From the agency perspective...

Organisational economics has two main streams. One, developed by Oliver Williamson, focuses on the governance of transactions. This addresses the circumstances under which it is more efficient for firms to extend their organisational boundaries, develop inter-firm collaboration or transact through the market. The other, exemplified by the dozen articles in this book by Michael Jensen and others, focuses on the principal-agent problem. This is concerned with the design of the internal organisation of the firm, how decisions should be decentralised under particular circumstances, and what systems are needed to measure performance, and reward or punish individuals.

For both approaches, the single key issue is specificity. Williamson argues that the more assets are specific to a trans-action, the greater the extent to which transactions are more efficiently internalised within the firm. Jensen argues that the more knowledge is specific, the greater should be the decentralisation of decision making and, therefore, the greater the principal-agent problem. This in turn means more elaborate systems to measure performance and generate incentives to overcome the agency issue.

This is the first time we have had a virtually complete exposition of Jensen's approach to what has become known as "positivist agency theory" between one set of hard covers. It should thus be extremely welcome to every scholar who wishes to incorporate an agency perspective in his or her work.

It has seven key elements. There is a model of the human agent as "resourceful, evaluative and maximizing" (the Remm model), claimed to be more complex than the usual economists' utility maximiser, but still explicitly rational and decidedly individualistic.

There is the distinction between specific and general information, the former of which is costly to move between people, implying that decisions have to be decentralised to the point where the relevant specific knowledge resides. But decentralisation generates the agency problem - people who act on behalf of others are likely to follow their own, rather than their principal's interests, and so create agency costs and a control problem.

Decisions taken in the market are controlled through the institution of alienable decision rights, but these are not available within organisations, so substitute mechanisms must be devised. These mechanisms have to serve three functions: allocate decision rights among agents in the firm, measure and evaluate performance in the firm, and reward and punish individuals for their performance. Allocating decision rights is at the heart of organisational design.

These seven elements allow Jensen and colleagues to theorise and advocate a range of organisational solutions for a variety of particular circumstances. The implications of this approach for many contemporary questions are developed in several articles in the book.

Though the strengths of this strongly rationalistic approach are obvious, it should be set alongside two other considerations. First, management is also a social and political process. Social values are incorporated in Jensen and Meckling's Remm approach to the extent that they allow altruism, but these are treated as exogenous. There is no suggestion that managers might try to reduce the agency problem through social processes, or that there may be other systematic institutional effects.

A more novel consideration is the challenge to his (Jensen's) assumption "that organizations are equilibrium systems". The more recent view of organisations as complex adaptive systems on the edge of chaos and far from equilibrium belongs to an entirely different paradigm from Jensen's neo-classical view of the world. But not all organisations are at the edge of chaos. There is, fortunately, quite a lot of rational behaviour and Jensen's book is a powerful tool in understanding the implications of this.

Arthur Francis is director, management centre, University of Bradford.