing to pay a premium to decrease the uncertainty surrounding probabilities and consequences of adverse events. What is the implication of this aversion for future events? The following section tackles this question and shows that we are not all that uncomfortable with future uncertainty.
AMBIGUITY AND THE FUTURE
Up to now, I’ve said little about how timing affects our reaction to uncertainty. For many decision problems we do not know when the event will be realized; examples include natural disasters such as earthquakes and hurricanes, health problems as a result of prolonged exposure to radiation, and life expectancy itself. How do we behave when faced with such uncertainty? Economic