The Evaluation of Risk in Business Investment

When a major capital investment decision is to be taken, much ofthe 'information' useful to the decision-maker will be speculative. The future is uncertain, and although experi? ence is valuable, no two major investment decisions are identical. There is the feeling that basic statistical techniques should be of assistance, yet the lack of data denies this facility unless we resort to subjective probabilities. The problem then consists of persuading the relevant experts to express their uncertainty about components of the problem in a quantitative form, converting these forms to distributions of the decision criteria and carrying out sensitivity analyses. This is the gist of Hull's very welcome book. It is a summary of the state of the art, largely dating from the publication of Hertz's case study1 in 1964. Topics covered include performance measures (largely net present value and internal rate of return), the assess? ment of subjective probability distributions, dependence between component variables, decision trees and the theory of utility. It is written in a direct and practical style and concludes with chapters covering a case study and implementation considerations. We could not expect Hull to provide answers to all the problems, and indeed he does not. An issue on which he left at least one reader rather confused was that of the method of obtaining a subjective probability distribution, where he appears to condemn the vari? able interval methods involving the identification of fractiles towards the tails of the distribution, yet uses such methods in subsequent analysis. Clearly much more will be said and written on this aspect of risk analysis; witness the recent articles by Mancini, Meisner and Singer2 and Moore.3 Nevertheless, this is a good book, useful to the practitioner and stimulating to the academic. The practitioner might with profit pass it on to his accountant colleagues, or indeed to general managers. The level of statistical knowledge required is basic, although the reader will need a familiarity with the subject's principles. It is relatively free from printing errors, but an index would have been useful. G. Gregory