Why is uncertainty important?

Uncertainty and risk in decision making

Integrated river basin management requires making a large number of decisions by operational agencies. A decision maker has to make decisions based on the available information. In most cases this information is deficient, incomplete and characterised by uncertainties of different kinds. How should this affect the decision making process? With increased uncertainty the chance of taking wrong decisions increases. Can the decision maker accept this? What can he or she do to make decisions that anticipate that outcomes can differ from what was expected? How should knowledge about the various uncertainties that characterise the available information be used to make better decisions? Nearly all information is uncertain. The main subject in this guidance document is methodologies and tools to characterise uncertainty and to assess the various sorts and sources of uncertainty, and the propagation of uncertainty through models to management information.

As one option, the decision maker may decide to postpone making a decision. This will allow more effort (money) to be put into collecting additional data or increasing modelling efforts, so as to improve the quality of information and, thereby, reducing the probability of taking a wrong decision. Further work on data collection and modelling involves costs in itself, and delays in the decision may also involve additional costs as well. The typical question "Do we have enough understanding to responsibly make a decision responsibly?" is thus in a rational sense a question on whether the risks of wrong decisions caused by imperfectness of the information are acceptable, or whether it is advisable to improve the quality of the information by, for instance, further data collection/modelling. One formal approach to address this question is through “value of information” analysis (VOI). Essentially, VOI encourages the prioritisation of research such that the expected value of additional information is compared against the opportunity costs of uncertainty, including the costs of making the wrong decision. This is sometimes difficult to do quantitatively in the environmental field because of problems associated with putting a value on natural resources. Nevertheless, even in the absence of explicit quantification, it is useful to remember that continuous attempts to reduce uncertainty by collecting additional information does not always make economic sense. Additional information—in the form of more data or more refined algorithms—almost always incurs costs. These costs must be balanced against the expected benefits of such information.

A note should be made here that higher quality of information does not automatically imply information with less uncertainty. It can also be information providing a richer insight in the sorts and magnitudes of uncertainty, so that the decision maker better understands what possible outcomes and risks to anticipate. This allows for contingent planning where options are kept open and flexible and where emergency risk management options can be prepared and kept on the shelf (compare to a fire extinguisher) so that they can be implemented immediately if that turns out to be necessary.

Another option open to the decision maker is to become risk-avers. This may result in incorporating a large safety margin or a ‘tolerance’, so that increased resources are spent on measures, such as clean-up of a site or water body, where there is a large probability that it may not be required to protect the water resources.

A decision maker can adopt one of the strategies developed in the field of decision theory if the ranges of outcomes for different options to choose from are known, but not the probabilities of each outcome. For instance, maximin is the strategy that chooses the option that has the best (that is: the least severe) worst case scenario. It makes sense if we have little to win and a great deal to loose, but it tends to prevent us from taking advantage of opportunities. Closely related to maximin is the difference principle: one society is better off than another if the worst-off members of the former do better than the worst-off in the latter. Maximin allows the most disadvantaged members of society to be harmed if the overall society benefits; the difference principle would forego an overall benefit to the society if it harmed the most disadvantaged members. The difference principle has been criticised for that it does not weigh limiting disadvantages to a subset of people against a possible increased average utility of society.

On the other extreme, maximax is the strategy that chooses the option that has the best best-case scenario. It is a risky strategy, often preferred by people with a risk seeking attitude that want to take advantage of opportunities and it makes sense if one has a great deal to gain and little to loose. Maximin can be seen as excessively pessimistic and maximax as excessively optimistic. An approach that attempts to balance between good and bad outcomes is the principle of insufficient reason: when we lack objective evidence to specify probabilities of outcomes, we should treat all outcomes as equally probable. (Resnik, 2003)

A widely advocated strategy is the Precautionary Principle if both the bounds on the outcomes and their probabilities are unknown. The Precautionary Principle grants greater benefit of doubt to the environment and to public health than to the activities that may be held to threaten these things (Stirling, 2003). Because the Precautionary Principle applies to those cases where serious adverse effects and surprises can occur with an unknown probability, it is rational to follow a better safe than sorry strategy. Failing to take precautionary measures in a timely manner could result in devastating and irreversible consequences (Harremoës et al., 2001). Such consequences might have been avoided by proactive and anticipatory interventions whose costs are justifiable in comparison to the damages and losses that could occur.

The above illustrates that decision makers should act differently under different situations of uncertainty. However, they will only be able to do this on a rational basis when they know how uncertain the available information is and when they know how to incorporate this in their decision making. Uncertainty is a difficult concept, and there is a need to educate and assist the decision maker working in a situation where there is uncertainty. In this document we will describe methodologies and tools for uncertainty assessment that may be used in the decision making process, allowing decision makers to take rational decisions on how to act under a situation of uncertainty.