With relatively few exceptions, like the economic theory of bureaucracy (Downs, 1957; Niskanen, 1972) and its derivative notion of bureaucratic failure, economists have ignored public sector organisations. If the core assumptions of economic analysis consist of rational maximising behaviour, stable preferences and comparable equilibria, and if neoclassical economics adds peripheral concepts like perfect information and costless market exchange (Eggertsson, 1990), then this reticence about examining hierarchical relationships characteristic of public bureaucracy would appear warranted. However, with the development of new institutional economics (NIE) and its emphasis on introducing institutional realism into economic analysis, especially agency theory, property rights economics and transaction costs economics, economists now possess powerful tools for the analysis of organisational behaviour and design. In common with several other contemporary commentators, including Bailey (1995), Frant (1991) and Hood (1998), we take up the question of NIE and its explanatory abilities in a public sector context.
NEW INSTITUTIONAL ECONOMICS
In common with the major earlier institutionalist tradition associated with Thorstein Veblen, Wesley Mitchell, John R. Commons and Clarence Ayres, NIE represents a loose collection of ideas aimed at bringing institutional characteristics back to the core of economic analysis (Rutherford, 1996). However, unlike the older tradition, NIE scholars have few problems with a priori deductive theorising. For example, Furubotn and Richter observe that:
The change in analytic approach adopted by the new institutionalists has not resulted from any deliberate attempt to set up a new and distinct type of doctrine in conflict with conventional theory. Rather, the tendency to introduce greater institutional detail into economic models has come about gradually over time because of the recognition that standard neoclassical analysis is overly abstract and incapable of dealing effectively with many current problems of interest to theorists and policymakers (Furubotn and Richter, 1992:1).
Given its somewhat disparate nature, NIE is difficult to define with any degree of precision. Frant (1991:112/113) has identified three "precursors" of NIE. First is Coase's (1937) focus on the importance of transactions and the related problems of whether to employ markets or hierarchies to handle transactions, which emphasised the costs attached to using the price mechanism. Second is the literature on the economics of property fights deriving from Coase's (1960) famous paper on how the assignation of property rights influences outcomes in the presence of externalities, which led to the so-called "Coase theorem". And third is Alchian and Demsetz' (1972) seminal attempt to apply the property rights paradigm to organisations engaged in productive activity, with the problems inherent in "team production", like "shirking" and "monitoring".
Disagreement exists on the major dimensions of NIE. For instance, Rutherford (1996:2/3) adopts a fairly broad view of the main elements of NIE by including the economics of property fights (Alchian and Demsetz, 1973), law and economics (Posner, 1977), rentseeking and distributional coalitions (Olson, 1982), agency theory (Jensen and Meckling, 1976), transaction costs economics (Williamson, 1979), game theory in institutional situations (Shubik, 1975), and the new economic history of Douglas North (1981). Matthews (1986) has developed a fourfold taxonomy of NIE in which institutions are viewed as property fights, as kinds of contracts, as conventions, and as governance structures. By contrast, Boston et al (1996) constrain their policy-orientated view of NIE to only two strands, namely agency theory and transaction cost economics. We follow a somewhat broader taxonomy of ingredients which …