• Choose a category
• All
• Classic-Fiction

In a decision context, ambiguity can be present in three ways, as pictured in Figure 12.1 and as summarized below:

1. unknown timing: ambiguity about when the event will happen

2. unknown probabilities: ambiguity about if and how often the event will happen

3. unknown stakes: ambiguity about how much will be at stake if the event happens

In the early 1960s, Daniel Ellsberg demonstrated aversion to the second type of ambiguity (unknown probabilities) in a simple experiment with two urns containing red and black balls.1 Indeed, people usually avoid options with unreliable information about probabilities. Maybe this is why bankers find comfort in highly sophisticated mathematical models; using such models they can assure themselves that they know precisely the risk they are taking when investing in mortgage-based securities, credit derivatives, and other alternative risk transfer mechanisms.

In many situations, ambiguity aversion is governed not only by unknown probabilities but also by unknown outcomes. For instance, we cannot put an exact dollar amount on the potential damage from a major quake in a particular region. In the case of global warming, we have only a vague idea about the chain of reactions it can cause, such as extreme weather conditions, rising sea levels, and so on. And, as this book is being written, the size of the global loss from the economic crisis is yet to be determined. (Of course, ten years from now history books will specify this loss, but for now nobody can describe it with any degree of confidence.)

At this point, we should note the difference between ambiguity aversion and risk aversion. Risk aversion is a preference for certainty over risk, whereas ambiguity aversion is a dislike of gaps and inconsistencies in information regarding probabilities or outcomes. For instance, a person may be willing to take risks in domains that she feels confident about (playing poker, participating in extreme sports, investing in volatile markets) but may be averse to situations where these risks are not accurately defined or are debatable (using cell phones, eating genetically modified food). Brain studies also confirm this distinction. Ambiguity lights up regions in our brain that are associated with negative feelings, such as fear, disgust, and displeasure, whereas these regions are not activated when we are faced with a risky choice.

Aside from Ellsberg’s classic urn problem, most cases of ambiguity have a dynamic component: The uncertainty is not resolved immediately upon the making of a decision. In fact, in many cases we do not even know when (if ever) the uncertainty will be resolved. Consider the following hypothetical earthquake scenario. Let’s say that seismic experts suggest there’s a 10 percent chance a major earthquake will occur in a particular region in the next twenty years, causing a loss of approximately $50 billion. First of all, the likelihood and loss figures themselves (i.e., ambiguous probabilities and outcomes) may be disputable. Second, it is very unlikely that the ambiguity will resolve today or tomorrow. Third, it is not clear whether one’s own house will be affected by this earthquake. When making a decision about buying earthquake insurance or investing in mitigation measures, a homeowner will need to evaluate costs and benefits for the next twenty years and beyond, not just for the present time. We know that people are generally uncomfortable with missing information about a current event. But how will they react when this uncertainty is prolonged into the future? Will those people who expect to live in their houses for the next twenty years be more or less tolerant of ambiguity? Will those who put their money into long-term investments be more or less concerned about ambiguity? Interestingly, as I will discuss later on, many people become more tolerant of future uncertainty. This means that our homeowner will feel less threatened by an earthquake the longer she plans to live in the area. The same