Consequences

That part of risk we have called 'consequences' does not have the philosophical problems associated with probability. There are, however, many practical problems. The consequences of an event must be measured in terms of a value system. The most straightforward value system is the binary measure of simple failure, where all that needs to be known is the probability of the event occuring. Either a building collapses, or it does not; or a train crashes, or a ferry is lost.

This sort of risk problem is relatively straightforward. Even so there is often the difficulty that the theoretical model used for analysing the problem does not show such clear-cut behaviour. For instance, there is no obvious way in which a computer model of a ductile building will 'collapse' in an earthquake, and so an arbitrary definition of collapse has to be made to relate the model behaviour to the required reality.

More often risk will be concerned with more fundamental results such as loss of life, monetary loss or environmental damage. The consequences must then be expressed in terms of a specific value measure. A difficulty arises if more than one has to be taken into account in the same analysis. This can be done by arbitrarily assigning monetary value to life loss or environmental damage. Sometimes it is necessary to do so. However, in many cases it turns out that the different measures of risk will be used in separate contexts.

This simplifies the problem in that the value measures can be kept separate. Even so, some things are not easy to quantifyinjury, for instance, as opposed to loss of life. Either a life is lost or it is not, but an injury could be of many degrees. Similarly, though some environmental matters can be quantified, such as an increase of carbon dioxide emission or noise level, others are not so easily handled-a loss of aesthetic value, for instance, or the long-term effects of destroying a forest or polluting a watercourse.

There is no simple answer to such problems, but methodologically there is less difficulty if the risk analysis is carried out in two parts, the first dealing with the probability of a disaster happening and the second analysing the consequences should it occur. There is also a problem in comparing long-term and immediate consequences.

The results could be quite different depending on the choice of discounting factor. In any case there is the question of whose view should be taken into account. Many people put little weight on long-term threats and give greatest importance to immediate risks. People are notably reluctant to give up smoking : the consequences are well documented and grave, but they are not immediate. However, we are now straying into the third component of risk, its context.