The history of writing in this area is characterised by a continuous search for meaningful taxonomies of government organisations based on a variety of defining characteristics. Most have looked at structures; some have attempted to look at functions. There have been commentators however who have doubted the very possibility of developing a classification system for British governmental organisations. Wilding, in reference to 'fringe' organisations, is of the opinion that 'it is not possible to make true statements about them all; the impossibility lies in their bewildering variety' and eschews taking on the 'mantle of a latter-day Linnaeus' (Wilding 1982: 34). Hogwood has recently restated this point: 'any attempt to define quangos by listing distinguishing characteristics will break down, since some of these characteristics will not apply to some quangos and some will be shared by other types of bodies. We can talk of relative frequencies, but it is difficult to classify bodies unambiguously by their characteristics' (Hogwood 1995: 209). What is more important perhaps is 'the extent to which a range of bodies exhibit varying combinations of characteristics with which we may be concerned, and what the implications of these are for policy delivery and accountability' (Hogwood 1995: 210).

A useful starting point when trying to bring order to large number of government organizations is the distinction made by Chester, between 'ministerial departments', 'local authorities' and the 'rest' (Chester 1953). Here ministerial departments are defined as 'a Minister of the Crown to whom powers have been given either explicitly by name of his office or in the name of a body which by convention or declaration is clearly understood to mean that Minister'. A local authority is defined as 'a council with its powers and duties confined to a local area and elected by the electors of that area'. The third category in his scheme is covered by 'any governmental administrative body which has its own statutory powers and responsibilities and is neither a Minister nor a local authority is part of the 'rest'. It does not matter if this includes a number of bodies which are in practice completely or very nearly subordinate branches of a ministerial department'.

• The Civil Service Yearbook
• Civil Service Statistics
• Annual list of Parliamentary Commissioner for Administration under 1947 Crown Proceedings Act
• List produced by Property Services Agency of 'Allied Services Departments' and 'Repayment Departments'
• Index of Hansard
• Treasury Memorandum on the Supply Estimates
• Audited accounts by Comptroller and Auditor General (about 50% counted as departments)
• HM Ministers and Senior Staff in Public Departments
• Departments under direct ministerial control listed in the List of Ministerial Responsibilities published since 1974

• Ministerial and non-Ministerial Whitehall Departments
• Agencies directly controlled by Ministers, staffed by civil servants, yet not counted as Ministries
• Agencies staffed by civil servants but not directly controlled by a Minister
• Agencies directly controlled by a Minister but not staffed by civil servants
• Agencies neither directly controlled not staff by civil servants.

His scheme excludes all quasi-non governmental organizations as he defines them, all agencies being 'fully public sector bodies, constituted by legislation or orders in council, funded by the Exchequer, and subject to some measure of direct parliamentary and ministerial control'. Categories A, B and D make up Dunleavy's central state, that is, those agencies and departments under some degree of direct ministerial control.

The concept of 'families' has recently been introduced into the vocabulary of Public Administration as a more useful unit of analysis (Hogwood 1995: 513). 'Families' have their focus on main departments headed by ministers and include:

• non-ministerial departments reporting to ministers, but for which ministers normally did not and do not have responsibility for detailed operational matters;
• associated executive non-departmental public bodies;
• other statutory bodies carrying out policy delivery;
• other bodies carrying out statutory or other functions not recognised by government as public bodies, but functionally acting as though they were.

• simple core-agency families;
• families with cores but also other organizations;
• families with few staff in agencies in the main department;
• families with residues larger than core functions and with complex structures;
• families with large organizations outside the civil service;
• federal families.

1. commercial operations 'designed to run an industry or public utility according to economic and commercial principles...'
2. social service corporations 'designed to carry out a particular social service on behalf of the Government'
3. supervisory public corporations having 'essentially administrative and supervisory functions'. They do not engage in commercial transactions either to fulfil their main objective or incidentally to the performance of a social service.

These types were later given the titles of 'managerial-economic', 'managerial-social', and 'regulatory-social' bodies (Street and Griffith 1952: 271-275). Greaves put forward an earlier classification based on the reasons why bodies were created. As such there were four types of bodies: 'regional', e.g. the many Port Authorities; 'quasi-judicial'. e.g. the Civil Service Commission, the Import Duties Advisory Committee, and the Electricity Commission; 'trusteeship' bodies (for the management of property on behalf of others), e.g. the Charity Commissioners; and 'administrative or managerial' (of a national service), e.g. the National Coal Board (Greaves 1947: 104ff). A later attempt by Hague, Mackenzie and Barker was made to distinguish between 'governmental', 'quasi-governmental', 'quasi-non-governmental', and 'non-governmental' organisations (Hague, Mackenzie and Barker 1975). The result and practical usefulness of this typology was limited. Significant problems existed in delimiting their categories. Anthony Barker, one of those who claim to have originated the term 'quango', later called it 'overused and uselessly vague'. (A glance at the OED will show that the earliest uses found are American; and that it is quite unclear whether the acronym stands for quasi non-government(al) organization or for quasi-autonomous national government organisation). Many 'fringe' bodies are far from being 'non-governmental'.

Official attempts to classify bodies in this category began with Gordon Bowen's survey commissioned by the Civil Service Department in 1978. The survey was based upon a questionnaire circulated to departments. Departments were left to decide what constituted a fringe body. As such this self-reporting was criticised for being an haphazard approach to the subject (Chester 1979). Furthermore, Bowen's analysis excluded bodies such as nationalised industries, tribunals, judicial and quasi-judicial bodies, and also those bodies responsible for health services which shared many of the characteristics as other fringe bodies studies. The final analysis revealed few defining characteristics shared between all bodies. Bowen drew out a range of characteristics that were found in the bodies he surveyed:

• set up by Act of Parliament
• probably non-Crown
• financed by grants-in-aid
• chairman appointed by a Minister
• staff are non-civil servants, recruited and employed by the board or council of a fringe body
• annual accounts submitted to the sponsoring Minister and laid before Parliament
• annual report published

Out of 252 bodies surveyed, only thirty-three displayed all seven characteristics highlighted by Bowen.

The later attempt by Sir Leo Pliatzky (1980) at classifying 'fringe' bodies was as equally arbitrary. Pliatzky suggested the term 'non-departmental public body'. He distinguished between 'executive bodies', 'advisory bodies', 'tribunals' and 'other bodies'. Pliatzky excluded from his study nationalised industries, some other public corporations and NHS bodies. He recognised the difficulties associated with producing a robust typology of fringe bodies due to there being 'no legal definitions to determine what should go in these lists or into some other category'. He identified 489 executive bodies, 1,561 advisory bodies and 67 tribunal systems reflecting the arbitrary nature of systems of classification.

Official classifications of 'fringe' bodies differ at the time of writing. Public Bodies 1995 follows Pliatzky's distinction between executive bodies, advisory bodies, tribunals and other bodies while including details of nationalised industries and other commercial organisations, certain public corporations, and NHS bodies. The annual publication provides a working definition of a non-departmental public body (the preferred term, which we use): 'an NDPB is a body which has a role in the processes of national government, but is not a government department or part of one, and which accordingly operates to a greater or lesser extent at arm's length from Ministers'. Yet at the same time, the Government's own advice to Departments on NDPBs separates out Royal Commissions under a separate heading (Cabinet Office undated). This document reflects the difficulty in establishing any agreed criteria upon which to classify individual bodies. In describing the creation of NDPBs the Guide hesitates in making anything other than generalisations:

• Advisory bodies are normally set up by administrative action.
• Royal Commissions are set up under a Royal Warrant under the auspices of the relevant Secretary of State.
• Tribunals are set up by statute, usually in the context of a wider legal framework establishing citizens' rights and obligations.
• Executive, etc bodies are usually legally incorporated, by one of:
• legislation
• Royal Charter
• registration under the Companies Act

• staff numbers;
• status of chief executive and nature of appointment;
• organizational origins: some agencies for instance were departments in their own rights prior to gaining agency status, some were separate units within departments, some were made up of units from two separate departments, some already had trading fund status, and some had previously been non-departmental bodies;
• funding; gross-cost, net-cost or trading fund regimes
• staffing: vary according to whether they employ civil servants and non-industrial and industrial civil servants;
• existence of boards in agency structures;
• monopoly status;
• function and task: single or multiple.

• mainstream agencies: which are fundamental to the main policy orientation of the parent department e.g. the Employment Service;
• regulatory and statutory agencies: execute statutory functions derived from the main aims of the parent department e.g. Vehicle Inspectorate;
• specialist service agencies: providing services to departments and other executive agencies e.g. Information Technology Services Agency;
• peripheral agencies: not linked to the main aims of a department but report to a minister e.g. HMSO.

• financial regime: whether the agencies were self-financing or not and
• market share: whether the agencies were monopoly providers.

But again, there is not attempt to produce a taxonomy.

In research on bureau-shaping models of bureaucratic behaviour, as well as distinguishing between different 'types' of budgets, Dunleavy identified 5 main types of bureaus and three additional categories, related to particular configurations and ratios of a bureau's core, bureau, programme, and super-programme budgets respectively (Dunleavy 1985, 1989). In this sense, there is a relationship between structure (the bureau's budget) and function. Dunleavy's scheme, is however, more akin to a typology, rather than mere taxonomy, in that it carries with it an assumption that there are certain relationships between concepts, constructs and variables. The main types of agency identified are:

• delivery
• regulatory
• transfer
• contracts
• control
• taxing
• trading
• servicing

1. Finance
2. National Defence and External Affairs
3. National Defence and External Affairs
4. Research and Information
5. Production (including Agriculture, Forestry, and Fisheries), Transport and Commerce
6. Employment
7. Supplies
8. Education
9. Health
10. Justice

• Purpose: such as furnishing water, controlling crime, or conducting education;
• Process: such as engineering, medicine, carpentry, stenography, statistics, accounting;
• Persons dealt with or served: such as immigrants, veterans, Indians, forests, mines, parks, orphans, farmers, automobiles, or the poor;
• Place where service is rendered: such as Hawaii, Boston, Washington, the Dust Bowl, Alabama, or Central High School.

Yet Hammond has demonstrated that Gulick was aware of the limitations of his four-fold classification of methods of departmentatlisation, recalling Gulick's comment reflecting the 'impossibility of clearly dividing all the work of any government into a few such major purposes which do not overlap extensively' (Hammond 1990: 163).

Hogwood has taken Gulick's classification and analysed contemporary government organisation against these principles (Hogwood 1992: 65-167). There is currently no government ministry or department organised according to the principle of process employed in delivering policy. Client group organization was used for example, in the creation of the Ministry of Pensions and some others proposed (e.g. Ministry for Women). The territory principle is used in the cases of the Scottish, Welsh and Northern Ireland Offices. However, even within the Scottish Office, functions are allocated according to other principles in the departmental subdivisions. The allocation of activities according to broad purpose or function served is perhaps more common. Nevertheless, as Hogwood points out ( 1992: 166):

Self's view is that despite these criticisms, the process principle is perhaps the most easily operationalised: 'If there has to be one dominant principle of organisation, it is not hard to show by elimination that major function or purpose should be one. It is most intelligible, and conducive to goal effectiveness, to base organisation upon output rather than process, ends rather than means.' (Self 1972: 56-57). Organizations may carry out particular functions, including statutory ones, but the goals of these functions may be numerous. In reality, departmental arrangements and allocation of functions may combine several elements of the Gulick classification. The British Ministry of Technology that existed between 1964 and 1970 could be argued to have employed three of the methods of departmentalisation - process, client group and purpose.

The subject of the allocation of government functions to organizational structures was revisited by the Committee on The Reorganisation of Central Government under the Heath Government in 1970. Geoffrey Rippon had called for a 'new Haldane Committee' while in Opposition. The Report supported the functional principle for the organisation of government departments (Prime Minister/Minister for the Civil Service 1970):

the object has been to ensure that the broad framework of the central machinery in terms of Ministerial and departmental functions complies with the Government's strategic policy objectives. In practical terms, this means the application of the functional principle as the basis for the allocation of responsibilities: government departments should be organised by reference to the task to be done or the objective to be attained, and this should be the basis of the division of work between departments rather than, for example, dividing responsibility between departments so that each one deals with a client group.

The White Paper was however thin in its analysis and justification of this approach for the basis of the allocation of functions. As Johnson makes clear, while Governments may find it practicable to allocate activities functionally between departments the choice of organizational design is dependent on a hierarchy of objectives (Johnson 1971: 5). Functions and 'function principles' are in Johnsons's words, 'elusive terms'. Taking housing as a function, Johnson notes that this might mean 'the maintenance of satisfactory housing conditions in society'. But this says little of specific tasks involved in achieving this which may be various. This policy goal might be achieved for instance by financial incentives through taxation and the 'housing function' might then be best placed inside a finance ministry.

• created to give extra weight to a given policy, or to mark a change of direction in an existing policy
• created with an eye to giving a favourable public impression of dynamism and progressive reform
• created to adapt to major changes in the environment beyond Westminster and Whitehall
• created for greater technical efficiency and/or administrative savings and/or better coordination
• created to administer functions largely new to central government
• created to ease the Prime Minister's problems in deploying and 'balancing the talents of his/her senior colleagues'

Pollitt and others have shown that three key actors are involved in machinery of government decisions. Decisions result from informal deliberations between the Prime Minister, the Head of the Civil Service, and the Cabinet Secretary (although these last two posts have been combined since 1981) - the 'triumvirate' (Pollitt 1984: 126). The Machinery of Government Division of the Cabinet Office (more recently OMCS, OPSS and OPS, although formerly a part of the Treasury), has been involved in MG decisions to varying degrees over time. MG Division Pollitt argues was consulted more regularly when William Armstrong was Head of the Civil Service. Their role was one of 'tidying up and adjustment' (Pollitt 1984: 131). The Division when Bancroft and Caulcott headed it provided significant briefing and support, and generally played a more proactive part in MG decisions. The Treasury's role in MG decisions has usually been in handling the transfer of funds between departmental votes.

As far as affected ministers and permanent secretaries are concerned, Pollitt suggests evidence that they often have little notice of what changes are proposed. This is confirmed in the case of the latest machinery of government change involving the creation of the Department for Education and Employment. Michael Bichard, permanent secretary of the new department in a recent speech states clearly that 'No one, including the two permanent secretaries (myself and Sir Tim Lankester) and the secretaries of state (Michael Portillo and Gillian Shephard) knew that the merger was to happen until an hour before it was announced' (Bichard 1996: 22). While therefore ministers and officials over the period have been involved to different degrees in MG decisions, the role of the Triumvirate is paramount and dictates the degree of consultation with other groups. The role of Parliament in MG decisions is similarly limited. Few debates have been held on MG issues. Finally, pressure group influence on MG decisions has also been modest.

Machinery of government reform in Britain presents one particular example of what has been recently labelled, 'executive self-regulation' (Page 1995). Most modern departments (and non-departmental public bodies) have been established by statute. The current practice since the 1975 Ministers of the Crown Act has been to transfer functions to a new department by the exercise of the prerogative, for instance in the creation of a new department of state with new secretary of state. This tends to be the case when there is a re-allocation of existing statutory functions rather than the creation of new ones.

Until the 1946 Ministers of the Crown (Transfer of Functions) Act, the transfer of powers between departments could only be carried out by primary legislation (Lee 1977). The 1946 Act provided for the transfer of functions between offices and departments by order subject to the negative resolution procedure. Recent legislation has brought with it greater flexibility in dealing with issues of function transfers. The Deregulation and Contracting Out Act 1994 makes provision for contracting out and transfer of certain functions defined as eligible by ministerial order to the private sector. Such ministerial orders are subject to the affirmative resolution procedure.

There are a variety of ways in which machinery of government change occurs:

Primary legislation is needed for new statutory functions. Bodies however may be created administratively and then subsequently become statutory as in the case of the Commission on Industrial Relations (established 1969; became statutory in 1971). Government functions may be undertaken purely administratively. Certain functions have also been created by administrative action pending statutory authorisation. The Criminal Injuries Compensation Board was set up in 1964 but not made statutory until 1988 under the Criminal Justice Act 1988 Part VII.

Non-ministerial bodies may be created by:

1. primary legislation
2. statutory instrument conveying or transferring existing statutory powers or functions
3. under prerogative powers (royal charters and warrants)
4. by administrative action (i.e. by action by ministers not under specific statutory powers and not in exercise of the prerogative powers of the Crown)

There is no need however under the Ministers of the Crown Acts for primary legislation to be enacted for the establishment of new departments under a new Secretary of State. Both the DEA and Welsh Office for instance were created under prerogative powers and therefore, there was no necessity for a statutory instrument to dissolve the DEA.

Barker, A. [1982] Quangos in Britain (London: Macmillan).

Bichard, M. [1996] 'Shake-up Inspires New State of Mind', People Management, 8, February, pp. 22-27.

Chester, D.N. [1953] 'Public Corporations and the Classification of Administrative Bodies', Political Studies, Vol. 1, No. 1, pp. 34-52.

Chester, D.N. [1979] 'Fringe Bodies, Quangos and All That', Public Administration, Vol. 57, Spring, pp. 51-54.

Machinery of Government Division, Methods of MG Change (London: MG Division).

Dunleavy, P. [1989] 'The Architecture of the British Central State, Part I: Framework for Analysis', Public Administration, Vol. 67, Autumn, pp. 249-275.

Efficiency Unit [1991] Making the Most of Next Steps (London: HMSO).

Friedman [1951] 'The Legal Status and Organization of the Public Corporation', Law and Contemporary Problems, Autumn.

Greaves, H.R. [1947] Civil Service in the Changing State (London).

Greer, P. [1992] 'The Next Steps Initiative: an Examination of the Agency Framework Documents', Public Administration, Vol. 70, No. 1, pp. 89-98.

Gulick, L. [1937] 'Notes on the Theory of Organization' in Gulick, L. and Urwick, L (eds.) [1937] Papers on the Science of Administration (Columbia: Columbia University).

Hague, D.C., Mackenzie, W.J.M. and Barker, A. [1975] Public Policy and Private Interests (London: Macmillan).

Haldane, Lord [1918] Report of the Machinery of Government Committee Cd 9230 (London: HMSO).

Hammond, T.H. [1990] 'In Defence of Luther Gulick's "Notes on the Theory of Organization"', Public Administration, Vol. 68, Summer, pp. 143-173.

Hogwood, B.W. [1995] 'The "Growth" of Quangos: Evidence and Explanations', Parliamentary Affairs.

Hogwood, B.W. [1992] Trends in British Public Policy (Buckingham: Open University Press).

Hogwood, B.W. [1993] The Uneven Staircase: Measuring up to Next Steps (Glasgow: University of Strathclyde).

Hogwood, B.W. [1995] 'Whitehall Families: Core Departments and Agency Forms', International Review of Administrative Sciences, Vol. 61, pp. 511-530.

Hood, C. and Dunsire, A. [1981] Bureaumetrics: The Quantitative Comparison of British Central Government Agencies (Farnborough: Gower).

Hood, C., Dunsire, A. and Thompson, K.S. [1978] 'So You Think You Know What Government Departments Are...?', Public Administration Bulletin, Vol. 27, August, pp. 20-32.

Johnson, N. [1971] 'Editorial: The Reorganization of Central Government', Public Administration, Vol. 49, No. 1, pp. 1-12.

Jordan, G. [1994] The British Administrative System: Principles Versus Practice (London: Routledge, 1994).

Lee, J.M. [1977] Reviewing the Machinery of Government 1942-1952 (London: privately printed).