Stochastic optimization and economic models (Q1118510)

Within the framework of economic policy intertemporal analysis seeks to understand basic characteristics of evolutionary processes in society, e.g. the time structure of change, needs, preferences, attitudes, decisions, and so on. The ultimate may be the design of a sequence of planning actions so as to replace possible but unwanted future events by more probable and desirable ones. However, to do so one needs methods of improving our knowledge and information on the states of reality, one needs to learn to adapt to predictible and nonanticipative events unfolding over time. Hence, one needs to increase our ability to cope with processes of change in the perspective of uncertainty, the latter to be understood as the lack of knowledge and information on the sources and states of change. Stochastic control theory in particular and stochastic optimization in general focuses on the design of models susceptible to precise mathematical description and seeks to understand fundamental principles for the guidance of processes taking place in the systems that these models represent. The book under review presents a rich manifold of economic models resulting from the application of stochastic and control techniques to a variety of problems arising in financial economics, in the theory of allocation and extraction of resources, cartelized markets and other areas. The unifying aspect is the emphasis on the economic implications of optimal decision rules. Chapters I and II give a quick introduction into the theory and art of applying stochastic processes to economics based on a purposely selected sample of examples. Chapter III centers on optimal control models aimed at illustrating the lines of activity in some major areas of current economic applied research, i.e. (1) models of differential market games, (2) stabilization policy in cartelized markets, (3) adaptive control in economic models and (4) model of renewable resources under uncertainty. The second half of the book is dedicated to various aspects of the highly interesting area of portfolio selection that have received in the last decades a great deal of attention. In portfolio theory, as in any field dealing with problems involving future uncertainty and a changing environment, the main efforts are concentrated on future modelling and risk management. It is in this sense that Chapter IV covers relevant issues on the estimation of risk, updating of an optimal allocation decision as additional information becomes available and on the associated robustness, decision rules for portfolio revision, diversification, mutual funds and the like. The conclusion that more theoretical and empirical investigations are needed on these and related questions closes the chapter. Chapters V and VI examine various theoretical and empirical aspects of the singularity and orthogonality associated with the problem of computing an efficient frontier in the mean-variance space and to evaluate diversification. The book ends with a chapter on efficiency measurement in non-market systems characterized by the fact that some of their basic variables are not directly observable, i.e. they are in a sense latent variables. After presenting an overview of the model of data envelopment analysis, for short DEA, and contrasting it with two other methods of efficiency measurement, the author considers several extensions of the DEA method for the stochastic and dynamic cases. Sengupta's book is certainly a welcome addition to the literature on stochastic optimization and should be read by theorists and applied oriented researchers alike, since it contains plenty of research suggestions that urgently ask for answers. The book is also suitable for less advanced readers, for the author presents a fairly clear and streamlined overview on the theory of random processes needed in stochastic economics and then moves skillfully to the forefront of this exiting and rapidly evolving field of research. However, we should say that a better prepared index would have considerably improved the usefulness of the book. There are some misprints although none of them of importance.