Public policies are, at once, the means for articulation of political philosophies and processes, the conduits for conversion of political and bureaucratic decisions into actions and the means by which the electorate can assess government performance. Public policy processes offer a means of achieving social and economic change and they are a primary justification for the existence of governmental systems. On these counts, identification of the elements of policy processes and the ways they interact with each other is essential to an understanding of the relationships between public policy decisions, systems of democratic government and their connections with wider society. This thesis goes behind the facade of public policy outcomes and analyses the processes involved in arriving at policy decisions. Linkages are traced between political theories, the processes of public policy decisions and final policy outcomes. This involves, first, an examination and critique of liberal-democratic theories. Second, there is detailed examination of pluralist democratic practice, which is the prevailing political paradigm of modern western liberal-democratic societies. The analysis finds substantial evidence of gross distortions in the process relative to normative theories. Plain causes are the institutionalisation of special interests to the exclusion of wider public interests and inadequate accountability of governments and bureaucracies for their actions. Policy processes in pluralist systems are examined and it is concluded that the social environment, institutional influences and factors which affect the behaviour of institutions are key elements explaining public policy decisions. The capacity for pluralism to significantly influence policy outcomes depends largely on the degree and nature of access to the public policy process at various points. In examining the role of government institutions in public policy processes, it is argued that a clear distinction between the elected legislature and the administrative bureaucracy is artificial and misleading. Further, there is evidence that public service bureaucrats can become captives of their particular client groups and, thus, less accessible to the full range of relevant interests. These problems are exacerbated by the two-party Westminster model of representative democracy which tends to concentrate power in cabinet government, resulting in a decline in the importance of parliament as a deliberative and scrutinising bodies. This dissertation develops the view that there are significant causal links between institutional philosophies and values and the dominant disciplines within institutions. It is also argued that growing professionalism in bureaucracies and a tendency for functional divisions of public policy to be in broad symmetry with the divisions of the professions, tends to intensify the influence of particular professional disciplines on related areas of public policy. The critique of liberal-democratic theories and the related discussion of factors affecting policy processes in a pluralist system are used to identify the essential elements of public policy processes. It is proposed that all policy processes contain the four elements of pluralism, access, accountability and planning which are interactively related. Differences in emphasis given to these elements in the policy process explains the nature of individual policy decisions. Thus, the normative policy process datum model provides both a static and dynamic framework for analysing policy decisions. In order to examine the theoretical arguments in an empirical context, the policy processes of the Australian Federal Government, in the area of industry assistance, are analysed. This policy arena contains all the 'raw material' of pluralist processes and is, therefore, a fertile area for analysis. Furthermore, operating within this policy arena is the Industries Assistance Commission [IAC], a bureaucratic institution which is quite unlike traditional administrative structures. The IAC has, prima-facie, all of the features of the policy process datum model; it operates in an open mode, it encourages a range of pluralistic inputs, it has a highly professional planning function and, because its policy advice is published, it encourages scrutiny and accountability of itself, other actors in the bureaucracy and the elected government. The IAC operates in a rational-comprehensive mode. The analysis concludes that the IAC was established in part to be a countervailing force to restore some balance in the industry policy arena. In this it has been partly successful - the distributive policy decisions of governments have come under much greater scrutiny than in the past and other areas of the bureaucracy have been forced to operate more frequently in a rational-comprehensive mode, rather than as advocates of sectional interests. The IAC has itself limited its range of objectives, however, and has tended to become a computational organisation, isolating its core economic [planning] technology from the interactive processes of the policy process model, i.e. pluralism, access and accountability. By protecting its essential philosophy in this way, the IAC runs the risk of becoming less influential in the overall policy process. Using the policy process model as a datum, and the empirical experience of the IAC and the policy arena in which it operates, several options for administrative reform are examined. A summary agenda for administrative change is proposed which revolves around ways of achieving balanced pluralistic inputs, a greater degree of access, better bureaucratic and government accountability and ways of exploiting but controlling technocratic planning expertise. Emphasis is placed on the need to achieve enriched interactive flows between each of these key elements. If these conditions can be met, it is proposed that a revised and improved administrative bureaucracy will emerge.