The object of this paper is to provide a method of cost-benefit analysis of earthquake prediction as a means of mitigation of earthquake effects. The research in earthquake prediction may or may not be successful and involves an initial cost. Earthquake prediction, if achieved, on the one hand provides society with information which allows it to take protective measures. On the other hand, each prediction involves the costs of those measures and the consequent disruption of economic life. The question is to assess the value of such information in a given state of the prediction technology.
The evaluation of a fault-monitoring program and its consequences for the public at the time of predictions is performed over a 50-year period. A rate of growth, a social rate of discount, and a rate of improvement over time of earthquake prediction techniques are assumed.
A model “TREE” is developed; it allows computation, for each year, of the expected value of the earthquake prediction information—expected costs minus expected benefits. The life component and the dollar component of the net result are kept separate throughout the evaluation.
The final result is an expected cost per life saved through the earthquake prediction program over a 50-year time period. This allows comparison with the results of earthquake engineering and building codes (see Paté, 1978). It also allows comparison with the results obtained in other public sectors involving risk mitigation—health and transportation, for example.
A numerical example has been worked out for the case of the San Francisco Bay Area; it gives a first approach to the results that can be expected from a prediction system with different assumptions on the success of research in that field.
This paper is based on the doctoral thesis at Stanford University of M-E. Paté, under the supervision of Professor H. C. Shah.