¹ See Dunleavy, Patrick and Ward, Hugh, ‘Exogenous Voter Preferences and Parties with State Power: Some Internal Problems of Economic Theories of Party Competition’, British Journal of Political Science, XI (1981), 350–80.Google Scholar

² Dunleavy, P., ‘Bureaucrats, Budgets and the Growth of the State: Part I, Existing Public Choice Approaches’, unpublished paper, available from the author at LSE.Google Scholar

³ Downs, A., Inside Bureaucracy (New York: Little, Brown. 1967)CrossRefGoogle Scholar; Niskanen, W., Bureaucracy and Representative Government (New York: Aldine-Atherton, 1973).Google Scholar Other less important work in the genre includes: von Mises, L., Bureaucracy (New Haven: Yale University Press, 1944)Google Scholar; Tullock, G., The Politics of Bureaucracy (Washington, D.C.: Public Affairs Press, 1965)Google Scholar; Buchanan, J. et al. , The Economics of Politics (London: Institute for Economic Affairs, 1978)Google Scholar; Breton, A., The Economic Theory of Representative Government (Chicago: Aldine, 1974), Chap. 9CrossRefGoogle Scholar; Migue, J. and Berlanger, G., ‘Towards a General Theory of Managerial Discretion’, Public Choice, XVII (1974), 27–43CrossRefGoogle Scholar; Borcherding, T. E., ed., Budgets and Bureaucrats (Durham, North Carolina: Duke University Press, 1977).Google Scholar

⁴ Niskanen, , Bureaucracy and Representative Government, p. 22.Google Scholar In his more popular exposition, Bureaucracy: Servant or Master? (London: Institute for Economic Affairs, 1973)Google Scholar, Niskanen admits (p. 11) that he sometimes uses the term in this specialized sense, and sometimes just to mean ‘civil servant’. Readers must presumably guess which is which, for the distinction is buried thereafter.

⁵ Lee, M., ‘Whitehall and Retrenchment’, in Hood, C. and Wright, M., eds, Big Government in Hard Times (London: Martin Robertson, 1981). p. 41Google Scholar, shows that in 1980 there were over 800 people in the top three ranks, spread across forty-seven departments. For more detailed figures, see Dunsire, A. and Hood, C., Bureaumetrics (Farnborough, Hants: Gower, 1980).Google Scholar

⁶ See, for example. Crouch, Colin, Trade Unions: The Logic of Collective Action (London: Fontana, 1982), pp. 67–74.Google Scholar Crouch's analysis of trade unions resembles at some points the approach used here.

⁷ For interesting comments on the interchangeability of coalition theory and collective goods models, see Abrams, R., Foundations of Political Analysis (New York: Columbia University Press. 1980). pp. 329–46.Google Scholar

⁸ Downs, , Inside Bureaucracy, p. 133.Google Scholar Downs continues: ‘If bureaus were really monolithic, control over nearly all their activities would be concentrated in the hands of their topmost officials. However, these officials must always delegate some of their powers to subordinates’.

⁹ This equation in the same general terms was, of course, specified by Downs, A., An Economic Theory of Democracy (New York: Harper and Row, 1957)Google Scholar as the calculus facing citizens deciding whether to vote in a democracy, and applied to the analysis of why people join interest groups by Olson, Mancur Jr., in The Logic of Collective Action (Cambridge, Mass: Harvard University Press. 1965).Google Scholar In each case the format is that the net utility gained from action, discounted by the probability of the action influencing the eventual outcome, must exceed the (non-discounted) costs incurred in taking the action.

¹⁰ See Bennet, J. and Johnston, M., The Political Economy of Federal Government Growth (College Station, Texas: Center for Education and Research in Free Enterprise, Texas A & M University, 1980), pp. 38–41.Google Scholar The problem with widely varying estimates here is in distinguishing between contracted-in staffs employed by the agency directly, and the much more massive phenomenon of contracted-out services being delivered by employees of firms or other agencies. We are concerned here with contracting in. Of course, contracted-out work often includes substantial staffing implications as part of larger contracts for products or service delivery. But since contracted-out staffs remain directly employed by other bodies, and since they often cannot even be sensibly construed as falling within the public sector, this kind of indirect employment is better analysed using the ‘contract agency’ classification of bureau situations in the typology developed below.

¹¹ Goodin, R., ‘Rational Bureaucrats and Rational Politicians in Washington and Whitehall’, Public Administration, LX (1982), 23–41.CrossRefGoogle Scholar

¹² Top bureaucrats are also likely to be much better informed about changes in the external environment than lower ranking officials, so that we should expect to see their reactions adapting much more quickly and accurately to external changes. In contrast, lower officials' behaviour may well perpetuate anachronistic attitudes into a new period, remaining conservative in their behaviour in environments favourable for organizational growth, but also carrying over a previously defined growth orientation into an era of cutback management.

¹³ See Jackson, P., The Political Economy of Bureaucracy (Oxford: Phillip Allen, 1982), Chap. 5 passim for a good review.Google Scholar

¹⁴ I would like to thank Jeremy Waldron of Edinburgh University for helping to clarify the issues involved here.

¹⁵ Both Downs and Niskanen shift between these meanings without apparently perceiving their different implications. Niskanen in particular explains agency behaviour quite largely in terms where ‘budget’ equals ‘bureau budget’. But in his vaguer accounts of how budget maximization creates state growth, ‘budget’ normally equals ‘programme budget’.

¹⁶ For the sake of argument I have assumed in constructing Table 1 that all these net utility aspects show gains with budget increases. In practice, it seems unlikely that this will be the case. For example, budgetary expansion may keep the peace inside agencies but simultaneously increase bureau inertia. Similarly, increasing bureau budgets may allow more agency patronage over contractors; but enhanced pressure for outputs may also increase an agency's dependence on external firms. Dunleavy, P., The Politics of Mass Housing in Britain. 1945–75 (Oxford: Clarendon Press. 1981)Google Scholar documents exactly this kind of power-dependency switchback between public housing agencies and construction corporations in Britain's post-war housing policy.

¹⁷ Advocacy costs of promoting budget maximization (on the right-hand side of bureaucrats' choice equation) need to be clearly distinguished from potential disutilities of budgetary expansion (which are incorporated into the net utility calculation on the left-hand side of the equation).

¹⁸ Taxing agencies might be regarded as a sub-type of regulatory agencies, but they are distinctive in raising revenues as a result of their activities, making it difficult to apply simple demand/cost analysis to them. Mixed agency types with a commercial role similarly involve complex modelling, although Niskanen drops a few hints about them.

¹⁹ In British local government, for example, administration cost ratios of 15–25 per cent of the programme budget are regularly applied where council staffs do agency work for other bodies (such as police or water authorities).

²⁰ For a more detailed discussion see Dunsire and Hood, Bureaumeirics. Dunsire and Hood's painstaking empirical work was clearly restricted in value by their decision to focus on an institutional category (‘the department’), rather than to develop statistics organized around more theoretically constituted variables.

²¹ On which see Kellner, Peter and Crowther-Hunt, , The Civil Servants (London: Macdonald, 1980), pp. 174–238.Google Scholar

²² A factor alleged to underlie the (partial) failure of giant departments in Whitehall. See Pollitt, C., Manipulating the Machine (London: Allen & Unwin, 1983).Google Scholar

²³ Dunleavy, , The Politics of Mass Housing in BritainGoogle Scholar shows, for example, that in 1966–67. at the height of the public housing boom, local authority architects designed just half of the dwellings involved, private architects 30 per cent, and contractors' architects 20 per cent. By 1973, when the public housing programme was half the size, local authority architects' share increased to 75 per cent. In addition, the work tasks which public architects did on housing changed dramatically. In one authority (Birmingham) they were doing solely contract drafting and supervision (plus landscaping) in the late 1960s, whereas by the mid-1970s the council architects had regained control of designing most of their department's housing. This kind of effect will exist wherever public agency staffing does not expand fully to accommodate workloads in ‘boom’ periods, so that the character of their work tasks tends to stretch to accommodate the variation. This phenomenon implies that staff utilities at middle or lower levels may rise quite sharply in periods of limited budgetary reductions when workloads are reinternalized.

²⁴ Bennett, and Johnson, , The Political Economy of Federal Government Growth, pp. 30–7Google Scholar demonstrate that the proportion of American federal employees in the bottom four of eighteen General Schedule grades declined from 40 per cent in 1959 to 21 per cent in 1978, while the proportion in the top five ranks grew from 6·5 per cent to 14 per cent in the same period. They compare this with the relatively static total of federal employees (up from 2·24 million in 1959 to 2·48 million twenty years later, and down as a proportion of the labour force from 3·2 to 2·4 per cent over the period), concluding that: ‘a massive shift in General Schedule grades has occurred towards policy-making and program administration levels’ (p. 38).

²⁵ See Gretton, J. and Harrison, A., How Much are Public Servants Worth? (Oxford: Blackwell, 1982).Google Scholar

²⁶ Cawson, A., Corporatism and Social Welfare (London: Heinemann, 1982)Google Scholar applies this idea to the social policy field in the United Kingdom.

²⁷ Hant, K. and Scharpf, F. W., eds, Interorganizational Policy-making (London: Sage, 1978).Google Scholar

²⁸ For example, in the United States federal government overall establishment numbers were frozen from 1977 to 1980; in the United Kingdom civil service manpower was first frozen in 1979, and then programmed to decrease by 7 per cent over three years (thanks primarily to the shedding of low-level line functions and to limited privatization). See Fry, G. K., ‘The Development of the Thatcher Government's “Grand Strategy” for the Civil Service’, Public Administration, LXII (1984), 322–36.CrossRefGoogle Scholar

²⁹ Kellner, and Crowther-Hunt, , The Civil Servants, Chap. 12.Google Scholar

³⁰ The criticism made here against the budget-maximizing thesis also applies with particular force to the aggregate-level explanation put forward by Tullock, G., ‘Dynamic Hypotheses on Bureaucracy’. Public Choice, XVII (1974). 128–32.Google Scholar

³¹ For discussions of this ‘Balkanization of the policy process’ see Self, P., Administrative Theories and Politics (London: Allen & Unwin. 1977). p. 279Google Scholar; Dunleavy, P., ‘The Limits to Local Government’, in Boddy, M. and Fudge, C., eds. Local Socialism? (London: Macmillan, 1984)Google Scholar, and ‘Quasi-Governmental Sector Professionalism’, in Barker, A., ed., Quangos in Britain (London: Macmillan, 1982).CrossRefGoogle Scholar

³² Hood, C., ‘Keeping the Centre Small: Explanations of Agency Type’, Political Studies, XXVI (1978), 30–46CrossRefGoogle Scholar; Dunleavy, P. and Rhodes, R., ‘Beyond Whitehall’, in Drucker, H., Dunleavy, P., Gamble, A. and Peele, G., eds. Developments in British Politics (London: Macmillan, 1983), pp. 104–13Google Scholar: Bennet, and Johnson, , The Political Economy of Federal Government Growth, passim.Google Scholar

³³ Jackson, , The Political Economy of Bureaucracy, p. 87Google Scholar, notes: ‘In the neo-classical system agents are treated as if they are mindless automata who respond in a fully programmed fashion to external stimuli such as price and quantity signals’. For comparison, see Piaget, J., Structuralism (London: Routledge and Kegan Paul, 1962)Google Scholar, which remains the best general introduction to the approach.

³⁴ Boudon, R., ‘Undesired Consequences and Types of Systems of Interdependence’, in Blau, P. and Merton, R., eds, Continuities in Structuralist Inquiry (London: Sage, 1981).Google Scholar See also Boudon, R., Effets perves el ordre social (Paris: Presses Universitaires de France, 1977).Google Scholar

³⁵ Warriner, C., ‘Levels in the Study of Social Structure’Google Scholar, defines inter-positional and inter-organizational levels of analysis, while Heydebrand, Wolfe V., ‘Marxist Structuralism’Google Scholar, discusses radical approaches, both in Blau, and Merton, , eds, Continuities in Structuralist Inquiry.Google Scholar

³⁶ Dunleavy, P., ‘Is There a Radical Approach to Public Administration?’. Public Administration, LX (1982). 215–33.CrossRefGoogle Scholar