the executive agency revolution in whitehall: public interest versus bureau-shaping perspectives...

Oliver James

The Executive AgencyRevolution in Whitehall

Public Interest Versus

Bureau-Shaping Perspectives

Transforming Government

General Editor: R. A. W. Rhodes, Professor of Politics, University of Newcastle

This important and authoritative new series arises out of the seminal ESRCWhitehall Programme and seeks to fill the enormous gaps in our knowledge ofthe key actors and institutions of British government. It examines the manylarge changes during the postwar period and puts these into comparative context by analysing the experience of the advanced industrial democracies ofEurope and the nations of the Commonwealth. The series reports the results of the Whitehall Programme, a four-year project into change in British government in the postwar period, mounted by the Economic and SocialResearch Council.

Titles include:

Simon Bulmer, Martin Burch, Caitríona Carter, Patricia Hogwood and Andrew ScottBRITISH DEVOLUTION AND EUROPEAN POLICY-MAKINGTransforming Britain to Multi-Level Governance

Nicholas Deakin and Richard ParryTHE TREASURY AND SOCIAL POLICYThe Contest for Control of Welfare Strategy

Oliver JamesTHE EXECUTIVE AGENCY REVOLUTION IN WHITEHALLPublic Interest Versus Bureau-Shaping Perspectives

David Marsh, David Richards and Martin J. SmithCHANGING PATTERNS OF GOVERNANCE IN THE UNITED KINGDOMReinventing Whitehall?

B. Guy Peters, R. A. W. Rhodes and Vincent Wright (editors)ADMINISTERING THE SUMMITAdministration of the Core Executive in Developed Countries

R. A. W. Rhodes (editor)TRANSFORMING BRITISH GOVERNMENTVolume One: Changing InstitutionsVolume Two: Changing Roles and Relationships

Martin J. SmithTHE CORE EXECUTIVE IN BRITAIN

Kevin TheakstonLEADERSHIP IN WHITEHALL

Kevin Theakston (editor)BUREAUCRATS AND LEADERSHIP

Patrick Weller, Herman Bakvis and R. A. W. Rhodes (editors)THE HOLLOW CROWN

Transforming GovernmentSeries Standing Order ISBN 0–333–71580–2(outside North America only)

You can receive future titles in this series as they are published by placing a standingorder. Please contact your bookseller or, in case of difficulty, write to us at the addressbelow with your name and address, the title of the series and the ISBN quoted above.

Customer Services Department, Macmillan Distribution Ltd, Houndmills, Basingstoke,Hampshire RG21 6XS, England

The Executive AgencyRevolution in WhitehallPublic Interest Versus Bureau-Shaping Perspectives

Oliver JamesLecturer in PoliticsUniversity of Exeter

© Oliver James 2003

All rights reserved. No reproduction, copy or transmission of thispublication may be made without written permission.

No paragraph of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under theterms of any licence permitting limited copying issued by the CopyrightLicensing Agency, 90 Tottenham Court Road, London W1T 4LP.

Any person who does any unauthorised act in relation to this publication may be liable to criminal prosecution and civil claims fordamages.

The author has asserted his right to be identified as the author of this work in accordance with the Copyright,Designs and Patents Act 1988.

First published 2003 byPALGRAVE MACMILLANHoundmills, Basingstoke, Hampshire RG21 6XS and175 Fifth Avenue, New York, N.Y. 10010Companies and representatives throughout the world

PALGRAVE MACMILLAN is the global academic imprint of the PalgraveMacmillan division of St. Martin’s Press, LLC and of Palgrave MacmillanLtd. Macmillan® is a registered trademark in the United States,United Kingdom and other countries. Palgrave is a registered trademarkin the European Union and other countries.

ISBN 0–333–99838–3

This book is printed on paper suitable for recycling and made from fullymanaged and sustained forest sources.

A catalogue record for this book is available from the British Library.

Library of Congress Cataloging-in-Publication DataJames, Oliver, 1971–The executive agency revolution in Whitehall : public interest versus

bureau-shaping perspectives / Oliver James.p. cm. – (Transforming government)

Includes bibliographical references and index.ISBN 0–333–99838–31. Organizational change – Great Britain. 2. Administrative

agencies – Great Britain. 3. Contracting out – Great Britain.4. Privatization – Great Britain. I. Title. II. Transforming government(Palgrave Macmillan (Firm))

JN329.073J36 2003352.3�67�0941—dc21 2003045688

10 9 8 7 6 5 4 3 2 112 11 10 09 08 07 06 05 04 03

Printed and bound in Great Britain byAntony Rowe Ltd, Chippenham and Eastbourne

Contents

List of Tables viList of Figures viiAcknowledgements viiiList of Abbreviations ix

1 Introducing Executive Agencies 1

Part I Perspectives on Executive Agencies 15

2 The Public Interest and Bureau-Shaping Perspectives 17

Part II Executive Agencies in Practice 39

3 The Process of Executive Agency Creation 414 The Practice of Executive Agency Working 705 The Performance of Individual Executive Agencies 886 Executive Agencies and Central Government

Systemic Performance 108

Part III Conclusions 125

7 The Perspectives, Findings about Practice and the Future Use of Executive Agencies 127

Appendices 151

Appendix 1: The Bureau-Shaping Model 153Appendix 2: The Formal Bureau-Shaping Perspective on Executive Agency Reform 154Appendix 3: Interviews Conducted for the Study 156Appendix 4: Census of 173 Executive AgenciesCreated between 1988 and 2001 158Appendix 5: Central Government Accounts Audited 166

Bibliography 167

Index 181

v

List of Tables

2.1 The public interest perspective on executive agencies 232.2 Characteristics of a bureau which are either

positively or negatively valued by officials in Dunleavy’s (1991) bureau-shaping model 25

2.3 The bureau-shaping perspective on executive agencies 363.1 Hypotheses about executive agency reform 423.2 Growth in executive agencies and civil servants

working in these bodies 1988 to 2001 573.3 Type of executive agency creation in non-trading

and trading agencies 603.4 Executive agencies in the DSS, date of establishment,

former location of activity and type of creation 623.5 Senior staff (Grades 1–5) in parent departments

in 1995 655.1 Hypotheses about the performance of individual

executive agencies 885.2 Non-trading and trading agencies with real

terms administrative cost rises and falls between 1995/96 and 1997/98 90

5.3 Ministerial satisfaction with executive agencies 94

5.4 Benefits Agency: Secretary of State’s targets achieved/targets set between 1991/92 and 2000/01 96

5.5 Percentage of fraud and error in payments before and after the establishment of the Benefits Agency 98

6.1 Hypotheses about central government systemic performance 108

7.1 Privatisation of executive agencies 136A4.1 Percentage of departments’ staff in

executive agencies and staffing figures 165A5.1 Central government accounts audited

1993/94 to 1998/99 166

vi

List of Figures

1.1 A parent department and non-trading and trading agencies 4

2.1 A senior official’s preferences for the organisation of the department 27

2.2 A senior official’s preferences for organisation under politicians’ constraint 28

2.3 Shift in politicians’ constraint and change in a senior official’s preferred form of organisation 29

2.4 Niskanen’s budget/output maximising model 333.1 Institutions and actors involved in bringing

about Next Steps 1986–92 444.1 Appointment of chief executives in post in 1998 734.2 Benefits Agency and DSS Headquarters’

organisational chart 745.1 Box-plot of percentage change in real terms

administrative costs of seventy-two non-trading and trading agencies 1995/96 to 1997/98 89

5.2 Economy of Benefits Agency 1992–2001 916.1 Economy of central government 1989–2002 1096.2 The Benefits Agency and public sector

externalities in the social security system 1157.1 Structures for co-ordinating and steering

executive agencies 1377.2 Department for Work and Pensions 145

vii

Acknowledgements

I’d like to thank Keith Dowding, Patrick Dunleavy, Rod Rhodes andWayne Parsons and colleagues at the University of Exeter, particu-larly Andrew Hindmoor, Stephen Wilks and Bruce Doern, for theircomments on this work. Many other people have influenced mythinking on these issues, particularly Christopher Hood, GeorgeJones, Colin Scott and Tony Travers – as part of the group exploring‘regulation inside government’ under the ESRC WhitehallProgramme. I’ve also discussed this topic with, and would like tothank, Ian Bartle, Arjen Boin, Andrew Chadwick, Chris Clifford,Francesca Gains, Carsten Greve, Adrian Haddock, Richard Heffernan,Derek James, Onder Kutlu, Martin Lodge, Nick Manning, AndrewMassey, Richard Parry, Christopher Pollitt, Nigel Pleasants, DaveRichards, Martin Smith, Morwena Stephens, Jim Stanyer, ColinTalbot, Rebecca Tunstall, Byeong-soo Yoon and several public officials (but, by convention, I do not name them here).

viii

List of Abbreviations

BEM Business Excellence ModelBFI Benefit Fraud InspectorateCADW Welsh Historic Monuments AgencyDSS Department of Social SecurityDSS HQ Headquarters section of the Department of

Social SecurityDWP Department for Work and PensionsEA (WW) Economic Affairs (Welfare to Work) CommitteeESRC Economic and Social Research CouncilFMI Financial Management InitiativeHMSO Her Majesty’s Stationery OfficeIAI Independent Administrative Institution (in Japan)ICI Imperial Chemical IndustriesIT Information TechnologyITSA Information Technology Services AgencyLCD Lord Chancellor’s DepartmentMAB Ministerial Advisory BoardMAFF Ministry of Agriculture, Fisheries and FoodNAO National Audit OfficeNPM New Public ManagementNPR National Performance ReviewNI Northern IrelandOECD Organisation for Economic Co-operation and

DevelopmentPAC Public Accounts CommitteePBO Performance Based Organisation (in the US)PCA Parliamentary Commissioner for AdministrationPIU Performance and Innovation UnitPSA Public Service AgreementPSX Public Expenditure CommitteeUK United Kingdom of Great Britain and Northern IrelandUS United States of America

ix

1

1Introducing Executive Agencies

The increased use of executive agencies as a way of delivering goodsand services that are under the control of central government politi-cians is a key part of ‘New Public Management’ (NPM) reform to pub-lic services. There is now a large, international, literature using theterm ‘NPM’ in a range of different ways (Barzelay, 2000). However, aninfluential definition of NPM identifies a pattern of reform awayfrom ‘traditional’ forms of public organisation that was evident in aset of OECD countries, including the UK, during the 1980s and 1990s(Hood, 1991; Hoggett, 1991; Dunleavy, 1994; Pollitt and Bouckaert,2000). In the context of central government, NPM characteriseschange especially away from the use of departments to deliver pub-lic services. Traditional departments are headed by ministers who areelected politicians and represent the interests of the public. Theydevelop policy based on their view of the public’s demand for ser-vices and levy general taxation to fund these services, controlling theimplementation of policy and provision of services through theirdepartments. Civil servants staff departments and pursue longcareers in the public service. They are organised in a hierarchicalmanner beneath ministers who monitor and command action on a day to day basis as they see fit. Traditional central government,as a whole, is subject to distinctive sets of rules and styles of man-agement, set and enforced by central units, that emphasise input controls on activities, especially detailed budget and staffingcontrols.

NPM reform involves radical change to these ‘traditional’ structures. First, splitting up departments into corporate units with

distinct tasks and resources. The units are free from many central government-wide rules and day to day interference by ministers andoperate with generic, rather than specifically central government,management styles. Second, changing accountability and control sys-tems. Instead of primarily controlling detailed inputs including budgetsections or staffing, the systems use contractual frameworks specifyingan overall budget level and a set of desired outputs against which theperformance of the unit is monitored. Third, more use of competitiveand market-like structures for provision of goods and services. Thesestructures include competition with private firms, splitting purchasersof services from providers and charging users individually for thegoods and services they receive rather than funding from general tax-ation. Fourth, reducing the reliance on a distinct group of civil serviceemployees to deliver services by opening competition for appoint-ments to candidates from other parts of government and the privatesector. Associated with this change is a strengthening of incentives forcivil servants to produce outputs, including pay and promotion linkedto performance (Hood, 1991, pp. 4–5; Dunleavy, 1994; Pollitt andBouckaert, 2000, p. 10; Christensen and Laegreid, 2001).

The analysis of NPM changes has developed at a high level ofabstraction, noting the international pattern of change and thespread of similar forms rather than systematically exploring elementswithin the package of reforms. The attention of government andmost academic commentators has tended to focus on the newest ini-tiative rather than exploring the outcomes of reforms. Pollitt andBouckaert (2000, p. 132) note that ‘the international reform move-ment has not needed results to fuel its onward march’, in the sense ofsupport for further changes based on evidence of the consequences ofreform. This book analyses executive agencies and executive agencyreform, a key part of NPM that, since the late 1980s, has been animportant development in UK central government. Although the useof similar forms predates the current wave of reform and is far fromexclusively a phenomenon found in the UK, the UK experience pro-vides a good starting point for accumulating knowledge that is of rel-evance to central government both in the UK and in other countries.

Section 1: The executive agency model

The executive agency model is defined from the recommendations of a report by the Prime Minister’s Efficiency Unit, titled Improving

2 The Executive Agency Revolution in Whitehall

Management in Government: the Next Steps and the Government’s state-ment accepting these recommendations (Hansard, 1988, col. 1149).The Next Steps report proposed that ‘agencies should be establishedto carry out the executive functions of government within a policyand resources framework set by a department’ (Efficiency Unit, 1988,p. 9). The executive agency model has two main features:

(1) An executive agency organisation, semi-detached from its parentdepartment, with its own budget, freedom from some departmen-tal regulations, freedom from ad hoc, day to day, intervention bythe department, freedom from some central government-wideregulation, with the organisation under the direction of a chiefexecutive recruited through open competition.

(2) Executive agency accountability for the performance of specificoperational tasks as a corporate unit, including output focusedperformance targets set by the parent department, and personalaccountability of the chief executive for performance.

The partial organisational separation and accountability arrange-ments are opposite sides of the same coin. The accountability struc-tures set out the bodies to which the executive agency is accountablein terms of authorisation for undertaking tasks, methods of reportingperformance and praising or blaming those responsible. Each execu-tive agency is held directly to account for ‘operational’ matters andoutputs as specified in its operating framework whilst ministers indepartments are held to account for deciding the overall aims of the executive agency and the broader policy framework in which thebody operates (Efficiency Unit, 1988, pp. 17–19). Chief executiveshave personal accountability for operational performance and reportresults, with individual responsibility for the consequences of theiractions and a portion of their pay, and in extreme cases continuedtenure of their post, linked to individual performance. Ministersretain the right to alter the system and reorganise the status of exec-utive agencies, for example, to fold it back into the department andto intervene in cases of extreme need.

The freedoms accorded to executive agencies differ between thosethat predominantly receive their budget through an allocation bythe parent department and those that have freedom to trade bycharging customers ‘fees’ for services (Efficiency Unit, 1988, p. 28).Trading agencies have freedom to raise revenue from their customers

Introducing Executive Agencies 3

and to alter services in response to customer demand, forming anadditional form of accountability to customers. Trading agencies pro-vide information about services and may be sanctioned by customerswho may choose not to purchase from the body. All executive agen-cies are subject to systems of central government-wide regulation,although some freedoms are granted from general rules on the use ofresources and there is little regular reporting of performance to, ordirection by, central units. Executive agencies are incorporated in thesystem of ministerial accountability to Parliament with chief execu-tives reporting operational performance and being held responsiblefor their individual contribution to operational performance whilstministers report strategic performance and remain responsible forstrategic issues. The relationships between non-trading and tradingexecutive agencies, a parent department, central units, Parliament,customers and users of services are set out in Figure 1.1.

The executive agency model embodies four aspects of the NPMforms. The executive agency organisational structure is semi-detachedfrom the parent department and is a corporate unit with distinct tasksand resources. The accountability system reflects the NPM focus onoutputs, rather than detailed inputs, and the use of a contractualframework. The divide between the parent department and the exec-utive agency is a form of split between purchaser and provider in thecase of non-trading agencies, with trading agencies having a marketmechanism through payments for services by customers. Finally,recruiting chief executives through open competition involves


Performancetargets

Budgetcontrol

Revenue from customers

Chief ExecutiveAgency 1 (Non-trading)

Chief ExecutiveAgency 2 (Trading)

Parent Departmentcontains ministers,sets the strategicframework

Users

Parliament(executiveagenciesaccount foroperationalmatters)

Treasuryand othercentral units

Figure 1.1 A parent department and non-trading and trading agencies

recruitment across the public and private sectors rather than relyingon a group of career officials to run central government services.However, despite common elements, executive agencies are a dis-tinctive combination of NPM forms requiring specific research ontheir development and operation.

The ‘contract’ between a department and an executive agency isdifferent from a contract between a department and a private firmdelivering public services. The executive agency ‘contract’ is betweentwo bodies that are part of central government. Contracts with a pri-vate sector body are enforceable in a court in the case of a dispute(Domberger, 1998, pp. 160–2; Pollitt and Bouckaert, 2000, p. 94).Formally at least, the ultimate right to change the status of the exec-utive agency rests with the minister in the department rather thaninvolving exchange and agreement between equal parties. Tradingagencies have additional contracts between the body and private cus-tomers which have stronger similarities with private sector contractsin this aspect of these bodies’ arrangements. However, trading andnon-trading agencies are not private bodies in the sense of ownershipof resources being in private hands. Whilst the Next Steps reportoriginally noted that executive agencies ‘could be part of govern-ment and the public service’ or might be better ‘outside government’(Efficiency Unit, 1988, p. 9) the official announcement of the reformde-emphasised the use of this alternative, noting that ‘agencies willgenerally be within the civil service, and their staff will continue tobe civil servants’ (Hansard, 1987–88, col. 1157).

The framework in which an executive agency operates is differentfrom the ‘contract’ in NPM ‘internal’ or ‘quasi-markets’ between pub-licly funded and owned purchasers and providers (Domberger, 1998,pp. 157–80; Pollitt et al., 1998, pp. 5–13). In these arrangements, thepurchaser–provider contract sets out funding levels and the servicesexpected in return and there is competition between providers,sometimes including competition with private providers. In contrast,the executive agency model does not, in itself, suggest competitionbetween bodies to provide services, with the possibility of budgetbeing switched to or from executive agencies depending on their performance.

Despite the strong association with NPM, the executive agencymodel has historical precedents in the UK and elsewhere. As withother aspects of NPM, it was not completely ‘new’ and the NPM


characterisation of ‘traditional’ central government was a simplifiedrepresentation. The authors of the Next Steps report noted similari-ties between their proposals and the ‘arm’s-length’ relationshipsbetween central government and local government, nationalisedindustries and other non-departmental forms. These relationshipsconsisted of frameworks setting out ‘policies, objectives, the resultsrequired, and the resources available’ with ‘monitoring’ of perfor-mance (Efficiency Unit, 1988, pp. 11, 17–18). However, the executiveagency model embodies this framework in a distinctive way and theNext Steps reform was supposed to mark a step change in the use ofthis form, further suggesting the need for specific research on execu-tive agencies.

There were substantial differences between local authorities deliv-ering central government funded services and the proposed execu-tive agency model. Local authorities had a statutory frameworkunderpinning their separation from central government and moresubstantial areas of autonomy in setting their own objectives.Although this autonomy was in decline during the 1980s it was stillsubstantial in many areas of activities (Wilson and Game, 1994, pp.105–15). Nationalised industries typically had more formal indepen-dence from ministerial intervention and freedom from central gov-ernment-wide rules than executive agencies, as well as sharing thecharacteristic of freedom to raise revenue with trading agencies. Theindustries had a statutory base rather than the framework in whichthey operated being open to direct revision by ministers (Prosser,1986, pp. 17–36). Non-Departmental Public Bodies were the closest‘arm’s-length’ form to the executive agency model and this similar-ity was noted in the Next Steps report. Non-Departmental Public Bodies had a role in the processes of national government but werenot government departments or part of one and operated, to agreater or lesser extent, at ‘arm’s length’ from ministers (Pliatzky,1980). Such bodies were established over a long period, includingbodies set up in response to the Fulton Report’s recommendation of‘hiving off’ routine operational tasks to ‘autonomous public boards’(Fulton Committee, 1968, pp. 61–2). However, they typically hadgreater formal autonomy from departments and ministers in settingtheir priorities, and had greater freedom from central government-wideregulations than the executive agency model. For example, the NextSteps report mentioned the Health and Safety Executive as having


some similarities with the proposed form of organisation (EfficiencyUnit, 1988, p. 18). However, unlike the executive agency model, thisbody’s constitution was set out in statute and it reported to aCommission of members appointed by ministers who proposed pri-orities for action. These structures involved a more indirect and con-strained mechanism of ministerial control over the body than thatcontained in the executive agency model.

Specific research on executive agencies is particularly importantbecause they have come to dominate the structure of UK central government. The Next Steps reform was formally launched in 1988and, in 2001, 126 executive agencies employed almost 60 per cent ofall civil servants. The change appears to be revolution for ‘Whitehall’,the colloquial term for UK central government after the region ofLondon in which many senior ministerial and official staff arelocated, although the bulk of officials in executive agencies work outside of this geographical area. Executive agency working is thecontext in which more recent central government reform initiativeswere introduced. The prospects for the use of information technologyto provide ‘e-government’ and to promote ‘responsive’ governmentthat ‘efficiently’ produces the goods and services that citizens want(Minister for the Cabinet Office, 1999, pp. 6–7) will be strongly influ-enced by these structures. The UK experience is of further importancebecause the executive agency model has been influential internation-ally. Several countries have emulated the UK reform to varyingdegrees or embarked on their own similar programmes (Pollitt andBouckaert, 2000; Pollitt et al., 2000; James, 2001; Talbot, 2002).

Section 2: The scope of the study

This book addresses three key questions about executive agencies asthey have been used in the UK since 1988.

First, why did the executive agency reform occur? At the start ofthe reform, a Parliamentary Committee described the Next Stepsreport’s proposals as ‘the most ambitious attempt at Civil Servicereform in the twentieth century’ (Treasury and Civil ServiceCommittee, 1990, p. v). However, the reform of UK central govern-ment is conventionally seen as very difficult to achieve, in partbecause of resistance from those operating the Whitehall machine(Hennessy, 1989, pp. 622, 627). The initial scepticism about the


prospects for change and the eventual, apparently radical, outcomeof the Next Steps reform presents the puzzle of identifying the causesof the reform.

Second, how has the use of executive agencies developed in centralgovernment? Commentators noted the outcome ‘… that Next Stepsmight constitute a substantively new and stable formula for publicmanagement into the next century seems to be the least probablemedium term fate for the programme’ (Hood and Jones, 1990, p. 82).However, the government declared victory as early as 1994, stating‘The success of Next Steps has been in its effective implementation’and claiming that ‘… Next Steps has had a major impact on the shapeand culture of the Civil Service’ (Prime Minister, the Chancellor ofthe Exchequer, and the Chancellor of the Duchy of Lancaster, 1994,pp. 12–13). Despite this self-congratulation, there has been little sys-tematic assessment of whether the reform transformed the formalstructures and working practices of central government. The execu-tive agency model offers a benchmark against which developmentscan be assessed.

Third, have executive agencies improved the performance of central government? The principal assessment by government con-cluded ‘The agency model has been a success [bold in the original].Since 1988 agencies have transformed the landscape of governmentand the responsiveness and effectiveness of services delivered by gov-ernment’ (Office of Public Service Reform and HM Treasury, 2002, p. 10). This emphatic and sweeping judgement was principally qualified by the view that executive agencies and departments had,in some cases, become too ‘disconnected’. However, the report didnot support its bold conclusion by providing systematic informationabout the performance of executive agencies. Performance is signifi-cant at two levels, the level of individual bodies and the level of cen-tral government systemic performance. Individual performancerelates to distinct executive agencies and systemic performance con-sists of the consequences of individual executive agencies’ actions forthe performance of other bodies, including departments, and theinteraction effects between executive agencies.

Performance at both levels can be assessed using conventional criteria of economy, efficiency and effectiveness. Economy is theinput used in the activity, including the budget and staffing of indi-vidual executive agencies and other parts of central government.


Effectiveness is evaluated in terms of the achievement of central government’s objectives for individual executive agencies and forother bodies in government. There are several ways of assessing efficiency and a distinction is often made between allocative effi-ciency, which assesses the extent to which people get what theywant, and productive efficiency, which is the ratio between inputsand valued outputs or outcomes (Dowding, 1995, pp. 35–41; Pollittand Bouckaert, 2000, pp. 97–133). Productive efficiency is the mainfocus of inquiry in this study, partly for the practical reason that evi-dence about economy and effectiveness can be combined to produceevidence about this form of efficiency and partly because many ofthe benefits from the reform suggested by the Next Steps reformersrelated to this aspect of performance.

As well as addressing these three questions, two perspectives onexecutive agencies are evaluated. The public interest and bureau-shaping perspectives are each set out in Part I to provide hypothesesabout the reasons for the executive agency reform, the outcomes ofreform and consequences for performance. To the extent that theperspectives’ hypotheses receive support from practice in the UKthey offer the potential to build cumulative knowledge about execu-tive agencies though evaluating them in other contexts. There aremany approaches to public sector organisation and reform that offerthe potential to generate such perspectives (for general surveys ofapproaches see Caiden, 1991; Lane, 1993; Howlett and Ramesh,1995). However, the public interest and bureau perspectives offercontrasting and influential views.

The public interest perspective was presented by the Next Stepsreport and the Government’s statement formally accepting its rec-ommendations in 1988. To explain the reform, the perspective sug-gests that politicians in the Government interpreted the publicinterest as being furthered by executive agencies. Politicians saw thisform of organisation as offering a substantial improvement on tradi-tional departments’ treatment of executive activity. Civil servantsthen neutrally implemented their plan. The perspective views thereform as a fundamental change, with the outcomes improving theeconomy, productive efficiency and effectiveness of public serviceshandled by individual executive agencies with beneficial effects onthe systemic performance of central government. There has been little attempt by government or academics systematically to assess


this official view and whether the expectations of the Next Stepsreformers were met in practice.

A radical alternative to the public interest perspective is offered by the bureau-shaping perspective, developed from the bureau-shaping (Dunleavy, 1985, 1986, 1989a,b, 1991) and budget maximising (Niskanen, 1971) rational choice models of the publicsector. Rational choice models focus on the instrumental action ofindividuals, and their interaction, within institutional structures thatset the opportunities available to them. Whilst other alternatives tothe ‘official’ public interest perspective are possible, a rational choiceperspective is particularly worthwhile because of the relative scarcityof such work on the UK public sector. Rhodes notes that althoughrational choice offers a ‘challenge to the orthodoxy’ there have been‘few such studies’ and any potential that the approach might have‘has not yet been realised’ in a ‘major way’ (Rhodes, 1997, p. 175). Inparticular, the potential of Dunleavy’s bureau-shaping model hasbeen noted (Parsons, 1995, pp. 317–20; John, 1998, pp. 129–36). Thebureau-shaping perspective seeks to release some of the potential ofthe rational choice approach for exploring executive agencies.

To explain administrative reform, Dunleavy suggested that officialsembark on ‘bureau-shaping strategies’ to shape their bureaus so thatthey undertake favourably valued policy work and offload undesir-able activity to other bodies (Dunleavy, 1991, pp. 200–5). Thebureau-shaping perspective on executive agencies develops thisinsight to suggest that senior officials are concerned to organise theirdepartments to maximise their policy work and, at the same time, tohave as large as possible a budget associated with routine operationalactivity. However, ministers, who are also located in the depart-ments, constrain the organisational options available to senior offi-cials. The perspective suggests that, during the 1980s, ministersbecame increasingly dissatisfied with the lack of attention senior offi-cials paid to the management of operational activity because of theirpreoccupation with policy work. Politicians demanded that seniorofficials pay more attention to managing the activity that theirdepartments were responsible for handling. However, instead of giv-ing up policy work to concentrate on management, senior officialspassed on operational activity to executive agencies as a ‘bureau-shaping’ strategy to concentrate on policy work. Unlike the publicinterest perspective, Government politicians did not have a specific


plan for an executive agency reform which was then implemented by civil servants. Instead, politicians triggered a change and seniorofficials were central to shaping its form as an executive agencyreform.

The bureau-shaping perspective is more pessimistic about the consequences of executive agency reform for performance than thepublic interest perspective. Instead of being the result of a search for an organisational form that would bring about a significantimprovement over existing structures, reform was the result of abureau-shaping strategy. According to the perspective, the incentivesfacing officials in individual executive agencies lead them to budgetmaximise, improving effectiveness but worsening economy and leav-ing productive efficiency unchanged. At the systemic level, as well asa deterioration of economy, executive agencies do not take intoaccount the consequences of their activities on other executive agen-cies and departments that are not part of their own narrow perfor-mance regimes, setting up public sector externalities that damagecentral government systemic productive efficiency and effectiveness.

Section 3: The methodological approach

Part II of this book addresses the three questions about executiveagencies and assesses hypotheses drawn from the two perspectives byexploring the practice of UK central government since 1988. Not allaspects of the perspectives are reflected in the narrower hypothesesbut they encapsulate key elements of each perspective. It is not pos-sible to use experimental studies or to fully control the context inwhich cases are explored, for example to evaluate the performance ofactivities identical to those handled by executive agencies under-taken by a traditional departmental form. Instead, the empiricalassessment uses a range of methods to evaluate the hypothesesincluding surveys of executive agencies and case studies. Obtainingthe sort of evidence necessary for such an empirical assessment isnotoriously difficult because sources are fragmented and incomplete(Pollitt and Bouckaert, 2000, p. 131). However, reform is charac-terised by the number and type of executive agencies, nature of executive agency creation and location of staff after the changes. The operation of the executive agency model and performance ofthese bodies is assessed in terms of changes in budget, staffing and


performance against targets. The survey is supplemented by morefine-grained case studies, in particular focusing on the social securitysector and the largest executive agency, the Benefits Agency. Thecase-study approach uses a range of sources of evidence includingmaterial from the survey, interviews and analysis of internal andpublished documents. Whilst conclusions from case studies must notbe over-generalised, the approach facilitates the gathering of evi-dence about relationships that are important for identifying theinteraction of actors and institutions that brought about the changesto evaluate the explanations of reform. The approach provides aninsight into the relationship between the executive agency model’sstructure and consequences for economy, productive efficiency andeffectiveness at the individual and systemic levels.

Assessing the performance consequences of reform is particularlydifficult because the objectives of government are difficult to identifywhere these are not written down, and written objectives do notalways reflect the full range of desired outcomes. Reasonable startingpoints are the ‘objectives’, ‘aims’ and ‘targets’ contained in executiveagency, departmental and programme reports. An assessment of thesystemic consequences is hampered by the lack of a document withan authoritative statement of the relation between objectives acrossgovernment that could be used to rank their importance where theyconflict. However, by using other measures of priorities includingstatements by ministers and officials, a judgement about tensionsbetween objectives can be made. This approach is used to explore theeffect of the Benefits Agency on systemic performance in the socialsecurity sector.

Official statements and interview material are not sufficient foruncovering what happened during the period, particularly for estab-lishing actors’ motivations and interests when they had an interestin conveying a particular version of events. For example, commentsby senior officials in the Cabinet Office that Next Steps was a success(Kemp, 1990; Butler, 1991, 1994) need to be viewed in the light oftheir responsibility for undertaking the reform and career and otherreasons for wanting to portray the changes as successful. Officialstatements and interview material is useful when it is cross-checkedwith other sources of evidence. Important alternative sources of evidence are the actions of individuals which usually, although notalways, have a higher cost for actors than making statements


because they imply that alternative actions must be forgone. Actionsmay be more likely to reflect underlying motivations and inter-ests than ‘talk’ which may be relatively ‘cheap’ in these terms. Actionsinclude choices about organisational arrangements and career choicesby officials. For example, whether officials who were involved inbringing about the reform eventually ended up working in execu-tive agencies is an indicator of whether they liked the type of workundertaken in these organisations, although not, by itself, enough toestablish this connection.

Actions, by themselves, may not always be a reliable guide to moti-vations and interests. The structure of a situation in which a choice ofaction is made may encourage strategic misrepresentation of intereststo try and gain longer-term advantage. Alternatively, the structuremay suggest a chosen outcome is more desired than one that is notchosen despite the fact that the one that is not chosen is more greatlydesired. An example of this problem occurs in situations resemblingthe well-known prisoners’ dilemma game. In the two player versionof this game, self-interested individuals A and B are faced with achoice between co-operation and defection with personal payoffsfrom the outcomes of their choices. If A co-operates and B does not,this is the worst outcome for A and best for B. If B co-operates and Adoes not, this is the worst outcome for B and best for A. The secondworst outcome for both is if they both do not co-operate. The secondbest outcome for both occurs if they both co-operate.

In the game, the individuals cannot co-ordinate their actions and,given the structure of the individual payoffs, they compare the out-come from co-operating or defecting under the possible outcomes ofchoices for the other player. This comparison leads each player sepa-rately to choose to defect. The resulting joint outcome of double defec-tion brings a lower payoff for each individual than would have beenachieved if they had chosen to co-operate. In this situation, observingthe action of the individuals, by itself, is not a good guide for workingout what is in the actors’ interests (Sen, 1973, pp. 249–53). Whilst thegame structure is hypothetical, the point that complex interactionsbetween choices may determine action which is not in the best inter-est of the actors involved in those choices appears likely to be highlyrelevant to complex organisations like central government.

Combining different sources of evidence enables cross-checking ofevidence relating to each question and hypothesis. In the light of this


evidence, answers to the three questions and conclusions about therelative merits of the public interest and bureau-shaping perspectivesare set out in Part III. This Part suggests that the perspectives and find-ings from this study are potentially useful for exploring the likelyfuture use of executive agencies in the UK and for exploring executiveagencies in other countries.


Part I

Perspectives on ExecutiveAgencies

2The Public Interest and Bureau-Shaping Perspectives

The two perspectives offer hypotheses about executive agency reformand the outcomes of reform including the performance of individualexecutive agencies and central government systemic performance.Section 1 sets out the public interest perspective which embodies theofficial view contained in the Next Steps report and the Governmentstatement accepting its recommendations. Section 2 sets out thealternative, rational choice, bureau-shaping perspective.

Section 1: The public interest perspective

Public interest explanations of policy are difficult to articulatebecause they involve a single overall assessment of benefits and coststo a collection of individuals or groups labelled ‘the public’ in termsof how the things these people want are affected by a policy (Barry,1990, pp. 187–206). However, the public interest perspective’s expla-nation of reform is that offered in the Next Steps report and in theGovernment statement to the House of Commons accepting thereport on 18 February 1988. These official statements contained spe-cific claims about why the reform was occurring.

The perspective explains the reform as the result of central government politicians, both at the centre of the administration andministers in departments, acting on behalf of the public, defined asa broad group including service users and taxpayers. The Govern-ment maintained that it directed the Efficiency Unit to come up with a reform that would significantly improve the performance ofcentral government. The Prime Minister commented that ‘I asked the

17

Efficiency Unit to report to me on the progress of managementreforms in the Civil Service’ and to make recommendations forimprovements (Hansard, 1988, col. 1149). On the basis of this advice,central government politicians made the key decision to pursue a spe-cific executive agency reform to improve the management of the civilservice.

The Next Steps report reviewed the implementation of earlierreforms in central government, assessed current practices and maderecommendations for improvement, principally to create executiveagencies. The report suggested a process for implementing the reform,with a central management team under the direction of the PrimeMinister and Head of the Civil Service to lead the change programme.A senior, Permanent Secretary level appointment was suggested as‘Project Manager’ to lead the team combined with support from ‘the highest level’ of government (Efficiency Unit, 1988, pp. 12–13).This team would set the framework within which departments identified and established executive agencies (Efficiency Unit, 1988,pp. 4–5 and 9–13).

The reformers suggested the potential for massive growth in theuse of executive agencies, noting that 95 per cent of the managementand staff in the Civil Service were ‘concerned with the delivery ofgovernment services’ in activities suitable for handling by ‘agencies’(Efficiency Unit, 1988, p. 3). The report commented ‘In someinstances very large blocks of work comprising virtually a wholedepartment will be suitable to be managed in this way. In otherinstances, where the scale of activity is too small for an entirely sep-arate organisation, it may be better to have one or even severalsmaller agencies within departments’ (Efficiency Unit, 1988, p. 9).

The reformers suggested that executive agencies would producesubstantial improvements over existing structures with better perfor-mance of tasks undertaken by these bodies and improvements tocentral government systemic performance. The benefits were to bebrought about through executive agencies’ management freedoms.Managers would be able to use the freedoms given to them to apply their skills and knowledge to achieve their specific tasks. Thevaried activities of the civil service were seen as too diverse to have a standard approach. The report noted the ‘frustration’ of thoseinvolved in delivering services directed at unnecessary controls andthe benefits from removing them (Efficiency Unit, 1988, pp. 3–4).


The importance of freedoms in recruitment, pay and grading wereespecially noted, allowing more appropriate management of staff.Executive agencies’ freedom from departmental ministers and offi-cials’ day to day, ad hoc, intervention in the running of the bodieswould allow managers to get on with directing resources.

There were additional benefits for the central government systemfrom this change. The reformers suggested that politicians were suf-fering from ‘ministerial overload’. Handing responsibility for day today management to executive agencies would reduce their workloadand enable them to concentrate on reporting to Parliament anddeveloping policy (Efficiency Unit, 1988, p. 4). However, the reportnoted that the reform should increase the involvement of all civilservants in the delivery of services, commenting ‘It is most importantthat there are not two classes of people in departments – those inagencies and those in the centre. The aim must be to have seniormanagers who at more junior levels have had substantial experienceof the skills and practical reality of management as well as effectiveexperience of the political and policy aspects of work in a depart-ment’ (Efficiency Unit, 1988, p. 12).

The accountability systems for executive agencies and their man-agers, especially the personal accountability of chief executives, wouldcontribute to better performance. The report commented that exec-utive agencies ‘need to be given a well defined framework in whichto operate, which sets out the policy, the budget, specific targets andthe results to be achieved’ continuing, ‘The management of theagency must be held rigorously to account by their department forthe results they achieve’ (Efficiency Unit, 1988, p. 9). There would bebetter information about progress in the performance of tasks with‘regular monitoring’ by the department, with ‘effective measures ofperformance’ including measures focused on outputs (EfficiencyUnit, 1988, p. 11).

The accountability systems would improve the allocation ofresponsibility for action and its outcomes. The executive agencywould account as a unit for operational matters whilst the depart-ment would be responsible for the strategic framework, clearly iden-tifying responsibilities for these different types of action. The ‘head ofthe agency’, later called the chief executive, would have ‘personalresponsibility for achieving the best possible results’ within the frame-work set for the body with ministers remaining ‘wholly responsible

The Public Interest and Bureau-Shaping Perspectives 19

for policy’ (Efficiency Unit, 1988, p. 10). Executive agencies would beaccountable as corporate units to other parts of government, users,other interested groups and Parliament. Chief executives would beheld accountable to Parliament in providing information and beingheld responsible for their individual contribution to operational matters (Efficiency Unit, 1988, p. 17).

The incentives created by these systems were not specified by theNext Steps report in great detail. However, the report noted the needto overcome the ‘lack of effective pressure to get better results’ and tocreate ‘real and sustained pressure on and within each departmentfor continuous improvement in value for money obtained in thedelivery of policies and services’ (Efficiency Unit, 1988, p. 7). Theimproved information about executive agencies’ performance couldbe used to allocate praise or blame and inform decisions aboutstaffing. The report observed that promotion was traditionally linkedto age rather than outcomes and cited the possibility of promotingoutstanding younger officials more quickly. This sort of career incen-tive was seen as an important way to encourage better performance,especially for the heads of executive agencies (Efficiency Unit, 1988,pp. 29–30). In announcing the reform, the Prime Minister noted thatexecutive agencies would help the government ‘obtain better valuefor money through personal responsibility’ for performance and thepotential for more ‘performance related pay’ as a way of ‘rewardingexcellent performance’ (Hansard, 1988, col. 1152–3).

Trading agencies have additional management freedoms andaccountability structures linked to raising revenues by charging cus-tomers for services, in contrast to non-trading agencies that receivegrants from their department to cover most of the cost of their activ-ities. Trading agencies’ partial dependence on revenues from cus-tomers was seen as providing an incentive for them to respond tocustomers’ requirements in a flexible way, for example, to expandservices in response to bottlenecks in supply (Efficiency Unit, 1988,p. 28). Customers might be able to choose alternative sources of sup-ply or not demand the bodies’ services, reducing their revenues. Atthe same time, their managerial freedoms would enable them tospend revenues in an appropriate manner to satisfy demand. Thepublic interest perspective hypothesises that trading agencies shouldbe used where such customer groups exist.


Advantages in economy, efficiency and effectiveness of individualexecutive agencies and central government systemic performancewere suggested. The reformers did not say when the improvementsshould appear. However, in talking about reforms since 1979, theysuggested the need for more ‘urgency in the search for better valuefor money and steadily improving services’ (Efficiency Unit, 1988, p. 1). From these comments, and the general tone of the report, itappears that improvements were expected within a few years ofestablishing executive agencies.

Improvements in economy, with fewer inputs, were suggested. Thereport commented that ‘it is difficult to put a figure on the benefitswhich should become available from our recommendations but the potential is obvious. Five per cent of Civil Service running costs amounted to £630 m in 1986–87, and experience elsewhere certainly indicates that when good management has the opportunityto perform well, percentage improvements larger than this areachieved’ (Efficiency Unit, 1988, p. 16). This statement suggest that5 per cent improvements in administrative costs were expected forindividual executive agencies within a few years. Executive agencieswere noted as a way of achieving ‘savings’ in government and con-tributing to ‘cost conscious’ working by civil servants (EfficiencyUnit, 1988, pp. 16, 21). These statements suggest that a reduction inprogramme costs, the budget handled by the executive agency but passed on in grants or contracts rather than being spent on direct administration, would flow from the reform. The report didnot explicitly consider trading agencies’ economy. Whilst the reportpointed to instances of these bodies raising revenue from customergroups, and noted that this was desirable activity, it did not discussthe possibility that this structure might result in lower economy ifadditional customer revenues were used to expand output and bud-get. Instead the general tone of the report was to note the potentialfor improvements in economy for all executive agencies. Similarly,the report suggested reductions in central government systemicadministrative and programme costs, with departments reducing‘overload’ by strategic rather than day to day monitoring of activitiesin executive agencies (Efficiency Unit, 1988, p. 4).

The reformers emphasised the potential benefits for effectivenessof service delivery that would flow from the reform. The report


viewed executive agencies as instruments for better achieving theobjectives of individual executive agencies and the overall systemthan the old arrangements. The report stated ‘With total programmeexpenditure of £128 b (1985–86), there is an immense opportunity togo for substantial improvement in outputs, with better delivery ofservices and reduced delays as an alternative to savings’ (EfficiencyUnit, 1988, p. 16). The report commented that the traditional systempaid ‘too little attention to the results to be achieved with theresources [of government]’ and there were ‘relatively few externalpressures [on parts of government] demanding improvement in per-formance’ (Efficiency Unit, 1988, p. 4). Executive agencies wouldimprove effectiveness by producing management ‘eager to maximiseresults, no longer frustrated or absolved from responsibility by cen-tral constraints and working with a sense of urgency to improve theirservice’ (Efficiency Unit, 1988, p. 16).

As well as improvements to economy and effectiveness, the reportnoted the potential for improvements in productive efficiency.Although the Efficiency Unit that produced the Next Steps reportcontained the term ‘efficiency’ in its title, suggesting a primary con-cern with this aspect of government performance, there was little dis-cussion of the meaning of this term in the report. However, thereport’s focus on reduced inputs and improvement in achievingobjectives suggests improved productive efficiency. The report didnot clearly state the relative contributions of economy and effective-ness to improved productive efficiency which could be improved by alterations to economy, effectiveness or both of these goals. For example, if budget rises but output rises faster than inputs then productive efficiency improves but economy worsens, andimprovements in economy could lower effectiveness but increaseproductive efficiency. However, the report claimed that improve-ment in administrative costs of individual executive agencies ofaround 5 per cent were possible, suggesting that improvements in services as an alternative to savings would occur once this thresh-old reduction had been reached. Systemic productive efficiencywould improve, in part through departments’ savings in administra-tive and programme costs through executive agency working. Themain hypotheses of the public interest perspective are summarised inTable 2.1.


Section 2: The bureau-shaping perspective

The bureau-shaping perspective offers a radical alternative to thepublic interest perspective incorporating insights from rationalchoice models of the public sector to generate hypotheses aboutexecutive agencies. There are many rational choice models of thepublic sector but very few were developed specifically for examiningreorganisation, and none were developed specifically to understandexecutive agency reform (for a survey see Mueller, 1989; Dunleavy,1991; Dowding, 1995; Weimer and Vining, 1996). The bureau-shaping perspective draws on two influential rational choice modelsof bureaucracy, the bureau-shaping model developed by Dunleavy(1985, 1986, 1989a,b, 1991) and Niskanen’s (1971) budget/outputmaximising model.


Table 2.1 The public interest perspective on executive agencies

Aspect Hypotheses

Reform 1. Government politicians control all aspects of departmentalprocess and organisation and suggest the executive agency reform outcomes plan in the public interest

2. Central units co-ordinate reform which is neutrallyimplemented by officials as part of a hierarchy and the use of executive agencies increases to around 95 per cent of the Civil Service within a few years

3. All civil servants have experience of working in executiveagencies which operate according to the model

Individual 1. Economy improves, with at least 5 per cent savings inexecutive administrative costs and savings in programme costsagencies’ 2. Effectiveness improves with better achievement of performance objectives

3. Productive efficiency improves including at least 5 per centsavings in administrative costs and savings in programme costs

Central 1. Economy improves with savings in administrative andgovernment programme costssystemic 2. Effectiveness improves with better achievement of performance objectives

3. Productive efficiency improves including savings inadministrative and programme costs

Dunleavy developed the explanatory part of the bureau-shapingmodel as a general model to explain a range of administrative situa-tions and experiences including inappropriate privatisation, admin-istrative reform and deinstitutionalisation across central and localgovernment (see Appendix 1). He suggested the potential of thebureau-shaping model to explain executive agency reform as abureau-shaping strategy by officials. Writing at an early stage of theNext Steps reform he argued that ‘most senior officials in other [non Treasury] Whitehall ministries have endorsed the government’sstrategy [for Next Steps]. The full implementation of the reformswould drastically reduce their departments’ personnel and core budgets … On the bureau-shaping model the hiving-off proposalsprovide senior bureaucrats with a unique opportunity to engage inwholesale reshaping of their bureaux to obtain the ideal form ofsmall, elite, staff agencies. The government report which triggeredthe reorganisation process candidly admitted that political and pol-icy proposals preoccupied policy-level officials and inherently werealways likely to do so’ (Dunleavy, 1991, p. 226). However, the origi-nal bureau-shaping model does not offer a fully developed explana-tion of executive agency reform. The model does not fully specifyindividual officials’ preferences for the organisation of the depart-ment in which they work and the relationship between officials andother actors in the system is not set out, particularly the relationshipbetween politicians and officials (James, 1995b, p. 616; John, 1998,pp. 133–4; Marsh et al., 2000, 2001, pp. 163–4).

The bureau-shaping perspective builds on the bureau-shapingmodel to generate hypotheses about the circumstances under whichsenior officials in departments will allocate activity to executiveagencies. The perspective assumes that central government is com-posed of separate departments staffed by officials and headed bypoliticians. Senior officials are given prominence in the explanationas people within central government whose rank within depart-ments and detailed knowledge of organisation enables them to influence reform. The perspective posits that individual senior offi-cials have preferences related to two principal features of the organi-sation of the department in which they work, the proportion ofpolicy work time in total work time, and the level of budget persenior official.


Senior officials are concerned about the proportion of policy worktime in their total work time, building on the insight that ‘welfaremaximising officials in policy making ranks are primarily concernedto improve their welfare by providing themselves with congenial workand a valued work environment’ (Dunleavy, 1991, p. 200). Positiveand negative assessments relating to the characteristics of work in anorganisation suggested by Dunleavy are summarised in Table 2.2.

The bureau-shaping perspective incorporates senior officials’ con-cerns about the nature of their work in a simplified way. Policy workis defined as involving innovation and entailing working in smallstaff units in close proximity to political power sources, such as min-isterial private offices. The perspective assumes that a senior officialwill prefer a higher proportion of policy work time in total work timeto a lower proportion if offered a choice between different levels. Byimplication, senior officials are assumed to desire as low a proportionof alternative work as possible. This alternative work is assumed to bemanagerial work involving the routine implementation of proce-dures, hands-on monitoring and direction of junior staff, and entailsworking in a large, extended hierarchy at the point of delivery,remote from political power sources. Having these concerns does notimply that senior officials are necessarily completely self-regardingactors. Instead, the reasons for the concerns may be quite complex,


Table 2.2 Characteristics of a bureau which are either positively or negativelyvalued by officials in Dunleavy’s (1991) bureau-shaping model

Positively valued Negatively valued

Staff functions including Line functions including routine work,innovative work, high level of low discretion and narrow tasksdiscretion and broadscope of concerns

Collegial atmosphere including Corporate atmosphere including largesmall work units and work units and non-elite and conflictco-operative work patterns ridden work patterns

Central location including Peripheral location includingproximity to political power remoteness from political power sources and high status sources and remoteness from highsocial contacts status social contacts

Source: Adapted from Dunleavy (1991, p. 202).

including an official feeling that this is the best way to deploy his orher skills for the good of society. However, for the purposes of thisstudy, the concern with having as much policy work time as possibleis the key assumption for generating hypotheses about reform.

The second aspect of departmental organisation about which asenior official is concerned is the level of budget and associated output per senior official. The concern with the level of budget/output draws on an insight from the budget/output maximisingmodel of bureaucracy (Niskanen, 1971). In Niskanen’s model, indi-vidual bureaucrats seek to maximise budgets because ‘Among theseveral variable that may enter the bureaucrat’s utility function arethe following: salary, prerequisites of the office, public reputation,power, patronage, output of the bureau, ease of making changes andease of managing the bureau. All of these variables except the lasttwo, I contend, are a positive monotonic function of the total bud-get of the bureau during the bureaucrat’s tenure in office’ (Niskanen,1971, p. 38). However, Niskanen states that officials are able to pushup budgets only through the production of more output for a spon-sor of their bureau. The need to produce output for the sponsormeans that the Niskanen model is really a model of budget/outputmaximisation rather than just budget maximisation. Orzechowskicomments that ‘Niskanen assumes that the goal of the bureaucrat isto generate the largest possible budget. This is equivalent to sayingthat bureaucrats maximise output subject to the bureau’s budget constraint. Carrying this to its logical conclusion would imply zerofiscal residuum remaining for such utility dimensions as salary, pre-requisites and security – dimensions which Niskanen assumes werethe consequences of budget maximisation’ (Orzechowski, 1977, p. 230). However, if the Niskanen model is interpreted as consistingof budget/output maximising motivation, with corresponding actionby officials, then this objection is overcome.

In the bureau-shaping perspective, senior officials prefer as high aspossible a level of budget per senior official associated with directlyproducing output by the department. A concern with output andbudget might be for reasons of personal benefit, with higher budgetsbeing associated with higher salaries as Niskanen suggests, or mightbe because a senior official perceives a government programme asworthwhile for society and wants to be in direct control of it. The keyassumption is only that senior officials are concerned about the level


of budget per senior official controlled directly by their departments.A senior official will prefer a higher level of budget per senior officialto a lower level in choices between different levels of budget andbudget is positively related to executive activity, rather than beingrelated to policy work time, because budgets are predominantly asso-ciated with executive programmes of service delivery rather than pol-icy work, which primarily entails only the relatively minor salarycosts of officials engaged in the policy work.

The preferences of an individual senior official for the organisationof the department in which he or she works are shown in Figure 2.1.The departments undertake a set amount of policy work and execu-tive activity, with its associated budget. Each indifference curve linksequally preferred organisational configurations. The curves areranked in order of preference, with curves furthest from the originrepresenting more preferred combinations (U3 is preferred to U2which is preferred to U1). The slope of each curve represents the rateof substitution required between levels of budget per senior officialand proportion of policy work time in total work time to maintainan equally preferred combination. The curves are convex to the origin because a diminishing marginal rate of substitution betweenproportion of policy work time and budget per senior official isassumed. At high levels of budget, an individual requires more addi-tional budget to compensate for the loss of a unit of policy work timethan at low levels of budget.

The bureau-shaping perspective specifies the structure in whichsenior departmental officials work which affects the way in whichthey seek to use executive agency reform as a bureau-shaping strategy.


U1

U2

U3

X= budget per senior official

Y=

pro

port

ion

of p

olic

y w

ork

time

in to

tal w

ork

time

Figure 2.1 A senior official’s preferences for the organisation of the department

Senior officials’ freedom to chose any form of departmental organi-sation, which would tend towards departments having all the possi-ble policy work and budget available, is limited by politicians’ viewson what organisational forms offer economic, productively efficientand effective combinations of activities. In this way, the perspectiveassumes that politicians have some interest in trying to promoteeconomy, productive efficiency and effectiveness, as in the publicinterest view. However, politicians do not develop detailed plans forthe organisation of their department, they have little direct interestin, or competence about, organisational matters beyond their day today interaction with officials in the department. Instead, their behav-iour is approximated as a rule of thumb about organisation thathigher levels of budget per senior official imply greater need for offi-cials to spend time on monitoring activities within the organisation,and less time on policy work, than lower levels of budget per seniorofficial. They surmise that it is not sufficient simply to employ moresenior officials within the department as managers, instead it is nec-essary for those that are at the heart of the organisation to knowthemselves what is going on down the line.

The preferences of an individual senior official operating under theconstraint in a department are shown schematically in Figure 2.2. Inthis Figure, Department A at X1, Y1 is the preferred and stable out-come. The alternative arrangement of X2, Y2 represents a department


Y =

pro

port

ion

of p

olic

y w

ork

time

X = budget per senior official

A

X1

Y1

1

Y2

X3X2

Y3

Figure 2.2 A senior official’s preferences for organisation under politicians’constraint

predominantly involved in policy work where ministers surmise thatit is detrimental simultaneously to be involved in a substantial pro-gramme of service delivery with a high budget. The alternativearrangement X3, Y3 represents a department heavily involved in asubstantial programme of service delivery, where ministers surmisethat it is detrimental also to be involved in a high proportion of policy work time.

The bureau-shaping perspective on the organisation of a depart-ment is presented formally in Appendix 2, which shows how a seniorofficial maximises budget per senior official and policy work time subject to politicians’ constraint. The perspective suggests severalhypotheses about the causes of executive agency reform including thetriggers, processes and outcomes of reform. The perspective does notsuggest that individual senior officials sit down and draw figures sim-ilar to those in Figure 2.2 to work out their best strategy but insteadthat they are concerned with these two aspects of organisation andwill pursue reforms that approximate to obtaining their most pre-ferred arrangement. Bureau-shaping strategies to change the organi-sation of a department from a stable arrangement at point A in Figure 2.2 are triggered by shifts in the constraint on senior officials.In Figure 2.3, under constraint 1, senior officials prefer point A toother possible organisational arrangements. However, a shift in the


Y =

pro

port

ion

of p

olic

y w

ork

time

X = budget per senior official

A

X1

Y1

1

Y2

X2

C

B

Constraint 2

Constraint 1

Figure 2.3 Shift in politicians’ constraint and change in a senior official’s preferred form of organisation

constraint alters the favoured organisational arrangement. Whilstthe constraint could shift either way, a shift from constraint 1 to constraint 2 in Figure 2.3 represents an increase in the proportion ofmanagement work time for each unit of budget per senior official.The shift makes A unobtainable and B the most preferred positionwithin the new constraint. Senior officials embark on a bureau-shaping strategy entailing changes from X1 to X2 and Y1 to Y2 inFigure 2.3 to bring about the organisational form corresponding topoint B, passing on responsibility for executive activity to executiveagencies and concentrating on policy work.

The creation of executive agencies passes executive activity andassociated budget to these bodies. At the same time, the departmentbrings policy activity within its direct purview and concentrates onsetting the policy framework in which the executive agency operates.More fundamental changes, such as outright privatisation of owner-ship with central government disengaging from the activity wouldnot be supported by senior officials because they would call thedepartment’s continued policy role into question. However, seniorofficials prefer an organisational form corresponding to point B toother organisational forms available to them under the shifted con-straint. For example, point B is preferred to the organisational formcorresponding to point C, an alternative form in which execu-tive activities and associated budgets are kept in-house with seniorofficials becoming hands on managers.

Bureau-shaping changes involve individual senior officials actingtogether to bring about the necessary alterations, suggesting thatofficials will discuss options for reform, formulate plans and imple-ment them. Reforms that are not consistent with senior officials’preferences will not be proposed or implemented. Bureau-shapingstrategies incur costs for individual officials in time and effort advo-cating reform. The strategies produce non-rival benefits that are con-sumed without implying a reduction in benefits received by othersenior officials. All the members of the shaped organisation benefitfrom the change, it being difficult to exclude non-contributorsbecause they are members of the shaped organisation. Such publicgoods, in the rational choice framework, tend to lead to free riderswho seek to benefit without contributing to costs, which may resultin the good not being provided at all. But bureau-shaping strategiesare assumed not to incur great costs for officials. Politicians set


the range of possible combinations of budget per senior official andproportion of policy work time and are indifferent between thesepoints.

Several methods could, theoretically, be used to evaluate thehypotheses in the context of Next Steps. An attempt could be madeto work out the precise trade-offs in choices of different bundles ofpolicy work time and budget, in terms of how much of one isrequired to compensate for a loss of a unit of the other. Alternatively,the responses in desired amounts of budget and policy work timerelated to shifts in politicians estimate of the time needed for budgetsupervision could be compared against actual responses. Assumingthe total time available to senior officials is fixed, the demand curvefor budget and policy work time relating quantity demanded to the‘price’ of budget as hypothesised by the perspective could be com-pared to data obtained from observing officials’ actions in practice.The perspective assumes that the ‘price’ of policy work time is fixedat the unit of time needed to undertake such work, so only the ‘price’of budget varies according to shifts in politicians’ constraint.Developing such an analysis, the expected elasticity of demand,defined as the percentage change in quantity demanded divided bythe percentage change in ‘price’, could be compared to that found inthe data from observations of officials. However, the evidenceneeded for such a detailed empirical evaluation would be very diffi-cult to obtain, requiring a very fine grained analysis of officials’ workover a long time period, including a period before this study was con-ceived. In any case, a less demanding method of evaluation is suffi-cient to assess the main hypotheses about key aspects of reformprocesses and outcomes.

The perspective produces the hypothesis that senior officials areconcerned with the organisation of their departments in a way setout in the indifference curve mapping. A bureau-shaping strategy istriggered by a shift in politicians’ constraint on officials with themagnitude of the shift determining the scale of activity passed on toexecutive agencies. For the perspective to be consistent with the NextSteps reform, the choices made by senior officials during the periodshould be consistent with the preferences set out in the perspectiverather than those preferences changing. In choices between differentorganisational forms in Figure 2.3, position A should be chosen over B and C under constraint 1, and B over C under constraint 2.


Senior officials should be observed both to attempt and succeed ingetting the best form of departmental organisation available to themwithin the constraints set by politicians. Under circumstances of aninward shift in politicians’ constraint from 1 to 2, the perspectivesuggests that there would be a passing on of activities to executiveagencies with an increase in the budget and other resources allocatedto executive agencies both within individual departments and forcentral government in general. The magnitude of these changeswould be positively related to the magnitude of the shift in politi-cians’ constraint. A substantial inward shift in the constraint is takento imply that between half and three-quarters of executive activityhandled by departments would be passed on. The perspective impliesthat working in departments is preferable to working in executiveagencies, which have a lower proportion of policy work time, andthat most senior officials would end up working in departments aftera reform.

The bureau-shaping perspective generates hypotheses about theperformance of executive agencies. Instead of being a better organi-sational form than the traditional department, the perspective suggests detrimental effects on economy, and some aspects of pro-ductive efficiency and effectiveness. The perspective assumes thatofficials working in executive agencies follow the best strategy avail-able to them which, as they cannot affect the distribution of tasksbetween departments and executive agencies, is to try to maximisethe level of the budget and associated output from the level origi-nally passed to the executive agency. The budget maximising strat-egy is suggested by Niskanen (1971) as a primary strategy for allofficials and is one of the strategies open to officials as an alternativeto bureau-shaping in the Dunleavy model (1991, pp. 193–4). Officialsin executive agencies’ concern with maximising budget/output doesnot necessarily imply that they are self-regarding, they may desirehigher budgets and outputs for a host of reasons including a com-mitment to the programme of activity they are involved in. The per-spective assumes only that they seek to maximise budget and outputand are likely to be able to push up these levels beyond that allocatedto them at the time of setting up the executive agency.

In suggesting budget maximising outcomes, Niskanen posited abilateral monopoly between a sponsoring legislature acting as thesole buyer of the good or service on behalf of the public and a bureau


acting as the only supplier. In market exchange, neither party to theexchange can exploit the other if they have perfect informationabout the product and exchange is voluntary. However, in conditionsof bureaucratic supply, officials can exploit the sponsor for their ownbenefit. In his model, the sponsor gives a budget to the bureau basedon the perceived public benefit from output of the bureau. Publicbenefits are assumed to increase with quantity, but at a decreasingrate, shown by the downward sloping line B1 in Figure 2.4. Thebureau has costs associated with marginal output which increase withquantity, shown by the upward sloping line C in Figure 2.4. Themodel assumes an information asymmetry such that the cost sched-ule is known by staff in the bureau but is not known by the sponsor.The sponsor sees only the total output of the bureau and its total budget. This feature of the sponsor’s position means that the sponsorcannot tell whether the output is efficient, in the sense that, at themargin, the benefits of increasing output of the bureau exceed thecosts of producing that extra output.

Niskanen introduced the possibility that the sponsor’s demand forthe bureau’s output might be so inelastic that the marginal benefit ofextra output would fall to zero before all the net surplus from initialproduction had been exploited in inefficient over-production,shown by line B2 resulting in output Q2 in Figure 2.4. At this pointthe sponsor is satiated and the bureau cannot deliver additional out-put, giving negative marginal benefits which would be noticeablybad to the sponsor. However, for simplicity, the bureau-shaping


B1B2

Q0 Q*

Quantity of budget/output

Mar

gina

l cos

ts (

C)

and

bene

fits

(B1,

B2)

C

Q2

S D

S = surplusD = deficit

Figure 2.4 Niskanen’s budget/output maximising model

perspective assumes the former case in which the declining marginalbenefits are represented by B1.

The perspective suggests that the department–executive agencyrelationship is analogous to the sponsor–bureau relationship whendepartments pass on budget to non-trading agencies. Officials inexecutive agencies can maximise their budget and associated output,rather than being able to affect the distribution of activities betweenthe department and the executive agency. Each executive agencyexpands production and budget beyond the allocatively efficient, interms of marginal benefit for the public equalling marginal cost, atlevel Q0, to the level Q*. At the level Q*, all the surplus of produc-tion where benefit exceeds cost up to level Q0 (area S) is swallowedup by the deficit of overproduction where cost exceeds benefit fromlevel Q0 to Q* (area D). Following the Niskanen model, the executiveagency can drive up the output to this level by offering this level ofoutput to the department rather than the alternative of not produc-ing any output at all. The department will prefer to have oversupplyto the alternative of no output because the surplus from productionup to level Q0 has not been completely eroded by the net deficitfrom oversupply beyond the level Q0 until level Q* is reached.

This situation enables the executive agency to produce twice thelevel of output that would have been supplied had the bureau pro-duction instead been produced by a body operating in a market. In amarket structure, individual consumers would only consume outputif the marginal benefit were greater or equal to the marginal cost.Consumers’ demand on a unit by unit basis would result in an output of Q0. However, the change brought about by the creation ofan executive agency working is not a move from a market to a bureaubut from a departmental arrangement for handling the activity in-house to a department ‘contracting’ with an executive agency.With departmental provision, the minister is able directly to see theoperation of activities within the organisation and has informationbased on day to day oversight. The creation of a semi-detached non-trading agency creates an information asymmetry which officials inthe executive agency can exploit to expand budget and output.Whilst this change will worsen allocative efficiency, this form of effi-ciency is beyond the scope of the inquiry and it is the hypothesisthat there will be a worsening of the economy of executive agenciesthat is the key suggestion of the perspective. Individual executive


agencies’ administrative and programme costs will increase overtime. Central government systemic administrative and programmecosts will rise as executive agencies’ budget maximising strategiesproduce more activity for departments and central units to monitor,although they will not be able to prevent economy worsening.

Officials in both trading and non-trading agencies will expandbudgets and output. The budget maximising behaviour of officials intrading agencies will be constrained by the need to raise revenuefrom individual consumers rather than from a departmental sponsor.Consumers will determine that spending is commensurate withdemand for the services which may increase or decrease the econ-omy of trading agencies depending on whether the former depart-mental arrangement was supplying at a level desired by consumers.However, the perspective suggests that officials in executive agencieswill only push to have their body designated as a trading agencywhere they think that they have the opportunity to expand servicesto increase budgets beyond that they might get from a department.In cases where they think customers will not pay they will not havesuch an incentive so the trading form will only be adopted where itfacilitates budget expansion. The perspective suggests that seniorofficials in departments will not be greatly concerned about whetheran executive agency has the freedom to raise revenue, their primaryconcern having been achieved by passing on executive activity.

The bureau-shaping perspective suggests that the effectiveness ofindividual executive agencies will be improved because budget andoutput is expanded. These benefits will accrue so long as the bureaudoes not overproduce to such an extent that marginal benefits fromits activities are negative, beyond the point that B1 intersects the ori-gin in Figure 2.4. However, the expectations for central governmentsystemic effectiveness are not so positive. The bureau-shaping perspective suggests that executive agencies will be exclusively con-cerned with achieving their own objectives. Such a pattern of behav-iour is likely to damage central government systemic performance bybringing about substantial public sector externalities. These publicsector externalities arise where the positive or negative effects of an organisation’s activities on other bodies are not incorporated inthe organisation’s own performance systems (James, 2000, p. 331).The perspective suggests that staff in executive agencies will not takeaction that would be of benefit to other parts of central government


if does not directly contribute to achieving the individual executiveagencies’ objectives, and that negative externalities harming otherbodies will be ignored. In particular, the perspective expects publicsector externalities in two areas. First, in the relationship betweenexecutive agencies and departments, where executive agencies battleto increase the scale of their activities and pursue their narrow agen-das, hampering departments in their efforts to make better policyand harming the communication between the bodies necessary forpolicy implementation. Second, in the relationship between execu-tive agencies and other bodies with operational responsibilities, thenarrow focus of each executive agency will lead to negative external-ities and the possibilities for beneficial co-operation to be ignored.


Table 2.3 The bureau-shaping perspective on executive agencies

Aspect Hypotheses

Reform 1. Senior departmental officials maximise the proportionprocess and of policy work time and budget per senior official subject tooutcomes government politicians’ constraint on departmental

organisation with a shift in the constraint triggering reform2. Executive agency reform is a bureau-shaping strategy by

senior officials (from point A to B in Figure 2.3) in whichdepartments pass on executive activity to executive agencies and concentrate on policy work. Senior officials oppose inconsistent reforms, including becoming hands on managers in integrated departments (from point A to C in Figure 2.3). A substantial inward shift in the constraint triggers between half and three-quarters of departments’ operational activity to be passed on.

3. Senior officials work in departments and junior officials inexecutive agencies

Individual 1. Economy in administrative and programme costs executive is reduced through the outcome of budget maximising agencies’ behaviourperformance 2. Effectiveness is improved through expanded output

3. Productive efficiency is unchanged

Central 1. Economy in administrative and programme costs government is reduced through the outcome of budgetsystemic maximising behaviourperformance 2. Effectiveness worsens because of public sector externalities

3. Productive efficiency worsens because of public sectorexternalities

The perspective suggests that the consequence of letting managersdirect resources inside executive agencies will not be to improve working but to raise output at the cost of greater inputs, with anunchanged ratio of inputs to outputs. The effect of executive agencyworking on individual bodies’ productive efficiency will be neutral. At the level of systemic performance, productive efficiency will beharmed by negative public sector externalities brought about by thenarrowly focused behaviour of executive agencies as they engage inbudget/output maximising in their own activities. The main hypothe-ses of the bureau-shaping perspective are summarised in Table 2.3.

Conclusion

The public interest and bureau-shaping perspectives offer substan-tially different hypotheses about executive agencies, as summarised inTables 2.1 and 2.3. In explaining reform, the public interest perspec-tive suggests that government politicians were in command of allaspects of departments’ organisation and suggested the Next Stepsreform plan in the public interest. In contrast, the bureau-shapingperspective suggests that government politicians were limited to set-ting the broad constraint on senior officials. These senior officialsexercised their concern with the organisation of their departments interms of the proportion of policy work time and budget per seniorofficial. Whilst the public interest perspective suggests that the reformwas neutrally implemented by officials, the alternative perspectivesuggests executive agency reform as a bureau-shaping strategy bysenior officials who, in response to a shifted constraint, preferred thisreform to the alternative of becoming hands on managers in inte-grated departments. Both perspectives hypothesise that executiveagency reform is a substantial change, in the case of bureau-shapingsubject to politicians shifting their general attitude to the manage-ment of executive activity. However, the bureau-shaping hypothesisis broader than the public interest perspective, indicating a generalband in response to a shift in the constraint rather than a preciselevel. Whilst the public interest perspective suggests that senior offi-cials work in both departments and executive agencies, the outcomeof bureau-shaping is that senior officials predominantly end upworking in departments.


The public interest perspective suggests improvements in economy, productive efficiency and effectiveness at the level of indi-vidual executive agencies and central government systemic perfor-mance, including administrative cost savings for individual bodies ofat least 5 per cent. The perspective suggests that freedom to managefacilitates better use of resources and a clearer focus on the task athand. The accountability systems, including performance targets foroutputs, enables better information about executive agencies to begathered. Combined with personal accountability of the chief execu-tive, these characteristics incentivise delivery and improve performance. The bureau-shaping perspective is much more scepticalabout benefits brought about by executive agency structures. Whilstit suggests the structures will be neutral for individual executive agen-cies’ productive efficiency it suggests that economy of administrativeand programme costs will worsen because of budget maximisingbehaviour. Whilst extra output is likely to improve individual leveleffectiveness, public sector externalities between departments andexecutive agencies and between executive agencies and other deliverybodies will worsen systemic effectiveness and productive efficiency.


Part II

Executive Agencies in Practice

3The Process of Executive Agency Creation

The pubic interest and bureau-shaping perspectives have contrastinghypotheses about the process of executive agency reform, sum-marised in Table 3.1. Section 1 examines the initiation of the reform.Section 2 explores the implementation of reform, especially in thecase of the Benefits Agency. Section 3 begins an assessment of theoutcome of the reform by examining the location of senior civil servants in departments and executive agencies after the changes.

Section 1: Initiation of the reform

Both perspectives suggest that actors within central government werethe main people involved in putting organisational reform on theagenda and deciding on a course of action rather than externalactors. The users of central government services and taxpayers didappear to have concerns about the performance of these activities in the mid 1980s, with 51 per cent of British feeling that the civil service was ‘not well run’. Although this figure was lower than thecomparable figure of 62 per cent for local government it reflectedwidespread critical opinion about the competence of administration(Social and Community Planning Research, 1991). However, therewas no large scale movement to lobby for reform in general or for theuse of executive agencies.

The issue of central government organisation and its performancehad low electoral salience, limiting the potential external influence on reform through the election of members of the governing Con-servative Party which was in power at the time of the Next Steps report.

41


Table 3.1 Hypotheses about executive agency reform

Public interest Bureau-shaping

Initiation of the 1. Government politicians 1. Senior departmentalreform control all aspects of officials maximise the

departmental organisation proportion of policy work and suggest the executive time and budget per agency reform plan in senior officialthe public interest subject to politicians’

constraint on departmentalorganisation

Implementation 2. Central units co-ordinate 2. Executive agency reform,of the reform reform which is neutrally triggered by a shifted

implemented by constraint, is a bureau-departmental officials as shaping strategy bypart of a unified machine. senior officialsUse of executive (moving from point A toagencies increases B in Figure 2.3) in whichincreases to around departments pass on95 per cent of the executive activity toCivil Service within a executive agenciesfew years, with trading and concentrate on agencies used where policy work. Senior customer groups exist officials oppose

inconsistent reforms, including becoming hands on managers in integrated departments (moving from point A to C in Figure 2.3). Asubstantial inward shift in the constraint triggers between half and three quarters of executive activity to be passed on.

Outcome of the 3. All civil servants will 3. Senior officials work inreform have experience of departments and junior

working in executive officials in executive agencies which operate agencies according to the model

There was no mention of executive agencies in the Conservative PartyManifesto for the 1987 election which noted only a general intentionto improve public services, stating ‘We will press on with long-termmanagement reforms in order to improve public services and reducetheir cost’ (Conservative Party, 1987, p. 42). The Conservative PartyManifesto for the 1992 election mentioned the desire to continue setting up executive agencies only in passing as part of a wider pro-gramme ‘to raise quality and efficiency in government’ (ConservativeParty, 1992, p. 15). The operation of executive agencies was not oneof the main issues discussed in material issued by parties or in themedia during the run-up to the two elections held in 1992 and 1997(Butler and Kavanagh, 1992, pp. 155–210; 1997, pp. 133–85).

The two perspectives have contrasting views of the role of differentactors within central government. The public interest perspectivesuggests that executive politicians had a specific executive agencyreform in mind. In contrast, the bureau-shaping perspective suggestsonly that these politicians became more concerned about a generallack of management effort and that senior officials were key to thechoice of the specific executive agency reform. In the late 1980s,there were around 110 executive politicians in central govern-ment, located in departments and in central units, including theCabinet Office and the Treasury (Hood and James, 1997, pp. 178–83).Defining senior officials as civil servants of Grade 5 (AssistantSecretary) and above and their equivalents in the bands introducedin 1996, there were 3613 senior officials out of a total of 492,527non-industrial officials in 1988 (Treasury, 1988b; Cabinet Office,1995b). The main actors involved in the Next Steps reform werelocated in different parts of central government as summarised inFigure 3.1.

Some authors have suggested that material from the memoirs ofpoliticians demonstrates that ministers originated and drove theNext Steps report and the changes that followed. These authors arguethat ‘More broadly [than just Next Steps], it is widely accepted thatthe reforms which the public service underwent during the lastConservative administration were driven by politicians and notbureaucrats. The most senior ex-Conservative ministers’ memoirs doindeed pay clear testimony to this fact’ (Marsh et al., 2000, p. 470).They cite three former senior politicians’ autobiographies. The first,by Michael Heseltine, was published before the reform began. Whilst

The Process of Executive Agency Creation 43

Heseltine argued that ministers were key in initiating many policyinitiatives and suggested government needed to be more business-like, he did not present any blueprint for Next Steps (Heseltine, 1987,p. 470). The second, by Nigel Lawson, is scarcely more forthcoming:the only substantial reference to Next Steps is on two pages out of1036 (Lawson, 1993, pp. 391–2). The third is Margaret Thatcher’sautobiography which mentions Next Steps only as a short footnotein a book of 862 pages of text, noting efforts to improve the man-agement of government. Thatcher did not suggest that her ministersor party originated the reform or ways in which the model was sup-posed to bring about improvements (Thatcher, 1993, p. 49). If politi-cians saw their role as significant they certainly do not seem to haveadvertised the fact.

Politicians’ behaviour suggests that they were less concerned witha specific organisational model than with the rule of thumb that civil servants with big programme responsibilities should devotemore time to these activities and less time to policy work than those with small programme responsibilities. However, politicians inthe Treasury and Cabinet Office seem to have been more interestedin matters of organisation than departmental ministers. The limited


Next StepsReportTeam

(1986–87)

CabinetOffice

Prime Minister’sAdviser onEfficiency

NextSteps

ProjectTeam

(from 1988)

EfficiencyUnit

Head of ProjectTeam and OMCS

OMCS

Cabinet Secretary/Head of Civil Service

Agency candidates(some part of departments

others free-standing)

Departments containingministers and senior officials

(asked to draw upagency

candidates from 1988)

Prime Minister and Cabinet(adopted the 1988 report as policy)

Parliament(especiallycommittees)

HM Treasury(approved agency plans)

Accountability relations

Figure 3.1 Institutions and actors involved in bringing about Next Steps1986–92

Note: OMCS � Office of the Minister for the Civil Service.

material on organisational issues in ministers’ biographies is mostevident for ministers who served in these central departments andstudies of the activities of central units broadly support this assess-ment (Thain and Wright, 1995, pp. 72–5, 147; Parry et al., 1997).

There is some support for the view that politicians shifted theirassessment of the time that should be spent on the management ofexecutive activity duirng the 1980s. This change in attitude has beendescribed as the ‘rise of neo-Taylorism’ (Pollitt, 1993, p. 53), involv-ing widespread adoption of the opinion that better monitoring ofactivities and employees was necessary to improve efficiency usingmodern technology. Politicians’ attitudes seem to have been influ-enced by the growing body of findings from scrutinies conducted bythe Prime Minister’s Efficiency Unit that were widely reported toministers (I1, Senior Official, Cabinet Office). A senior minister,Michael Heseltine, commented in 1987 that ‘When the literacies of the Civil Service and the generalities of intention are turned intotargets which can be monitored and costed, when information isconveyed in columns instead of screeds, then objectives becomeclear and progress towards them becomes measurable and far morelikely’ (Pollitt, 1993, p. 58).

Further evidence for the shift in politicians’ general attitudetowards management was their support for other reforms rather thana specific executive agency reform. In the early 1980s, central gov-ernment politicians supported the idea of the Financial ManagementInitiative (FMI) which sought to improve management within exist-ing integrated departmental organisations. The FMI suggested anorganisational system for each department giving managers a clearview of their objectives and measures of how far they were achievingthem, well defined responsibility for making the best use of resources,and better information and training, especially information aboutcosts (Prime Minister and Chancellor of the Exchequer, 1982, p. 21).Whilst the FMI had some aspects in common with executive agencyreform, including better management information systems, it differedsubstantially from Next Steps because it did not suggest passing onresponsibility for executive tasks to separate bodies. Instead thereform involved senior civil servants taking more time directly tomanage the activities for which they were responsible. In terms of theorganisation of the department outlined in Figure 2.3 in Chapter 2,whilst Next Steps involved a move from A to B, FMI involved a move


from A to C. Politicians’ support for Next Steps and FMI suggests thatboth reforms were on the shifted constraint curve and both werebroadly consistent with their concerns.

The interests of senior officials did seem to be highly significant inbringing about the executive agency reform rather than a full imple-mentation of FMI within integrated departments. Senior officialswere recruited after university and had long careers in the service,making them a cohesive group with a common set of interests (Hecloand Wildavsky, 1981). In the 1980s, Permanent Secretaries, the mostsenior grade with the posts usually held by individuals in their fifties,had, on average, each worked in the civil service for thirty years(Theakston, 1995, p. 37). A survey found that 82 per cent of themhad been continually in the civil service since the age of twenty-seven (Barberis, 1996, p. 153).

There is some indication that senior officials had preferencesbroadly consistent with the bureau-shaping perspective. Civil ser-vants’ attitudes to the FMI suggested that they were happy to runorganisations with large budgets but at the same time reluctant to get involved in management. Sir Robert Armstrong, a former CabinetSecretary and Head of the Civil Service, commented on senior offi-cials’ continued interest in policy work despite the attempts ofreforms to increase the profile of management tasks (Hennessy, 1989,pp. 737–8). A Deputy Secretary complained that ‘Management is atiresome business, nobody goes into it unless they have to’ (Zifcak,1994, p. 40). Some interviewees for this study suggested that manycivil servants shared the, usually unspoken, view that their positionas policy-makers for government made the best use of their skills and resulted in the best form of professional policy-making for thecountry (I1 and I2, Senior Officials, Cabinet Office). Officials demon-strated a concern to maximise their policy work role five years afterNext Steps began. A survey of officials in the top three grades andfast-stream entrants at lower grades found that the interesting natureof strategic work was one of the aspects of their jobs which they likedthe most (Efficiency Unit, 1993, pp. 105–8).

A recent study appeared to find evidence conflicting with this viewof senior officials’ concern with maximising policy work time. Thestudy suggested ‘To the extent that a simple dichotomy between thepolicy advice and the management function is possible, some civilservants certainly appear to prefer policy work, others management


work and others a balance between the two’ (Marsh et al., 2000, p. 468). The study examined Permanent Secretaries’ work, splittingtheir tasks into managing the policy process, administering thedepartment, and devising and implementing specific policies. Theauthors concluded ‘In our view, borne out by our interviews, PSs[Permanent Secretaries] manage in two senses. They administer thedepartment and they manage the policy making process’ (Marsh et al., 2000, p. 480).

Several of the tasks described as ‘management’ work by Marsh et al. have the characteristics of tasks described as ‘policy’ work in thebureau-shaping perspective. They state that the administration of the department involves ‘managing senior staff and resources in thedepartment, being the department’s representative/spokespersoninside and outside the Westminster/Whitehall community and act-ing as department accounting officer’ (Marsh et al., 2000, p. 466).These activities involve proximity to political power sources in the‘Whitehall village’ and beyond and seem to exclude routine man-agement of non-senior staff and resources in the department. Theyfurther suggest officials’ role in ‘managing’ the policy process, ensur-ing ‘long term policy strategy is being effectively managed’, choosingofficials to deal with policy issues, ensuring progress on policy issuesand ‘attempting to prevent political problems resulting from ‘fail-ures’ in the policy process’ (Marsh et al., 2000, p. 466). These tasksappear to involve innovation, working in small staff units or dealingwith political power sources. The study did not find evidence of offi-cials claiming to like working in a large extended hierarchy ratherthan in small staff units, and working at the point of delivery, remotefrom political power sources. Such a finding would constitute moreof a puzzle for the bureau-shaping perspective.

The Marsh et al. study relied mainly on interview material, but this form of evidence may not be sufficient to settle the issue, to theextent that there appears to be interview evidence both for andagainst senior officials’ concern for policy work time. However, evi-dence of senior officials’ career choices provides further support forthe bureau-shaping hypothesis. There was a general disinclination toacquire management skills, and tendency for officials to move to jobsinvolving policy work. The Next Steps report team found that ‘Thebusiness of top managers in many departments is still dominated bythe policy and political support tasks’ and continued ‘Many people


commented to us that too few senior civil servants showed the qualities of leadership which would be expected from top managers inorganisations outside the Civil Service’ (Efficiency Unit, 1988, p. 24).A survey of the choices made by officials in their careers conductedduring the early 1990s remarked on a disinclination to acquire man-agement skills (Efficiency Unit, 1993, pp. 105–8). Three quarters of the Permanent Secretaries in 1980 and 1993 had worked in ministers’private offices involved in policy advice and ministerial support and such jobs were prized by potential ‘high flyers’ (Theakston, 1995,pp. 37–43).

Evidence about senior officials’ concern to have as large an alloca-tion of budget and associated activity as possible is less comprehen-sive than that for their liking of policy work time. Much of theresearch in this area focuses on the outcomes of budget change ratherthan distinguishing between attempts to increase budget and budgetoutcomes, which may or may not increase (for a survey see Blais andDion, 1991). However, research on the outcomes of budget changesin central government during the 1970s and 1980s provides partialevidence. During times of budgetary retrenchment, departmentalofficials tended to cut their own budgets last, suggesting that theirown budgets and programmes were of particular concern (Dunsireand Hood, 1989, pp. 98–111).

Studies of departmental officials’ relationships with the Treasury in the spending round provide evidence of their attempts to push upbudgets and programmes, even if these were not always successful.Officials in each department usually suggested expansion of theirprogrammes in periods of budgetary growth and attempted to protect them against cuts in periods of retrenchment (Heclo andWildavsky, 1981, pp. 88–103; Thain and Wright, 1995, pp. 142–98).Similarly, interviewees for this study noted that departmental officials generally proposed increases in their own programmes andassociated budgets (IA3, Senior Official, Cabinet Office).

Rather than the reform being developed by senior officials in thedepartments it was suggested by officials in the Efficiency Unit.These officials’ behaviour was not fully consistent with direct bureau-shaping expectations. The authors of the Next Steps report were amix of career civil service and outsiders, most were not working forthe civil service by the mid-1990s (I1 Senior Official, Cabinet Office).At the time of the report, the Unit was supervised by the Adviser to


the Prime Minister on Efficiency, Sir Robin Ibbs. He was not a careercivil servant, having spent much of his career in private industry atImperial Chemical Industries (ICI), and had little reason to embarkon a bureau-shaping strategy. However, the reform was a form ofindirect bureau-shaping because officials in the Unit had strongincentives to come up with new initiatives that would be successfullyimplemented because they gained individual career benefits in termsof enhanced promotion prospects. Officials in the Unit were policyentrepreneurs and the Unit was described as a ‘ginger group’ and a‘place for strategic thinking’ by officials who worked in it (I1, SeniorOfficial, Cabinet Office; I2, Senior Official, Cabinet Office).

The civil service members of the Unit had an interest in the organ-isation of the departments to which they might move later in theircareers, which gave them an indirect interest in departmental organ-isation. However, more fundamentally, the choice of reform propos-als was limited by the need to take into account the wishes of seniorofficials in the departments. These officials’ support was needed toimplement any reform successfully, because departments were thebodies that carried out plans for reform.

Efficiency Unit officials undertook extensive consultation withsenior civil servants in the departments about possible reformoptions in conducting their research for the Next Steps report. Visitswere undertaken to see different parts of the system in action. Whilstthe team spoke to staff at all levels, senior officials were particularlyinfluential. Interviews were conducted with twenty-six PermanentSecretaries or equivalents and twenty-six Grade 2s. Five group dis-cussions were held with thirty-five Grade 3 officials and seven groupdiscussions were held with fifty officials of Grades 5–7 (EfficiencyUnit, 1988, Annex C, pp. 33–4). The Grade 7s included ‘high flyers’just starting their careers who expected to be promoted quickly inthe following years. The influence of the anticipated reaction ofdepartmental officials to reform proposals was noted by an officialinvolved in the study (I16, Senior Official, Cabinet Office). Anotherofficial responsible for implementing the reform noted that ‘Its [theNext Steps Report] persuasiveness lay not in the presentation ofdetailed evidence but in its description of a state of affairs whichmany readers [civil servants] said that they recognised readily fromtheir own experience; and in an analysis with which most of theminstinctively agreed’ (Goldsworthy, 1991, p. 8).


The Efficiency Unit team observed from the history of reform thatchanges to increase senior officials’ involvement in hands-on man-agement were likely to be greeted with hostility and were unlikely tobe successfully implemented. Full implementation of the FMI reformto make senior officials hands-on managers would have involved amove from position A to C in Figure 2.3 in Chapter 2. Such a reformwould have involved officials giving up policy work time to protectbudget in response to the inward shift in politicians’ constraint.Instead, executive agencies marked a step change from FMI andinvolved delegation to bodies responsible for managing executiveactivity. An investigation of FMI found that ‘the continuing devotionof senior officials to advising ministers and developing policydoomed the FMI to fail as an exercise to turn senior civil servants intoresource managers and directors of executive operations… . …topmanagement systems [in FMI] were given only passing attention insenior echelons’ (Zifcak, 1994, p. 40). The Efficiency Unit concluded‘Our evidence suggests that Top Management Systems are seen ashaving more relevance in executive functions and in the regions,than in headquarters or policy functions’ (Efficiency Unit, 1988, p. 22). Officials were blinkered in the range of evidence they drew on and did not look very far into the performance consequences ofthe structures they proposed. The study only lasted for ninety daysover the period November 1986 to March 1997, costing just £50,000(Efficiency Unit, 1988, pp. 33–5). There was a considerable delaybefore publication and several drafts were produced and revisedbefore publication. The desire to produce a reform that was accept-able to senior officials rather than weighing up all available optionswas reflected in the way the report’s authors interpreted the ideasand experiences examined in the study.

Several sources of ideas for New Public Management (NPM) reformhave been identified, especially new institutional economics andbusiness managerialism (Hood, 1991). However, the Next Stepsreformers’ had a particular notion of what making government moremanagerial and ‘business-like’ entailed, influenced by an Anglo-American big business practice of multi-divisional firm organisation(James, 2001). Multi-divisional firms are conventionally interpretedas a way of running diverse activities under an overall umbrella formof organisation in the interests of shareholders. In such firms, thecentral headquarters delegate activities to distinct business units and


give managers in these units considerable freedom in the use ofresources. At the same time, the centre monitors the performance ofthe units on a few key, output focused, criteria. Individual businessunits and the overall firm have chief executives which locates respon-sibility for different parts of the firm to individuals (Charkham, 1994).The Next Steps report team examined this organisational model in a private company, ICI, and a public corporation, British Rail(Efficiency Unit, 1988, p. 35). At the time, ICI had multi-divisionalstructure and British Rail had just imported a similar model into thepublic sector by splitting its organisation into separate business units.The Cabinet Secretary and Head of the Civil Service drew explicit parallels between parts of the civil service and sections of firms likeICI when voicing his support for the Next Steps reform (Butler, 1991).

The adoption of Anglo-American business practices was a highlyselective view of what government becoming more business-likeentailed. The Anglo-American practices were very different from thosefound in other systems, notably Germany or Japan which are conven-tionally seen as offering very different business models. Whilst allthese systems made use of the multi-divisional form, there is tradi-tionally more emphasis on collegiate responsibility and longer termmeasures of performance in German and Japanese firms (Charkham,1994). Reading across the ideas to central government, the systemswould suggest less strict separation of executive activity from policywork and more informal assessment of performance rather than theuse of performance targets (James, 2001). However, these insights werenot pursued, instead the reformers appeared to be fixated with the sep-aration of tasks between departments and executive agencies and con-trol of these bodies through performance targets and chief executives.

Much of the language in the report was economistic, in the senseof reflecting a concern with the need to provide incentives for indi-viduals to perform their tasks. New institutional economics, particu-larly public choice, transaction cost and principal agent theory are allrelevant to executive agencies but their full implications were notexplored. The reformers appeared to adopt some of the concerns of‘principal–agent’ analysis. This analysis is concerned with the speci-fication of a contract for performing a task by an agent on behalf ofa principal in conditions where the agent has better informationthan the principal and parties are prone to opportunistic behaviour(Milgrom and Roberts, 1992, p. 214).


Principal–agent analysis suggests possible undesirable outcomes inconditions where contracts cannot link an agent’s reward directly tothe interests of the principal, especially where agents may not bewilling to bear the risk of his or her reward being linked to an out-come that may not be completely under the agent’s direct control.The outcomes of the inability to set a contract that protects both parties’ interests include adverse selection, where the best agents donot end up performing tasks for principals, and moral hazard, whereagents do not perform in the way desired by principals. A range oforganisational solutions are suggested to this problem includingforms of performance pay, different ownership structures and usingcompetition between agents to uncover information, for example,using comparison between agents’ performance of the same task(Milgrom and Roberts, 1992, pp. 214–41).

The framework between the department and agency resembles theprescriptions of principal–agent analysis in specifying a ‘contract’with incentives for performing tasks. The post of chief executiveincorporates incentives for good performance including performancerelated pay and continued tenure of the post related to performance.However, principal–agent analysis points to the importance of com-petition as a way for principals to uncover information and ensureagents act in their interests. Competition between executive agenciesand between executive agencies and other bodies are not core features of the model and the authors of Next Steps did not explorethe full prescriptions of the analysis.

The Next Steps authors were selective in their use of evidence aboutthe performance of structures similar to those that they proposed. Agovernment report had previously noted some of the difficulties with‘arms-length’ structures in central government, in particular the diffi-culty of clearly dividing responsibility for tasks between departmentsand such bodies. The author of this earlier report commented withsurprise that the Next Steps reformers ‘did not even mention the 1980report on non-departmental bodies’ or the concerns about this formof organisation (Pliatsky, 1992, pp. 557, 561). The use of nationalisedindustries was being markedly reduced in the period of executiveagency reform, in part, because of dissatisfaction with their perfor-mance. Privatisation involving transferring ownership from the pub-lic sector to a private firm occurred in telecoms (1984), gas (1986),water supply and sewerage (1989), electricity (1990–93) and railways


(1994). Around 800,000 employees moved from the public to the pri-vate sector in the 1980s (Pollitt and Bouckaert, 2000, p. 273). Just asnationalised industries were being dismantled, structures with somesimilarities were being created in the executive agency initiative.

The bureau-shaping perspective is not sufficient to explain whyexecutive agency reform was preferred as bureau-shaping strategyrather than certain alternative strategies. The potential bureau-shaping option of increasing the use of local authorities to delivercentral government services was not acceptable because, as indicatedin the Chapter 1, this form of delivery involves bodies which have,in broad terms, more autonomy than executive agencies. However,there is also evidence to suggest that, whilst ministers did not have aspecific form of organisation in mind, they would have vetoed theuse of local authorities because many of these bodies were controlled by opposition parties and because they lacked confidence in theircompetence (I16, Senior Official, Cabinet Office).

A potentially more viable alternative bureau-shaping option toexecutive agency reform was the option of increasing the use of pri-vate firms under contract. Such a form of delivery maintains seniorofficials’ policy control without the responsibility for hands on man-agement of executive activity. This shaping strategy was pursuedsimultaneously with executive agency reform, although to a muchmore limited extent. The 1988 report suggested the possibility of‘agencies being located outside of government’ and, whilst thischange did not take place, widespread contracting out of activity toprivate firms was evident during the period. The market testing orCompeting for Quality initiative was launched in 1991; 935 markettests of central government activity were conducted resulting in£1632 m of activity being given to private firms under contractbetween 1992 and 1995 (Efficiency Unit, 1996, p. 93). The total valueof contracts averaged just over 1 per cent of civil service administra-tive expenditure per year, although there were further contracts withprivate firms that were not part of this programme.

Consistent with the bureau-shaping hypothesis, there was very little market testing or contracting out of policy work. Instead, officeservices and estate management accounted for over 40 per cent of allthe competitions that were conducted. Other services providedunder contract included engineering and maintenance, printing andphotocopying (Efficiency Unit, 1996, p. 95). The lack of widespread


contracting out to private firms as an alternative to executive agencies appears to have occurred because of potential oppositionfrom unions and concerns about the lack of firms able to take on theactivities (I1, Senior Official, Cabinet Office). Unions consistentlystated their opposition and the potential for industrial unrest(Council of Civil Service Unions, 1996). The best way for senior offi-cials to be sure of passing on executive activity, and to avoid becom-ing hands on managers of integrated departments, was throughexecutive agency reform.

The bureau-shaping perspective is more consistent than the public interest perspective with the importance of senior officials inoriginating the executive agency reform. However, there was some diversity of interests within the group of senior officials which is notdirectly consistent with either perspective. The Treasury was a rivalto the Efficiency Unit as a source of ideas about how to organise andcontrol central government, for example, the major public expendi-ture survey of the 1960s drew heavily on the experience of seniorTreasury officials (Thain and Wright, 1995). Officials in different cen-tral units competed to propose reforms and were committed to par-ticular viewpoints about their desirability (I3, Senior Official, CabinetOffice). The reform also appeared to make the Treasury’s mission tocontrol public expenditure more difficult because there was someuncertainty about how the model was to be incorporated in existingpublic expenditure control systems (IA1, Senior Official, Treasury).Senior officials in the Treasury were important in the early stages of the reform because of their responsibility for administering budgetary and accounting systems in central government and couldhave delayed the changes. There were disputes between CabinetOffice and Treasury officials about the desirability of the reform in its early stages (Zifcak, 1994, pp. 81–4; O’Toole and Jordan, 1995, pp. 3–17). Theakston emphasised clashes between ‘managerial radi-cals’ and so-called Treasury consolidators in the early days of thereform (Theakston, 1995, pp. 131–3). Greer concurred, describingevents as a power struggle between the Cabinet Office units, espe-cially the Office of the Minister for the Civil Service and the Treasury(Greer, 1994, pp. 45–58). The Head of the Next Steps Project TeamKemp eventually resigned because of the conflicts, prompting him toattack his former colleagues and the reforms ‘There are welcomebreakthroughs but they do not add up to a revolution. At the centre,


in Whitehall, old attitudes and the old guard prevail. The only headsthat have rolled have been among the revolutionaries themselves’(O’Toole and Jordan, 1995, p. 9). Whilst some of this assessmentexpressed personal frustration, there were undoubtedly conflictsbetween parts of the service.

The Efficiency Unit staff eventually won the battle and Treasuryofficials relaxed their initial opposition. The executive agency modelwas made consistent with the traditional central government sys-tems of expenditure and accounting control. The Treasury enshrinedthese principles in a White Paper, The Financing and Accountability ofNext Steps Agencies, published in December 1989 and used by depart-ments in setting up executive agencies (Treasury, 1989). A similarlogic was eventually accepted in a wider review of Treasury expendi-ture control across central government (IA1 and IA2, Senior Officials,HM Treasury). Under the Treasury’s fundamental ExpenditureReview, the expenditure control system changed towards fewer, morestrategic, ‘bottom line’ controls on expenditure with more freedomfor departments to use expenditure as they wished within these overall limits (Parry et al., 1997).

Section 2: Implementation of the reform

The public interest perspective suggests that officials in departmentsacted under strong central direction to implement the reform. Incontrast, the bureau-shaping perspective stresses that the interests ofsenior officials in departments undertaking bureau-shaping strategieswere key to pushing through the changes. There was some centraldirection to the reform, with a Next Steps Project Team in theCabinet Office having formal responsibility for direction of thereform, as shown in Figure 3.1. These officials had a particular inter-est in fulfilling their mandate of implementing the reform. A formermember of the Team noted that the Team wanted a rolling pro-gramme of implementation in departments (Goldsworthy, 1991, p. 20). The first Project Manager aggressively promoted the estab-lishment of executive agencies. In evidence to a Treasury and CivilService Committee inquiry in 1989 he set an unofficial target that atleast twenty executive agencies would be established by April 1990.In June 1990 he stated that the Government expected that at least


half the Civil Service would be operating in executive agencies by theend of 1991 (Treasury and Civil Service Committee, 1990, p. 1).

Whilst superficially consistent with the public interest perspec-tive’s expectation of central direction of the reform, the Next StepsTeam’s staff were concerned about the reception of the reform bydepartmental officials. The Team initially consisted of the ProjectManager, Sir Peter Kemp (a former senior Treasury official) and threecivil service assistants. These officials had an interest in the organi-sation of departments in which they might later work. The Team wastoo small to carry out the work itself, never consisting of more thantwenty people. The vast bulk of the work of implementing thereform was done by officials in the departments. The Team pursuedits aims by producing guidance and acting as a central point for mon-itoring departments’ progress and encouraging them in the reform.There were regular meetings with departmental officials, intended to‘foster support for and a sense of purpose about Next Steps at a seniorlevel’, during implementation of the reform and, overall, the viewwas that ‘effective practical proposals for reform should originatefrom those who have responsibility for bringing them about, i.e.those working within the services concerned. Only in this way willthose who are responsible for making these changes feel that theyown them and are committed to them’ (Goldsworthy, 1991, p. 22).

The public interest perspective suggests a substantial growth inexecutive agencies from a low base to cover 95 per cent of the civilservice. Given the sizeable shift in politicians’ attitudes, the bureau-shaping perspective suggests a substantial shift with between halfand three quarters of executive activity being passed on to executiveagencies. A census of all bodies called ‘agencies’ set up under NextSteps reveals that 173 executive agencies were created in the period1988 to 2001. A list of all executive agencies by department, withtheir dates of establishment and type is given in Appendix 4. Allthese bodies had the two core features of the executive agencymodel, management freedoms and an accountability framework. Thenon-trading type was the most numerous with 154 agencies or 89 percent, compared to nineteen trading agencies or 11 per cent of thetotal. There was a similar proportionate split in staff numbers, with90 per cent of staff in non-trading agencies and 10 per cent in trad-ing agencies. The total number of executive agencies, number of newbodies created each year, and their relative importance in centralgovernment staffing is shown in Table 3.2.


Tabl

e 3.

2G

row

th i

n e

xecu

tive

age

nci

es a

nd

civ

il s

erva

nts

wor

kin

g in

th

ese

bod

ies

1988

to

2001

Yea

rTo

tal

nu

mb

erE

xec

uti

veC

ivil

ser

van

tsA

ll c

ivil

Per

cen

tage

of

exec

uti

veag

enci

esin

ex

ecu

tive

serv

ants

of

civi

lag

enci

escr

eate

dag

enci

es(t

ho

usa

nd

s)se

rvan

ts i

n(i

ncl

ud

ing

(th

ou

san

ds)

exec

uti

vetr

adin

g fu

nd

s)ag

enci

es (

%)

1988

33

(3)

658

0�

119

8910

7 (1

)9

569

119

9035

25 (

9)11

456

220

1991

5924

(1)

200

554

3619

9276

17 (

2)21

056

537

1993

9216

250

554

4519

9410

214

268

540

5019

9510

914

(1)

267

517

5219

9613

330

275

495

5619

9713

410

285

475

6019

9813

84

277

463

6019

9913

65

271

460

5920

0012

63

279

475

5920

0112

62

(2)

277

483

57

Sour

ce:

Ap

pen

dix

4 c

ensu

s of

all

exe

cuti

ve a

gen

cies

an

d C

ivil

Ser

vice

Sta

tist

ics

(fig

ure

s fo

r ci

vil

serv

ants

in

G

reat

Bri

tain

).

The growth in executive agencies was not enough to reach 95 percent of the whole service but was broadly consistent with the bureau-shaping perspective. In the first year of the reform, there were justten executive agencies with 9000 staff. There was then a period ofsubstantial growth, with the number rising to seventy-six executiveagencies and 210,000 staff by 1992. By 1998, there were 138 bodiesemploying 277,000 staff, amounting to 60 per cent of all civil ser-vants. In 2001, the total stood at 126 executive agencies with277,000 civil servants, or 57 per cent of the total, including nineteentrading agencies employing 29,000 staff. Trading agencies employed6 per cent of all civil servants. The total number of executive agen-cies created is higher than the number in existence at the end of theperiod because of mergers and the abolition or privatisation of somebodies.

The number of executive agencies created each year initiallypeaked at twenty-five in 1990. The creation of trading agencies gotoff to a faster start than non-trading agencies; four trading agencieswere created by 1989, with the number of new bodies peaking atnine in 1990. However, some of these bodies had trading status con-ferred on them only after being established as executive agencies.There was a second, higher, peak of newly established executiveagencies in 1996, with thirty new bodies. But this figure was, for the greater part, the result of sixteen new bodies in the Ministry ofDefence that were relatively small, most having fewer then 1000staff, and included re-launches of units designated as executive agen-cies in previous years. The number of new bodies trailed off after theend of 1997. Whilst the reform was petering out, fourteen executiveagencies were created between 1998 and 2001.

The pattern of executive agency creation over time was broadlysimilar in different departments, with most bodies in each depart-ment set up between 1990 and 1996. However, there were two ‘inno-vators’, the Department of Trade and Industry, with nearly all itseleven executive agencies set up between 1988 and 1991 and theDepartment of the Environment, which created its five initial execu-tive agencies from 1989 to 1992. There were no obvious laggards,although the Ministry of Defence, which eventually created the mostbodies, was relatively slow to start. Beginning with two in 1990 andsix in each of 1991 and 1992 the number of new agencies fell backto two in 1993 before rising, on average, to six a year between 1994


and 1999. Similarly, in the Northern Ireland group of departments,sixteen of its twenty-four executive agencies were created in theperiod between 1995 and 1998.

There was a territorial dimension to the creation of executive agen-cies. About seventy executive agencies dealt with parts of the UKrather than the whole area with these bodies covering only England,Scotland, Wales, Northern Ireland or some combination of theseareas (see Appendix 4). Whilst executive agencies in England, Walesand Scotland were staffed by civil servants, those in Northern Irelandwere staffed by members of the Northern Ireland Civil Service. Thispattern reflected long running fragmentation of the system ratherthan a new departure linked to executive agency reform.

Both the public interest and bureau-shaping perspectives charac-terise executive agency creation as a revolutionary change in theorganisation of central government. A variety of organisationalchanges were involved in the application of the term ‘agency’ as partof Next Steps and departments were not monolithic structures beforethe reform (Hogwood, 1997). Categorising the creation of executiveagencies based on the structures handling activities ‘before’ and‘after’ the move to executive agency status reveals the pattern ofchange. The main forms of executive agency creation are defined as

(1) Pure creation involving establishing an executive agency whenthe activity was previously substantially integrated into a depart-ment without a specific set of management freedoms and a specific accountability framework.

(2) Mitigated creation consisting of establishing an executive agencywhen the activity previously had a partially developed specific setof management freedoms or a specific accountability framework.

(3) Nominal creation consisting of establishing an executive agencywhen the activity had a specific set of management freedoms anda specific accountability framework.

(4) New activity creation consisting of setting up an executiveagency to perform an activity not previously conducted by central government.

The greatest proportion of changes between 1988 and 2001 weremitigated executive agency creation, as summarised in Table 3.3; 121 cases or 70 per cent were of this type. This type of change was


followed by nominal creation with twenty-seven cases or 16 per centtaking this form. Pure executive agency creation was evident intwenty-two cases or 13 per cent and new activity creation was theleast common form with only three cases or 2 per cent. There was no obvious pattern in the distribution of types of executive agencycreation over time, except that there were no cases of ‘pure’ creationin the first two years of the reform. Instead, in 1988 and 1989, thereform involved activities that were already associated with clearlydefined organisational arrangements focused on a particular set oftasks or were already separate bodies with external accountabilitysystems. Overall, Next Steps was a substantial, revolutionary, changerather than a nominal reform because the changes were either pureor mitigated in 83 per cent of the cases. However there was a differ-ence between non-trading agencies that were the result of nominalchange in only 13 per cent of cases and trading agencies, where 37 per cent of these bodies arose from nominal change. Over a thirdof activities put into trading agencies had organisational arrange-ments closely resembling the executive agency model prior to thereform.

The two perspectives focus on change rather than the initialendowment of executive activity handled by departments. Both perspectives suggest that departments pass on a substantial portionof whatever executive activity they were initially responsible formanaging to executive agencies. Departments with more executiveactivity to start with pass on more activity than those with less executive activity. In broad terms, the pattern of executive agencycreation was consistent with these expectations. The main exception


Table 3.3 Type of executive agency creation in non-trading andtrading agencies (with percentages)

Non-trading Trading Total

Mitigated 109 (71%) 12 (63%) 121 (70%)Pure 22 (14%) 0 (0%) 22 (13%)Nominal 20 (13%) 7 (37%) 27 (16%)New activity 3 (2%) 0 (0%) 3 (2%)Total 156 (100%) 19 (100%) 173 (100%)

Source: Appendix 4 and survey of the origins of activities handled by executive agencies.

to the characterisation of the reform as a substantial change were thenominal changes that occurred in 37 per cent of trading agencies. Thepressure for these changes mainly came from a combination of offi-cials in the Next Steps team wanting to chalk up some early successesin the reform initiative, and the enthusiasm of officials in potentialcandidates. This latter group of officials were keen to get more free-dom to raise their own revenues (I2, Senior Official, Cabinet Office).The case of the Benefits Agency illustrates the process of executiveagency creation for a non-trading agency that was the result of miti-gated change. Almost two-thirds of all executive agencies had thesecharacteristics as shown in Table 3.3. The Benefits Agency was set upas part of a process to split the old Department of Social Security (DSS)between a departmental Headquarters and seven executive agenciesas summarised in Table 3.4.

The DSS conducted reviews of its activities in the light of the NextSteps report. Senior officials’ responsibility for choosing candidatesand setting them up could have lead to delays if they had objectedto the executive agency reform. The reviews included considerationof ‘prior options’ for the activities before setting up an executiveagency. The options included abolishing the activity, privatisationthrough sale of assets and contracting out the service to a privateorganisation. The reviews were dominated by civil servants ratherthan the three different holders of the post of Secretary of State forSocial Security in the main period of change between 1988 and 1992.The most senior official, the Permanent Secretary, was particularlyinfluential in getting the process underway (I13, Senior Official,DSS). A review team led by senior officials produced the Agency StudyReport, sometimes called the Hickey Report after the official who ledthe team. The report examined proposals for separating the activitiesof paying benefits and collecting social security contributions fromthe rest of the Department.

The senior officials involved in the review made statements thatwere partly consistent with the public interest perspective. Theynoted that ‘It is government policy that the executive functions ofDepartments should be carried out by agencies operating within apolicy and resources framework and using the resources set by the Secretary of State’ (DSS, 1993, p. 3). The change was noted asbringing ‘delegation with clear lines of responsibility and account-ability, clearly defined targets and challenging performance targets,


published annual reports, making the business more open toscrutiny, greater emphasis on the business and customer service’(DSS, 1993, p. 3). The Benefits Agency commented on similar advan-tages, allowing the organisation to focus on the needs of customerswhilst working within a framework set by ministers (Benefits Agency,1991, p. 5). Despite these statements, the limited scope of the reviewwas consistent with bureau-shaping motivations. The two mainalternatives under consideration were the executive agency model or some form of contracting out. Contracting out would have beenconsistent with the bureau-shaping perspective, which suggests that


Table 3.4 Executive agencies in the DSS, date of establishment, former location of activity and type of creation

Executive Date Location of activity Type ofagency established before executive creation

agency creation

Resettlement May 1989 Units under MitigatedAgency control of DSS

Information April 1990 Information MitigatedTechnology Technology ServicesServices Agency Directorate within (ITSA) the DSS

Contributions April 1991 Branches of the DSS PureAgency

Benefits Agency April 1991 Regional Directorate and PureRegional Offices in the Social Security Operations Group within the DSS

Child Support April 1993 New activity New activityAgency

War Pensions April 1994 War pensions and PureAgency general benefits

sections, Social SecurityOperations Group in theDSS (Benefits Agencyfrom April 1991)

Appeals Service April 2000 Sections handling MitigatedAgency appeals within

the DSS

Source: Appendix 3.

senior officials wanted to pass on executive activity to other bodies,and inconsistent with the public interest perspective which suggeststhat the executive agency form was the only one desired by politi-cians. There was an extensive discussion of options for contractingout at a senior level at the time of setting up the Benefits Agency andagain in 1995 during the review of the Agency’s status (I15, SeniorOfficial, Benefits Agency). In the end, the option of contracting outwas not pursued. The rejection of this option was not because ofopposition from senior officials or politicians but because of theopposition of officials in candidates for contracting out (I14, SeniorOfficial, DSS). The possibility of managing in a ‘hands on’ way withinan integrated structure was taken much less seriously than the execu-tive agency and contracting out options. Instead, much of the discus-sion around the time of the review concerned how to split the tasksup between different executive agencies (I13, Senior Official, DSS).

The depth of senior DSS officials’ support for the executive agencyform of organisation is further demonstrated by the similarity betweenthe Next Steps proposals and their own reform proposals which they developed before the publication of the Efficiency Unit report. An internal departmental report, produced in 1988, recommended setting up structures to handle the implementation of social securitypolicy, separate from the policy core of the Department (Moodie et al.,1988). Both this proposal and Next Steps built on the existing regional structure of local offices delivering benefit which were already partiallyseparate from the policy core. The Benefits Agency brought togetherthe regions and central centres for paying out benefits, separate froma policy core of the Departmental Headquarters (DSS, 1993, pp. 26–9).Most senior officials in the Department expected to find themselveslocated in the DSS Headquarters after the reform and consequentlycarrying out favourably valued policy work tasks.

The officials working in potential executive agency candidatesappear to have had much less influence on the course of reform thansenior staff at the centre of the Department. The first chief executiveof the Benefits Agency, Michael Bichard, came from local govern-ment and was not part of the creation of the Agency. The unionswere sceptical about the benefits of the reform and the consequencesfor junior staff’s pay and conditions, but were in a weak position. TheCivil and Public Servants Association were not completely hostile tothe reform, but the National Union of Civil and Public Servants


opposed the change. The unions were consulted in the process of setting up executive agencies but were not key to their design(Goldsworthy, 1991, p. 16). However, partly because of the splitsbetween the unions, and partly because of their reduced influence inthe 1980s, the unions were not able to mount a campaign against theexecutive agency initiative (Greer, 1994, p. 54). The junior officials inthe DSS had more influence in blocking the alternative proposal ofshifting the administration of welfare payments to private firmsunder contract, which staff were more implacably opposed to, andsuggested industrial unrest might follow such a decision (I15, SeniorOfficial, Benefits Agency).

Some staff in candidates for executive agency status were able toinfluence the boundaries between different executive agencies. Theorganisational boundary created between the Benefits Agency andITSA was heavily influenced by officials in the InformationTechnology Directorate of the DSS. The Directorate was staffed by IT professionals who worked in a labour market which cut across thestandard civil service career structures, more frequently moving inand out of government than other officials (Greer, 1994, p. 33).These specialist staff wanted more freedom to develop their specificskills and higher pay, seeing their own executive agency as a way toachieve this (I1, Senior Official, Cabinet Office). The IT staff werepart of the report team that recommended the establishment of theseparate agency (DSS, 1993, pp. 30–1). ITSA took on IT systems andhad a role in contracting out IT functions. The Agency was able tomake use of freedoms to group staff by function and respond moreflexibly in pay and conditions to recruit and keep specialist IT staff(ITSA, 1996, p. 20).

Section 3: Outcome of the reform

According to the bureau-shaping perspective, senior officials in theDSS and other departments predominantly end up working in the policy core of departments rather than in executive agencies. In contrast, the public interest perspective expects a more even split of senior officials across departments and executive agencies.The location of senior officials (Grades 1–5) immediately after the most intensive period of executive agency reform is summarised inTable 3.5. In three-quarters of departments, over 70 per cent of senior


staff worked in the non-agency parent section of the department.The Department for Education and the Welsh Office had 98 per centof senior staff in the parent department, although these departmentsdid not have many executive agencies. The four departments withpercentages ranging from 55 to 61 per cent had lower percentagesthan the others in large part because of several Grade 5 officials work-ing in executive agencies. This grade was the least senior of the seniorgrades; excluding these officials to focus on the most senior officials


Table 3.5 Senior staff (Grades 1–5) in parent departments in 1995

Department Per cent in Number in Number inparent parent executive

department department agencies

Department for Education 98 67 1Welsh Office 98 66 1Scottish Office 93 162 13Foreign and 91 43 4Commonwealth Office/Overseas Development Administration

Department of Employment 88 183 24National Heritage Department 85 23 4Department of Health 82 192 41Department of the Environment 82 164 36Ministry of Defence 80 338 85Ministry of Agriculture, 77 117 34Fisheries and Food (MAFF)

Department of Trade and 75 231 76Industry

Treasury 73 103 39Cabinet Office 71 102 41Northern Ireland departments 67 18 9Department of Transport 61 95 61 (inc. 53

at Grade 5)Lord Chancellor’s 58 42 30 (inc. 25Department (LCD) at Grade 5)

Home Office 57 113 87 (inc. 69at Grade 5)

DSS 55 66 55 (inc. 42at Grade 5)

Source: Cabinet Office, 1995b.

raises the proportion working in parent departments to well overthree quarters in each of these cases.

In the DSS, 55 per cent of 121 senior officials worked in theHeadquarters. In contrast, 3 per cent of the DSS’s 89,075 junior staffworked in the DSS Headquarters with the rest in executive agencies(Cabinet Office, 1995b, pp. 21–3). Of the senior DSS staff who workedin executive agencies, over three-quarters were at the lowest seniorlevel of Grade 5. Many of these officials would have expected to moveto the Headquarters or to other departments in Whitehall later intheir careers, which was not the expectation of most of the less seniorofficials in executive agencies. The level of budget per senior officialin the Headquarters was about one-tenth of that in the DSS prior tothe creation of executive agencies (James, 1995b, p. 622). However,this straight comparison overstates senior officials’ involvement inmanagement work before the reform because there was some partialorganisational separation of operational activities within DSS prior tothe changes.

Overall, about 10 per cent of executive agencies were involved inactivities with the characteristics of policy work, contrary to both theperspectives’ hypotheses. However, these bodies were overwhelm-ingly small research and development organisations handling spe-cialist, often scientific, tasks which generalist civil servants did nottraditionally undertake. A more important test of the perspectives’expectations is whether or not politically sensitive tasks requiringfrequent interaction with ministers were put into executive agencies.The location of this activity is particularly important for evaluatingthe bureau-shaping perspective because it was traditionally a key partof policy work for senior officials. A rough indication of political sen-sitivity is given by the number of Written Parliamentary Questionsasked by Members of Parliament and directed at ministers relating to the activity. Policy work activities of political sensitivity have a higher number of Written Parliamentary Questions than routineexecutive activities despite MPs having some interest in routine executive activities on behalf of their constituents.

The political salience of activities handled by executive agencies asa group was much lower than activities handled by departments.Taking 1995 as a census year, questions relating to activities handledby executive agencies were just 7 per cent of the total for theGovernment. The results for individual executive agencies reflected


low saliency. The greatest number of questions to executive agencieswere for HM Prison Service with 613 questions, followed by the ChildSupport Agency with 231 and the Benefits Agency with 217. Therewere thirty-two executive agencies with fewer than this number butmore than fifteen questions; the remaining sixty bodies had betweenzero and fifteen questions (Hogwood et al., 1997, pp. 109–12).Adjusting Written Questions to take account of expenditure empha-sises the relatively low salience of activities in executive agencies. Therewere 0.3 questions per £m of running costs for executive agencies as agroup, in contrast to eleven questions per £m of running costs for par-ent departments. No executive agency had more than eight questionsper £m and over three-quarters had fewer than two questions per £m.

Senior officials in most departments were able to maintain theirpolicy work position in the 1980s and 1990s. Very little policy workwas contracted out to outside providers. Whilst increasing use wasmade of special advisers and outside sources of policy advice therewere usually only one or two special advisers per minister, a verysmall percentage of those working on policy in departments. Thecharacter of civil servants interaction with ministers may havechanged from the traditional way of working, including ministersnot necessarily accepting the options put to them by departmentalofficials (Richards and Smith, 2002, pp. 220–3). However, despite thischange, policy work remained a non-routine activity and involvedworking close to political power sources.

Senior civil servants’ interest in undertaking policy work in depart-ments was maintained by the establishment of the Senior CivilService consisting of officials of Grade 5 and above and by the limitsplaced on the scope for using executive agencies. A White Paper com-mented that ‘The Government does not, however, envisage extend-ing the formal establishment of agencies into areas of the CivilService primarily concerned with policy … For day to day support forMinisters on policy matters, policy divisions on existing lines willcontinue to be the preferred model, adapted by departments as nec-essary to deal with changing workloads’ (Prime Minister et al., 1994,p. 13). In the new Senior Civil Service, executive agency chief execu-tives formed less than 5 per cent of the total, the vast bulk of seniorofficials being located in departments. In the mid 1990s, there werereductions in senior staff through delayering as part of departmentalSenior Management Reviews (Cabinet Office, 1995a, p. 17). However,


there was relative stability in senior civil service numbers over theperiod as a whole. In 1988, before Next Steps, there were 3613 officials of the senior grades out of a total of 580,000 (Cabinet Office,1989, 2000, p. 44). In 1999, membership of the Senior Civil Servicestood at 3570 out of 460,000 officials. Over the period, whilst thetotal number of officials declined by 20 per cent, the number ofsenior officials declined by just 1 per cent.

Conclusion

Both perspectives are consistent with the dominant role of actorswithin, rather than outside, central government in the process of theNext Steps reform. However, the process was more consistent withthe bureau-shaping perspective than the public interest perspective.Senior officials responded to politicians’ pressure for managementimprovement through executive agency creation rather than politi-cians having a plan to create executive agencies with civil servantsimplementing it neutrally. Consistent with the bureau-shaping per-spective, civil servants preferred policy work time to managementwork and there is more limited evidence that they preferred higherbudgets to lower budgets. In terms of Figure 2.3 in Chapter 2, oncethe initial status quo position A was no longer possible given the new constraint, the Next Steps reform to bring about position B waschosen rather than the alternative reform of moving to point C bydeveloping the FMI to make senior officials hands-on managers ofexecutive activity in integrated departments.

The constellation of interests that brought about the reform wasmore complicated than that suggested by either perspective. The ini-tiation of the reform by officials in the central Efficiency Unit was aform of indirect bureau-shaping. These officials wanted their plans tobe implemented and needed a proposal that would have the supportof senior departmental officials. Contrary to the public interest per-spective, the report team drew on a limited range of ideas rather thanembarking on a thorough review of possible structures for centralgovernment. It had a particularly Anglo-American view of businessmodels and did not follow many of the prescriptions of principal–agent analysis. The report team were selective in their choice of evidence about the performance of similar arms-length bodies, espe-cially non-departmental bodies and nationalised industries, whenthis evidence did not provide support for their proposals.


The reform was a revolutionary change, the reform proposals weretaken up and implemented by departments; there was a massiveexpansion in the use of executive agencies with 173 bodies createdbetween 1988 and 2001. In 2001, 57 per cent of civil servants workedin 126 executive agencies. The change was not as great as the publicinterest perspective’s hypothesis that 95 per cent of central govern-ment activities would be put in executive agencies but consistentwith the half to three quarters of operational activity suggested bythe bureau-shaping perspective. The changes were substantial, only16 per cent of cases of executive agency creation were nominalchanges, although 37 per cent of trading agencies were the result ofnominal change. Officials in the Next Steps project team had aninterest in presenting the reform as a success and this partiallyexplains nominal changes, especially in the early days of the reform.

The practice of reform was generally consistent with the bureau-shaping perspective’s hypotheses. Whilst the details of the reformvaried between executive agencies, senior officials in departmentsappear to have consistently attempted to pass on executive activityrather than policy work. In the case of the Benefits Agency, the NextSteps proposals were similar to senior DSS officials’ own plans forchange. Most junior staff in sections of work proposed for executiveagency status were unenthusiastic about the change, and stronglyopposed the alternative of contracting out work to private firms.However, a group of IT professionals supported the introduction ofthe separate ITSA which enabled them to pursue their specialistcareers across the public/private sector boundary.

Consistent with the bureau-shaping perspective, most senior offi-cials expected to end up working in parent departments rather thanexecutive agencies and this outcome occurred. On average, 70 percent of senior officials were located in departmental headquartersrather than in executive agencies after the reform. Whilst some exec-utive agencies handled non-routine work, many of these bodies wereresearch and development organisations requiring specialist scien-tific skills and did not handle the policy work with high politicalsalience that was traditionally handled by generalist senior officials.An indicator of the political salience of activities based on WrittenParliamentary Questions reveals that activities in departments produced eleven questions per £m of administrative costs and amuch lower figure of 0.3 questions per £m of administrative costs foractivities in executive agencies.


70

4The Practice of Executive Agency Working

A large number of bodies were set up with the two main features ofthe executive agency model. This chapter continues the examinationof the outcome of the reform by assessing whether executive agen-cies’ procedures and general working practices were consistent withthe model, or whether the changes were superficial. If the changeswere superficial then these bodies’ performance does not revealmuch about executive agencies and the performance expectations ofthe two perspectives are much less relevant. Section 1 assesses thesemi-detached organisational arrangements between executive agen-cies and their parent departments. Section 2 examines the account-ability framework. Section 3 explores executive agencies’ freedomfrom central government-wide regulations and accountability toother parts of central government, Parliament, customers and usersof their services.

Section 1: Semi-detached organisation

The executive agency model consists of management freedom andfreedom from ad hoc intervention by departmental staff and minis-ters with resources under the overall direction of a chief executiverecruited through open competition. Trading agencies have furtherfreedom to raise revenue and expand activities covered by feescharged to customers. There was an initial discussion of an annualquasi-contractual arrangement between departments and executiveagencies to set out the objectives and structure in which they oper-ated (Talbot, 1996, p. 43). However, it was quickly decided instead to

The Practice of Executive Agency Working 71

have three, and later five, year period ‘Framework Documents’ which ‘publicly set out the respective roles of ministers … and thoseof agencies’, the ‘boundaries between policy and service deliveryfunctions’ and the ‘financial and personnel management flexibilities’of each agency. The aim was to make reporting lines ‘transparent’,clearly setting out ‘accountabilities within government and toParliament’ (Treasury, 1989, p. 17). Framework Documents were notlegal documents but were described as containing durable elementsdescribing the relationship between the executive agency and otherbodies (Goldsworthy, 1991, p. 26). The Documents were supposed tobe flexible enough to fit the different situations of each agency, fol-lowing the view of the Next Steps report that ‘the detailed nature ofthe relationship between a department and an agency will vary withthe job to be done or the service to be delivered’ (Efficiency Unit,1988, pp. 9–10).

There were disputes in drawing boundaries between strategicframeworks and day to day operations. These problems were greatestin ‘mainstream’ agencies that were key to achieving the main aims ofdepartments including the Benefits Agency, Employment Service andPrison Service. In contrast, for trading agencies and other executiveagencies more ‘peripheral’ to departments’ concerns there were fewerdisputes, examples of these bodies included Her Majesty’s StationeryOffice (HMSO) and the Central Office of Information (EfficiencyUnit, 1991). The relationships between departments and executiveagencies became more clearly defined over time as the details of thedivision were worked out (Trosa, 1994, pp. 11–13). The Governmentacknowledged that the departmental side of the relationship was stillin flux during the early 1990s. ‘At the same time [as agency develop-ment], the development of a more structured relationship betweenthose at the centres of departments and those in agencies has increas-ingly focused attention on the nature of headquarters tasks and howthey can best be discharged, whether in policy making or the strate-gic management of agencies’ (Prime Minister et al., 1994, p. 13).However, despite disputes at the boundary, blocks of work were successfully parcelled out to executive agencies along with substan-tial management freedoms.

Executive agencies’ freedoms from departmental regulations tooktime to develop; there was usually no big bang at the launch of thebody (Efficiency Unit, 1991, pp. 6–11; Trosa, 1994, p. 29). However,


by the end of the 1990s, 80 per cent of chief executives declaredthemselves broadly content with the delegations granted to them,although nearly 40 per cent proposed additional delegations beyondthose that they had been given (Cabinet Office, 2001). The day today running of some central government-wide systems was handledby departments. Executive agencies generally submitted requests forresources to their departments rather than being able directly toargue their case with the Treasury. This arrangement was a cause ofsome frustration amongst chief executives when they disagreed withtheir departments about the level of resources they required (IA1,Senior Official, Treasury; IA4, Senior Official, Treasury). The Treasurygranted departments end of year flexibility to allow them to carryforward an under-spend to the following year’s budget. Chief execu-tives expressed dissatisfaction with the way some departments oper-ated end of year flexibility because they did not always pass on thisfreedom to executive agencies. In other cases, departments wereunclear whether they would allow their executive agencies to havethis freedom, making it difficult for these bodies to plan accordingly(Office of Public Service Reform and HM Treasury, 2002, pp. 40–1).

The degree to which executive agencies were successful in devel-oping their own patterns of working, distinct from their depart-ments, varied. Some chief executives expressed frustration at lack offlexibility in departmental pay and grading systems and the need toseek approval for management decisions (Efficiency Unit, 1991;Trosa, 1994). Distinctive systems of working practices appeared mostdeveloped in trading agencies (Mellon, 1993, pp. 25–31). For exam-ple, HMSO developed its own procedures to be more focused on itscustomers including a revised staff appraisal system (Dopson, 1993,pp. 17–23). Flexible working focused on the task at hand was lessextensive in many non-trading agencies but greater than under theold systems (Trosa, 1994, pp. 28–51).

The post of chief executive was adopted for all executive agencies.Open competition was used in 65 per cent of appointments between1988 and 1995, although all the appointments made in the first twoyears were civil servants (Horton and Jones, 1996). In 1998, aroundtwo-thirds of chief executives in post had been appointed throughopen competition and one-fifth of appointments had been made tooutside (i.e. non-civil service or military) candidates, as summarisedin Figure 4.1. Many chief executives developed more ‘managerial’


styles of working than traditional civil service practices, with moreconcern about operational practice, human resources and a focus onthe organisation’s aims (Horton and Jones, 1996). A survey found thatchief executives felt that they were able to devote the bulk of theirattention to developing their own internal management systemsrather than cross-departmental management issues (Price Waterhouse,1993a,b). For example, the chief executive of the Employment Serviceargued that he had been able to develop a clearer ‘vision’ of the busi-ness and styles of working focused on these aims which would nothave been possible within the old departmental arrangement (Fogden,1993, pp. 10–11).

Ministers and other staff in parent departments did not alwaysrefrain from ad hoc, day to day, intervention in executive agencies.Centres of departments were found to intervene to provide assurance‘not just on the adequacy of control systems, but also on the activi-ties of the Agency’ (Efficiency Unit, 1991, p. 19). Ministers did notrespect the frameworks as barriers to intervention when they felt nec-essary and there was generally more intervention in ‘mainstream’non-trading agencies than in trading agencies. The degree of contactbetween ministers and chief executives varied, from once a year dis-cussions about the strategic plan for the executive agency to regularweekly meetings in some ‘mainstream’ bodies where ministers hadparticular concerns (Trosa, 1994, p. 11).

The executive agency model did not set out a mechanism for‘enforcing’ the freedoms given to these bodies. Ad hoc intervention

61

31

39

7

Open: internalOpen: externalClosed: internalActing

138 chiefexecutives

Figure 4.1 Appointment of chief executives in post in 1998

Source: Cabinet Office (1998).


by departmental staff and ministers was usually accepted by chiefexecutives because of the difficulty of separating tasks fully. The tra-ditional ‘Whitehall village’, noted by Heclo and Wildavsky (1981),involving informal negotiation between all the key actors involvedin running central government over a prolonged period was found inexecutive agency working. The ways of the village encouragedFramework Documents to be seen as flexible statements of the tasksof executive agencies, subject to revision, and being overriddenwhere necessary. Chief executives recruited from outside the civil ser-vice expressed greater frustration with intervention in the running oftheir organisations than traditional civil servants who were moreused to informal ways of working (Mellon, 1993, pp. 26–7).

The case of the Benefits Agency illustrates the operation of semi-detached organisational structure. The formal organisational separa-tion between the Headquarters and the Benefits Agency is shown inFigure 4.2. The purpose of the Agency was stated as ‘the adminis-tration of social security benefits and other services’ (BenefitsAgency, 1991, p. 4) and, subsequently, as a mission to ‘support theGovernment in establishing a modern welfare state … by helping to create and deliver an active and modern social security service.The service will encourage independence and pay the right money tothe right person at the right time, all the time’ (Benefits Agency,

From April 1996 the number of territories inthe Benefits Agency was reduced from three to two

Department ofSocial SecurityHeadquarters– Secretary of State and ministers– Permanent Secretary and senior policy staff

Benefits AgencyChief ExecutiveCentral Services– Finance– Personnel– Policy and Planning– Medical Services– Change Programme

TerritorialDirectorates– Scotland and North– Southern– Wales and Central England (each containing local districts)

Performancetargets

Budget

Performanceand otherinformation

Services tobenefitclaimants

Working withother parts ofgovernment

Figure 4.2 Benefits Agency and DSS Headquarters’ organisational chart

Source: Benefits Agency (1992, pp. 33–5, 1997, p. 7).


1999a, p. 5). At the end of the 1990s, the Agency administeredtwenty-four social security benefits, provided information and adviceabout benefits, handled claims, appeals and reviews, and made pay-ments and recoveries for benefits (Benefits Agency, 1999a, p. 31). TheAgency was supervised by the Department of Social Security (DSS)Headquarters which was responsible for overall social security policyand implementation, supporting ‘ministers on policy and legislation’and providing ‘corporate management of the Department’ andsupervising ‘the allocation and accounting for resources, targets andperformance measurement’ (DSS, 1999a, p. 1).

The Framework Document suggested that the Department shouldconsult the Agency before making policy changes and that the chiefexecutive could propose changes, but that the Department took the lead on policy matters. The Agency should ‘contribute to theDepartment’s development and evaluation activities; provide infor-mation on the operational implications of current and alternativeprogramme characteristics and provide, to an appropriate level ofquality, such data as Ministers and the Permanent Secretary mayrequire to support the monitoring, evaluation and development ofpolicy and the monitoring and forecasting of benefits expenditure’(Benefits Agency, 1991, p. 5 and 7). The Framework Documentemphasised that the Secretary of State, Permanent Secretary and DSS Headquarters (DSS HQ) did not ordinarily intervene in the dayto day running of the agency but reserved the right to do so, statingthat the Agency was ‘acting on behalf of, and in accordance with anydirections, where appropriate, of the Secretary of State for SocialSecurity’ (Benefits Agency, 1991, p. 4). The Agency was required to‘maintain the capacity to respond quickly to Ministerial direction’(Benefits Agency, 1991, p. 5). However, staff in the Benefits Agencycomplained of unjustified interference by departmental staff beyondthe terms of the Framework Document. There were frequent and in-depth meetings between senior Benefits Agency staff and the DSS,including meetings with ministers (I10, Senior Official, BenefitsAgency). However, the vast bulk of staff in the Agency were involvedin routine administration of benefit payments in the network of localoffices and had little contact with the Headquarters.

The chief executive post in the Benefits Agency operated in a waythat was generally consistent with the model. The chief executiveswere recruited by open competition on the basis of suitability for


running the Agency and were on non-permanent contracts (BenefitsAgency, 1991, p. 15; 1995b, p. 7). The first chief executive came fromlocal government, the second from the private sector, whilst thethird was from the civil service. The chief executives had responsi-bility for setting management structure, creation and retrenchmentof posts, promotion and pay arrangements and general personnelissues outside of the most senior posts. The first chief executive’sleadership style, including the use of logos for the Agency and directcommunication with junior staff through newsletters, was noted asparticularly important in establishing a distinctive identity for thebody (I13, Senior Official, Benefits Agency).

Section 2: The accountability framework

The second part of the executive agency model consists of anaccountability framework including performance targets and a chiefexecutive with personal accountability for performance. The systemsof reporting performance to departments varied with annual, quar-terly and other reports, regular meetings and ad hoc interaction.Each executive agency produced annual reports and corporate andbusiness plans. Some officials were designated as contacts betweendepartments and each executive agency, called ‘Fraser’ figures afterthe author of the report that suggested them (Efficiency Unit, 1991,p. 7). Such figures were broadly consistent with the thrust of theexecutive agency model, enhancing the quasi-contractual arrange-ment between two semi-detached bodies. However, the figures some-times acted as representatives for executive agencies, potentiallyplacing him or her ‘between’ the department and the executiveagency. Occupants of the role found it difficult to carry out rigorousoversight and at the same time argue the executive agency’s case indisputes with the department (Trosa, 1994, pp. 13–16).

Executive agency performance targets were used extensively tomeasure performance and exercise influence by specifying desiredlevels of performance. All executive agencies had some system of per-formance targets, with the number varying from just one to overtwenty, with a mean of seven targets per body. Targets were usuallyset by ministers who formally approved targets proposed by the exec-utive agencies. Departments often had to rely on the executiveagency for ideas about suitable targets and target levels, or sometimes

operated rules of thumb such as always asking for more than theexecutive agency proposed (Trosa, 1994, pp. 61–2).

There were initiatives to improve the advice available to ministersin supervising executive agencies, including the use of MinisterialAdvisory Boards (MABs). MABs included both executive members ofagency management, civil servants from the departments and non-executive directors, often with a private sector background. MABswere found to be more effective than Fraser figures where a widerange of skills were needed to set targets for an executive agency(Next Steps Project Team, 1995b, pp. 16–21). A report suggested thatMABs should be extended to all departments and the proportion ofoutsiders should be increased to one-third, partly to increase theirindependence from executive agency management (Trosa, 1994, pp. 19–22). In 1997, about half of executive agencies were covered by MABs.

There were large gaps in the performance target regimes of manyexecutive agencies, with outcomes affecting valued aims and objec-tives not having corresponding targets. In some trading agencies,overall goals were expressed as a single target. For example, the RoyalMint was tasked with making a percentage return on assets. In part,this limited target regime reflected the alternative accountabilitymechanism of customers paying for services. However, it was oftendifficult to encapsulate departmental interest in activities using tar-gets, particularly in non-trading agencies. A survey of executive agen-cies in the early 1990s that few had clear targets for outputs. Mosttargets were for financial performance and inputs (Pendlebury et al.,1992, p. 47). Even where output targets existed for production, thesedid not always map on to outcomes. A study in the mid-1990s exam-ined all executive agency targets. It found that 59 per cent related to outputs (usually units of different goods or services produced in a year), 17 per cent related to efficiency (often a measure of unit cost of output), 12 per cent referred to processes (often the achieve-ment of an administrative task) 9 per cent related to inputs (usuallybudget or staff levels) and only about 1 per cent directly outcome targets (Talbot, 1996). Taking the Home Office and Department ofSocial Security’s executive agencies as examples, the study found that 47 per cent of these bodies’ aims and objectives were not covered by performance targets and a further 31 per cent were only partiallycovered (Talbot, 1996). Not only were targets limited, they changed



from year to year, hampering a comparison of the performance overtime. There was a 70 per cent turnover of executive agencies’ targetsin the first half of the 1990s (Talbot, 1996, pp. 49–51). The level of tar-gets set by departments did not seem to be consistent with attempt-ing to challenge the organisations to improve their performance. In 2000, 16 per cent of targets were set at a lower level than in theprevious year (Cabinet Office, 2001).

The chief executives had higher powered incentives than those tra-ditionally given to officials. The appointments were on the basis ofpersonal contracts and were initially for a period of three years, thenon a rolling basis. The remuneration consisted of a basic salary,which before 1996 corresponded to a Grade in the civil service andafterwards was linked to a pay band level in the Senior Civil Service,and performance related pay. The chief executives were more pub-licly visible than traditional officials, their pictures were printed inannual reports and in the annual reviews of executive agencies pub-lished by the Cabinet Office. They had a personal responsibility forreporting performance, explaining why performance outcomes tookthe course that they did. In some cases the chief executive reportedregularly and directly to the minister, but in many other instances heor she reported to the Permanent Secretary or lower ranking officialsin the department.

Chief executives continued tenure was, at least nominally, relatedto acceptable performance. A report found that chief executives were‘acutely aware of their visible personal responsibility and account-ability for the success of their Agencies’ (Efficiency Unit, 1991, p. 5).However, few chief executives were publicly criticised by ministers orremoved for poor performance. Chief executives’ tenure was nor-mally considerably greater than three years. In 1996, only one-thirdof executive agencies had experienced a change in chief executiveand changes were often associated with an alteration of the status ofthe organisation (Horton and Jones, 1996). Even when an executiveagency performed poorly, it was difficult to separate out the contri-bution of chief executives’ action from other factors. There wasuncertainty about whether poor executive agency performance wassufficient to invoke blame or whether the chief executive’s personalcontribution had to be clearly identified.

Disputes between departments and executive agencies aboutresponsibility for outcomes were normally discussed internally.

The main exceptions to this generalisation involved chief executivesappointed from outside of the civil service. The most notable publicdispute occurred over the removal of the chief executive of the PrisonService, Derek Lewis, in 1995. The case was unusual, but illustrates thedifficulty of identifying responsibility for outcomes. Lewis wasremoved from his post by the Home Secretary after the Learmontreport criticised his removal of the Governor of Parkhurst prison fol-lowing the escape of some prisoners. He claimed that he had beenforced to remove the governor by the Home Office and had been sub-ject to day to day interference from ministers despite the supposedoperational discretion of the Prison Service. Interventions includedsummoning Lewis to the Home Office to discuss operational matterson a daily basis. Lewis issued a writ for unfair dismissal that was even-tually settled out of court (Lewis, 1997; Barker, 1998; Polidano, 1999).

In the case of the Benefits Agency, the DSS held the Agencyaccountable for performance against targets set by the Secretary ofState. A steering group advised the Secretary of State annually on setting targets and sought advice from both the department and the Agency (NAO, 1998b, p. 37). The targets were mainly quantita-tive targets for outputs with a few milestone targets related to theachievement of specific procedures. The Agency provided theDepartment with performance information each month which fedinto three meetings each year between the Permanent Secretary andthe chief executive in which the factors affecting performance out-comes were discussed (NAO, 1998b, pp. 39–40). The external targetswere supplemented by internal management targets within theAgency, cascaded from the Secretary of State’s targets to cover moredetailed aspects of performance. The internal targets were set by thechief executive and approved by the Secretary of State and developedthe logic of the target regime throughout the organisation.

The Secretary of State’s targets did not cover some important partsof the agency’s work, for example, the accuracy of the administrationof Retirement Pensions (NAO, 1998a, p. 56). There were few mea-sures of whether benefit went to the right people and the systemslargely ignored the issues of productive efficiency and the quality ofcustomer service (NAO, 1998a, p. 39). The Agency had an indicatorof cost of administration per pound of benefit expenditure, whichcovered all spending of the agency on benefits. However, this mea-sure left aside whether the right people were getting the benefit,


which should have been a consideration in a fuller indicator of productive efficiency. There were frequent changes to performancetargets. Some targets, for example, the Secretary of State’s targets forspeed of payment, were downgraded after 1992/93. There weremovements in target coverage, definitions and the introduction ofnew targets related to new tasks and occasional mis-reporting of per-formance against targets (NAO, 1998a, p. 16). The Agency and thechief executive were only held responsible for performance againsttargets to a limited degree. The Agency was vulnerable to changes inits environment affecting its performance, such as the level ofunemployment, and it was difficult to separate out the contributionof Agency staff to performance (I10, Senior Official, Benefits Agency).The Department’s assessment of the Agency and the chief executive’sperformance was much broader and more nuanced than simple performance against targets.

Section 3: Relationships between executive agencies and other bodies

Beyond the parent department, the executive agency model consistsof freedom from some central government-wide regulations includ-ing controls over expenditure, pay and grading. These freedoms arecombined with accountability of the executive agency and its chiefexecutive to other parts of government, Parliament, the users servicesand citizens. Trading agencies have additional accountability to paying customers who provide a large part of their revenues.

Executive agencies were incorporated within the existing expendi-ture control framework with freedoms granted in some areas. Non-trading agencies’ financial flexibility varied. Some were financed bya ‘gross’ budget passed on from the department and were required topay any surpluses into general receipts, for example, the WelshHistoric Monuments Agency (CADW). Some were supply financedbut operated on a net running cost regime where they covered thecosts of providing certain specified services by charging, for example,the Vehicle Certification Agency. The freedom to have net controlrather than gross was widespread with about a quarter of executiveagencies recovering some or all of their costs from charging (CabinetOffice, 2001). There were freedoms in use of budget, 85 per cent of a sample of executive agencies in 1991 had complete or partial


freedom to vire money between parts of their budget; and only 25 per cent were not allowed some form of carry forward of unspentbudget from one year to the next (Pendlebury et al., 1992, pp. 42–4).In the case of the Benefits Agency, the body was financed by moneythat was approved by Parliament under gross control. The Agencyorganised requests for expenditure through the Department ratherthan dealing directly with the Treasury (Deakin and Parry, 2000, p. 132). The Agency was able to carry over 0.5 per cent of runningcosts each year, could carry out capital projects up to a limit, andswitch between sub-headings of running and capital costs within a limit (Benefits Agency, 1991, p. 14).

The system of expenditure control was not well integrated with thesystems of performance targets. There was very little development ofperformance budgeting, for example, to analyse how changes inexpenditure affected performance against target (Talbot, 1996, p. 43).Central policies on expenditure were imposed across governmentwhich limited executive agencies’ discretion. There were blanketpublic sector pay settlements and periodic freezes on running costs.A survey of chief executives found dissatisfaction with the limits cen-tral policies placed on their ability to pursue their objectives (PriceWaterhouse, 1993). The greatest financial freedoms were given totrading agencies, in part by extending the 1973 Trading Fund Act.Trading agencies were, by and large, able to use their authority toraise income to alter services and to plough back cash generated byproviding services into developing their businesses. Trading agencieswere allowed to invest from accumulated cash balances to improvetheir services and chief executives were generally content with theoperation of these delegations (Office of Public Service Reform andHM Treasury, 2002, p. 42).

Executive agencies were incorporated in the system of accountingfor expenditure, which placed responsibility on officials designatedas ‘accounting officers’. Accounting officers were conventionally the heads of departments, usually designated permanent secretary,and reported to the Public Accounts Committee (PAC) of the Houseof Commons for the money voted by Parliament to the department.Where an executive agency had the status of a department the chiefexecutive was designated as a full accounting officer. In other cases,the chief executive was generally appointed as an additionalaccounting officer for the executive agency with the Permanent


Secretary of the department remaining as the overall accounting officer. In the Benefits Agency, the Permanent Secretary of the DSSwas Accounting Officer for the DSS and for administrative spendingand payments from the National Insurance Fund. The PermanentSecretary appointed the chief executive as Agency AccountingOfficer for this expenditure where it was handled by the Agency(Benefits Agency, 1991, p. 10; 1995b, p. 4). The Treasury appointedthe chief executive Additional Accounting Officer in accounting forthe Social Security Benefits Vote and the Social Fund Vote (BenefitsAgency, 1991, p. 10; 1995b, p. 5). The role as Additional AccountingOfficer gave the chief executive a more distinct set of responsibilitiestied to the Agency than if he or she was just an agency accountingofficer (I13, Senior Official, Benefits Agency).

Flexibilities were granted to executive agencies in employment systems, although staff remained civil servants or military staff ontransferring to these bodies. The Treasury and Cabinet Office granteddelegated authority for recruitment which was revocable anddepended on departmental and central bodies’ support. The devel-opment of recruitment and pay bargaining systems specific to eachexecutive agency challenged the traditional conception of the civilservice as a single employer (Clifford, 2000, pp. 18–20). A study ofthirty-four executive agencies, in 1991, found that 82 per cent hadpartial freedom over recruitment. Specific systems of pay took severalyears to develop; almost three years after the initiative began, HMSOwas the first executive agency to develop a separate pay system. In1994, thirty-two executive agencies had the freedom to develop theirown pay and grading policies; this freedom was extended to all thebodies in 1996. The Benefits Agency gained responsibility for payarrangements in April 1994. The chief executive carried out negotia-tions with employees in pay bargaining which gave the body moreflexibility. However, pay bargaining was carried out in a way consis-tent with the DSS Personnel Guiding Principles and the system wassemi-detached rather than completely separate from that elsewherein the department (DSS, 1995).

There were several centrally directed public sector reform initia-tives that limited the ability of executive agencies to develop theirown systems. In procurement of goods and services, 85 per cent oftwenty-six executive agencies surveyed in 1991 had partial freedomto develop their own systems, although restrictions often applied to



items over £0.25 m of expenditure (Pendlebury et al., 1992, p. 45).However, the Treasury introduced the Competing for Quality pro-gramme in 1991. The programme required departments to proposeactivities to be ‘market tested’ against alternative provision by privatecontractors. Many small executive agencies felt driven to conduct testsfor activities in which the process incurred costs that were out of proportion with potential benefits (Efficiency Unit, 1996, pp. 29–31).Even in large agencies like the Benefits Agency, managers felt underpressure to meet central targets for testing rather than undertaking thereviews to improve working within the Agency (I13, Senior Official,Benefits Agency).

A second central initiative was the Citizen’s Charter programmewhich developed from 1991 and was led by the Cabinet Office. Itincluded a requirement for bodies that dealt with the public todemonstrate performance against standards of service. The CabinetOffice required Framework Documents and plans to reflect Charterstatements about the quality of service (Chancellor of the Duchy ofLancaster, 1994, p. ii). The Citizen’s Charter Unit was described as animportant influence on standard setting in the Benefits Agencybecause the standards had to be approved by the Unit (I13, SeniorOfficial, Benefits Agency). The scheme presented less conflict withthe discretion afforded to executive agencies than Competing forQuality because the standards were largely integrated within existingperformance targets (IA9, Senior Official, Cabinet Office). However,according to senior officials responsible for implementing NextSteps, the fact that these central initiatives were being interpreted asan unnecessary burden reflected a change in attitudes. Chief execu-tives were questioning the relevance of central initiatives for ‘their’distinct organisations rather than accepting them as part and parcelof running bodies that were part of central government (I2, SeniorOfficial, Cabinet Office).

In the executive agency model of accountability to Parliament, thebodies were to be held accountable for routine ‘operational’ matters,especially through performance targets, with the chief executive having personal accountability. Ministers were to be accountable forthe overall strategic framework, giving information to Parliament,being open to questioning, taking responsibility for their actionsand, on occasion, resigning for their contribution to poor strategic performance. However, the convention of ministers taking personal

responsibility for the poor performance of their departments andresigning did not operate for central government in general duringthe period since 1988. Ministers did not resign for problems traceableto their role in setting the framework in which executive agencieswere supposed to operate (Dowding, 1995, pp. 161–78). As well asthe dispute between a minister and the chief executive of the PrisonService in 1995 which ended in the removal of the chief executive,the chief executive of the Child Support Agency, Ross Hepplewhite,complained that her Agency had been set an impossible task by itsdepartment, which explained its very poor performance. Despite crit-icism of the framework by Parliament, no minister resigned. Instead,the chief executive of the Agency was removed (Harlow, 1999).

Executive agencies and their chief executives were held to accountfor operational matters, in the sense of providing information abouttheir performance as a unit. There was limited sporadic monitoringof executive agencies’ ‘operational’ tasks by Parliament. Chief execu-tives answered MPs questions relating to operational matters, refer-ring them to ministers if they raised matters that were deemed torelate to policy. For most executive agencies, the standard practicewas for the minister to see a copy of the chief executive’s answer but not to intervene (Hogwood et al., 1997, p. 20). However, MPsrequests for information and demands for action generally related toparticular concerns raised by constituents and occasional flurries ofinterest in cases of very poor performance (Evans, 1994).

Executive agency officials answered questions at Select Committeehearings on the same basis as departmental officials, subject to the‘Osmotherly Rules’ which stated that civil servants answered ques-tions on behalf of ministers and that ministers made and interpretedpolicy (Natzler and Silk, 1994, p. 73; Dowding, 1995, p. 156). In casesof very poor performance, chief executives were sometimes ques-tioned as accounting officers by the PAC. The PAC drew on theNational Audit Office (NAO) ‘financial audits’ of expenditure and‘value for money’ studies. Each executive agency’s accounts was sep-arately checked by the NAO, reflecting the emphasis on corporateresponsibility in the model. It was often difficult to separate outresponsibility for tasks between the executive agency and depart-ment and chief executives often appeared alongside their depart-mental Permanent Secretary. Other committees touched on the work of executive agencies where there were perceived to be severe


problems of performance (Hogwood et al., 1997, p. 15). An examplewas the Social Security Committee’s interest in the Child SupportAgency which was the subject of several reports because of excep-tionally poor performance (e.g. Social Security Committee, 1993,1994, 1995).

In the case of the Benefits Agency, MPs questions about routinematters were handled by the chief executive. Despite a few initialcomplaints from MPs, there was not great disquiet about the systemdespite the Agency having the third highest volume of questions ofany executive agency (Hogwood et al., 1997, pp. 12–14). TheAgency’s accounts were assessed by the NAO as a distinct unit. TheNAO found problems with both financial accounting and value formoney. The head of the NAO, the Comptroller and Auditor General,qualified his opinion on financial accounts of the Agency for thetenth successive year in 1997–98, because of errors (PAC, 2000, Secs 1–3). However, the difficulty of fully separating tasks betweenthe DSS Headquarters and the Agency was reflected in accountabilityto Parliament, especially to parliamentary committees. The Agencywas asked to present its case jointly with the Department, for exam-ple, in giving evidence to the PAC about the high level of fraud and error in the programme (PAC, 2000).

Executive agencies developed relationships with the users and customers of their services. Their annual reports were usually pub-licly available although corporate plans were sometimes confiden-tial for commercial reasons (Prime Minister and Minister for the Civil Service, and the Minister of State, Privy Council Office, 1990,pp. 7–8). The trading agencies were responsive to customers who pur-chased their services, seeking to explore their demands by variousmeans, including surveys and attempting to respond by altering thegoods and services they provided (Dopson, 1993). A further group ofexecutive agencies had users of their services within government.These bodies increasingly made use of user surveys, written feedbackand complaint monitoring in attempting to respond to user groups(Office of Public Service and HM Treasury, 2002, p. 67).

An indication of orientation towards customers is provided by the extent to which executive agencies were covered by the CharterMark scheme. Under the scheme, public bodies were encouraged toapply for external assessment and recognition of services focused onthe customer. Those given the award had to demonstrate that they


published explicit standards, were open to customers, offered choice,operated in a courteous and helpful way, redressed complaints andprovided value for money. They also had to show improvement inthese areas over time. Charter Mark was mostly about processesrather than outcomes and the majority of activities handled by exec-utive agencies were not covered by awards. However, sections of fifteen executive agencies received the Charter Mark by 1994, risingto forty-nine in 1995, seventy-one in 1996 and ninety-one in 1997(Chancellor of the Duchy of Lancaster, 1994, pp. ii–iii; Next StepsTeam, 1996, p. 17; 1997, p. 12; 1998, p. 14). These figures suggestthat an increasing number of bodies had a customer focus in theperiod.

The Benefits Agency developed accountability systems focused onthe users of its services. The Agency published annual reports, busi-ness and corporate plans and other documents setting out its strate-gies and publicly stating its aspirations for performance to clientsand the general public (Benefits Agency, 1991, p. 8; 1995a, p. 5).Officials in the Agency suggested that it was keen to try and projecta ‘corporate’ image as a distinct organisation to people claiming ben-efit including refurbishing offices in a distinctive style, giving staffname badges and producing promotional material (I4, Official,Benefits Agency; I13, Senior Official, Benefits Agency). The Agencyproduced an internal survey of customer satisfaction, monitoredcomplaints and undertook mystery shopper exercises. However, thesurvey of satisfaction was dropped, in part, because of worries aboutwhether it reflected performance or was being manipulated by staffto meet the target. The Chief Executive commented that ‘I felt thatmaybe the national survey was designed to try to get a figure of 86 per cent [the target]’ rather than reflecting actual satisfaction(Social Security Committee, 1998, Q174).

Conclusion

There was considerable variety in the operation of executive agenciesand the working practices within central government were far fromuniform prior to the reform. However, freedom from departmentalsystems was established for most executive agencies and depart-ments’ intervention contrary to the spirit of semi-detached operationwas sporadic. Intervention was more significant in bodies that were



‘mainstream’ to a department’s aims than in trading agencies or otherperipheral executive agencies. Freedom from central government-wide systems were granted in expenditure, pay and grading and otherareas; trading agencies’ right to raise revenues from customer groupswas, by and large, respected. All executive agencies had chief execu-tives with non-permanent contracts; about two-thirds of appoint-ments were made subject to open competition, with 20 per cent ofappointed chief executives coming from outside the civil service.

Accountability systems were varied, although all executive agen-cies established under Next Steps had performance targets. Tradingagencies typically had a few targets, often including a target for mak-ing a return on the capital they employed, whilst non-trading agen-cies had a greater number of more varied targets for different outputs.However, the targets were not comprehensive of the outcomes ofexecutive agencies’ activities and they changed over time. Instead,performance against targets was one part of departments’ assessmentof executive agencies. Chief executives took personal accountability,in terms of providing information about performance, but their rolein taking personal responsibility for operational performance wasmore limited. There were high-profile disputes about responsibility,such as occurred between the head of the Prison Service and theHome Office, although such disputes were unusual. In general, chiefexecutives were not held directly responsible for their organisation’sperformance against targets, although this performance was used inassessing performance pay.

Executive agencies provided more information on their perfor-mance as a distinct unit to Parliament, users of services and cus-tomers than under the old arrangements. MPs addressed questionson operational matters to chief executives, although the perfor-mance matters that were of interest to committees usually requiredboth departmental and executive agency staff to attend. The con-vention of ministerial resignation for poor performance did not operate in general during the period and no ministers resignedbecause of failings in the strategic framework set for an executiveagency. The development of accountability to users of executiveagencies’ services was restricted with the greatest strides made intrading agencies.

88

5The Performance of IndividualExecutive Agencies

The two perspectives’ hypotheses about the performance of individualexecutive agencies are summarised in Table 5.1. Section 1 examinesthe economy of executive agencies. Section 2 looks at their productiveefficiency and effectiveness. Section 3 assesses the effect of executiveagency structures on the performance of the Benefits Agency.

Section 1: Economy

The public interest perspective suggests that savings of at least 5 percent of individual executive agencies’ administrative costs occur

Table 5.1 Hypotheses about the performance of individual executive agencies


Economy 1. Economy improves with at 1. Economy in administrativeleast 5 per cent savings in and programme costs isadministrative costs and reduced through the savings in programme costs outcome of budget

maximising behaviour

Effectiveness 2. Effectiveness improves 2. Effectiveness is improvedwith better achievement of through expanded outputobjectives

Productive 3. Productive efficiency 3. Productive efficiency isefficiency improves including at least unchanged

5 per cent savings inadministrative costs andsavings in programme costs

The Performance of Individual Executive Agencies 89

within a few years of executive agencies being set up. In contrast, thebureau-shaping perspective hypothesis is for a rise in administrativecosts. An analysis of executive agencies from 1995/96 to 1997/98reveals changes in real terms (in 1996/97 prices) administrative costs.This period was in the mature phase of the reform, when most bodieshad been created, and is not so long such that reorganisations obscuretrends in the use of resources. There were seventy-two executive agen-cies that were broadly stable in their organisation in the period. Intotal, as a group, their costs fell 4.6 per cent from £9127m in 1995/96to £9079m in 1997/98. In these terms, the executive agency initiativedid improve the economy of these activities by nearly as much as the5 per cent expected by the public interest perspective. However, the mean score was a small administrative cost rise of 1.7 per cent. This increase reflected modest rises in several bodies and a few largeincreases, the largest being the Forensic Science Service which increasedits budget by 75.5 per cent. The largest budget fall was in theInsolvency Service which fell by 25 per cent. However, the medianchange in administrative budget was a fall of 4.4 per cent, almost ashigh as the 5 per cent expected by the public interest perspective.

There was variation in the change in administrative costs betweentrading and non-trading agencies. As shown in the box-plot in Figure 5.1, the median score for non-trading agencies was a 5.3 per cent

1260

100

80

60

40

20

0

–20

–40

TradingNon-tradingN =

Per

cen

t

Figure 5.1 Box-plot of percentage change in real terms administrative costsof seventy-two non-trading and trading agencies 1995/96 to 1997/98 (withnumber of cases)


reduction. In contrast, trading agencies had a median increase of 8.2 per cent. An assessment of the direction of changes in adminis-trative costs reveals that there were reductions in costs in 57 per centof executive agencies. However, whilst 60 per cent of non-tradingagencies experienced falls, there were falls in only 42 per cent of trading agencies, as summarised in Table 5.2. In terms of both thedirection and size of the cost changes, non-trading agencies weremore consistent with the public interest perspective and tradingagencies were more consistent with the bureau-shaping perspective.

In the Benefits Agency, real terms (in 1996/97 prices) adminis-trative cost fell by £114.5 m or 4.4 per cent between 1995/96 and1997/98, nearly the amount expected by the public interest perspec-tive. However, Figure 5.2 illustrates that in the period 1991/92 to2001/02 the picture was mixed for both real terms administrativeand real terms total costs (including both administrative and pro-gramme costs). Real terms administrative costs rose from £2018 m inthe first year to peak at £2608 m in 1995/96 before falling back to£2194 m in 2000/01, an increase of 9 per cent over the period as awhole. Staffing showed a similar pattern to administrative spending,as shown in Figure 5.2. Staffing rose from 68,000 in 1991/92 to peak at 77,435 in 1996/97 before falling back to 70,000 in 2000/01,a rise of 3 per cent over the period as a whole. The figures for 1996/97and subsequent years included several thousand staff transferredfrom the Employment Service to the Agency to handle work associ-ated with the Job Seekers Allowance benefit. Adjusting the figures forstaffing and expenditure to remove the effect of this transfer suggeststhat the increases in the early 1990s were partly balanced out by improvements in economy in the latter half of the 1990s.

Table 5.2 Non-trading and trading agencies with realterms administrative cost rises and falls between 1995/96and 1997/98 (with percentages)

Cost fall Cost rise Total

Non-trading 36 (60%) 24 (40%) 60 (100%)Trading 5 (42%) 7 (58%) 12 (100%)Total 41 (57%) 31 (43%) 72 (100%)

Source: Survey of seventy-two executive agencies.


The process of setting up the Benefits Agency itself created adminis-trative costs. It is difficult to put an exact figure on this expenditurebecause the activities were not accounted for separately and manydifferent sections were involved in the process. However, the cost ofestablishing the Agency was ‘considerable’ according to interviewees,running to ‘several million pounds’ (I14, Senior Official, Departmentof Social Security (DSS); I15, Senior Official, Benefits Agency).However, these costs seem unlikely to have been a sizeable propor-tion of the total administrative costs for the period since 1991.

The total expenditure of the Agency was dominated by expendi-ture on the programme of social security welfare payments, withadministrative costs forming only around 3 per cent of the total, asshown in Figure 5.2. Real terms (in 1996/97 prices) total spendingrose from £65,858 m in 1991/92 to peak at £78,246 m in 1997/98before falling back to £74,194 m in 2000/01, a rise of 13 per cent overthe period as a whole. On this indicator of economy, the Agency did not perform well, although the rate of growth fell from the mid-1990s.

Comparing the costs of activities before and after the creation ofthe Benefits Agency indicates whether executive agency working was

Year

£ m

at 1

996/

97 p

rices

and

sta

ff nu

mbe

rs

total costs admin costs total staff

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

100,000

80,000

60,000

40,000

20,000

0

Figure 5.2 Economy of Benefits Agency 1992–2001


associated with a different pattern of budget change to the old system. Such a comparison is hampered by difficulties in identifyingseparate budget figures for the relevant parts of the old Departmentof Social Security. However, in the three years before the creation ofthe Agency, the Department’s total real administrative costs, whichwere predominantly the costs of the activity that became the Agency,rose from £2500 m to £2684 m, an increase of £184 m or 7.4 per cent(Chancellor of the Exchequer, 1994, p. 41). There was a rise of about15 per cent in real terms total social security payments in the sameperiod (Chancellor of the Exchequer, 1998, p. 57). In the first threeyears of the post-agency system, there was a rise in real terms admin-istrative costs of about 16 per cent and real terms total spending of 19 per cent. These figures suggest no improvement in the rate ofincrease. The change in costs both before and after the creation of the Agency was closely and positively related to workload linkedto the economic cycle. Executive agency working, with the exceptionof a period in the mid to late 1990s, was not associated with a consistent improvement in economy.

Section 2: Productive efficiency and effectiveness

According to the public interest perspective, executive agenciesimprove productive efficiency and effectiveness in their use ofresources to achieve their aims as set out in their FrameworkDocuments, annual reports, business plans and corporate plans. Inparticular, improvements in productive efficiency arise from reduc-tions in administrative costs of at least 5 per cent and reductions inprogramme costs. In contrast, the bureau-shaping perspective sug-gests that productive efficiency is unchanged, although effectivenessimproves because of expanded output. There is a chronic lack of governmental or academic research on productive efficiency andeffectiveness in central government, including executive agencies(Trosa, 1994; Pollitt, 1995; Talbot, 1998). However, some indicationis provided by bringing together information from a fragmented setof sources and through undertaking new surveys.

An indicator of effectiveness is provided by the National AuditOffice’s (NAO) audit of financial regularity in executive agencies.These surveys assessed whether money provided by Parliament wasused for the purposes intended, with due regard for propriety and


whether the financial statements of executive agencies were freefrom material misstatement, whether caused by fraud, error or someother irregularity. Where these problems were seen as significant or‘material’ auditors ‘qualified’ the account. If the problem was lesssevere, auditors produced reports referring to the problem andattached it to their ‘audit statement’ about the account. The resultsof audits of executive agencies and central government’s accountsfrom 1993–99 are given in Appendix 5. On average, 6 per cent ofexecutive agencies’ accounts were either qualified or had a reportattached to them, with between one and two executive agencies’accounts qualified each year, and with between none and six execu-tive agencies having reports attached to their audits. These resultswere not greatly different from the pre-executive agency arrange-ments which exhibited a similarly low level of qualification ofaccounts (Hood et al., 1999, pp. 87–9). However, because this infor-mation focused on financial control, it provides only a partial pictureof effectiveness.

A supplementary source of evidence is offered by the ‘benchmark’scores for executive agencies collected for the ‘Business ExcellenceModel’ (BEM) run by the Civil Service College (Minister for theCabinet Office, 1999, p. 9). The BEM was developed by the EuropeanFoundation for Quality as a framework for assessing the performanceof public and private sector organisations across a range of activitiesto help them measure and improve performance. Performance wasassessed on nine criteria relating to a mixture of procedures and per-formance outcomes. The model measured procedural performance inleadership, policy and strategy, people management, resources andprocesses. There were outcome focused measures of people satisfac-tion, customer satisfaction, impact on society and business results.The BEM process produced a profile of scores for individual executiveagencies on each of the criteria.

A pilot BEM scheme in 1996 involved thirty executive agencies,about a third of the total existing at that time. Data was held by theCivil Service College on a confidential basis. However, publishedinformation was available for executive agencies as a group.Comparing the aggregate score to benchmark scores for private sector performance reveals mixed results. Executive agencies laggedbehind on scores for leadership, impact on society and people satisfaction. On this last criterion, executive agencies achieved just


over half the private sector score although they scored higher thanthe private sector on customer satisfaction (Samuels, 1997, p. 4).These findings suggest some evidence of a customer focus in execu-tive agencies but they do not demonstrate a clear improvement associated with executive agency working.

Evidence about customer satisfaction over time is complicated by the wide range of people benefiting directly or indirectly frommost executive agencies’ activities. The annual reviews of executive agencies produced by the Cabinet Office discussed benefits for groups including firms, charities, other organisations and individuals, taxpayers, elected representatives, other government bodies and ministers (Minister for the Cabinet Office, 1999). However, ministerslocated in parent departments were a key group of beneficiariesaccording to the executive agency model. An indication of minister-ial satisfaction is given by executive agencies’ performance against thetargets set by ministers. This performance is summarised for a three-year period between 1995/96 and 1997/98 in Table 5.3. A large major-ity of executive agencies were in the ‘satisfactory or above’ category,indicating achievement of most or all targets. However, six executiveagencies, 8 per cent of those surveyed, were in the ‘some cause for concern’ category, reflecting achievement of just over half theirtargets. A further three executive agencies, 4 per cent, were in the

Table 5.3 Ministerial satisfaction with executive agencies (with percentagesand examples)

Number Examples(with %)

Satisfactory or 63 (88) Benefits Agency, Information above Technology Services Agency (ITSA),

Driver Vehicle Licensing Agency

Cause for concern 6 (8) HM Prison Service, Training and Employment Agency, Valuation Office,Public Trust Office, Intervention Board

Serious cause for 3 (4) Child Support Agency, Fire Serviceconcern College, Student Awards Agency for

Scotland

Total 72 (100)

Source: Survey of seventy-two executive agencies from 1995/96 to 1997/98.

‘serious cause for concern’ category, with half or more targets notachieved or very serious failure to hit targets.

Taking an executive agency from each category illustrates some ofthe performance issues associated with the category. First, the satis-factory or above category included the ITSA in the DSS. The Agencyhad a measure of effectiveness in the level of service to customerswithin budget, on time and according to contract. The target wasraised and met in each of the years from 1995/96 to 1997/98. A per-centage efficiency improvement target was met in each year. In addi-tion, there were several milestone targets for performance whichdiffered from year to year but were all met (ITSA, 1996, 1997, 1998).Over the three years, real terms administrative costs fell by 17 percent, which, when combined with the achievement of targets, sug-gests productive efficiency improvements. In 1997, the Secretary ofState for Social Security commented that the Agency was ‘demon-strating year on year performance improvements in efficiency andquality of service delivered to its customers’ (Chancellor of theDuchy of Lancaster, 1997, p. 193).

The ‘some cause for concern’ category included HM Prison Servicewhich achieved four of eight targets in 1995/96, six of eleven targetsin 1996/97 and eight of ten targets in 1997/98 (Minister for theCabinet Office, 2000, p. 180). The Agency had a crude measure ofproductive efficiency in cost per prisoner place. Performance on thismeasure ranged between £23,940 and £24,271 over the period, belowthe target level of £24,615 (Prison Service, 1998, p. 7). However, theChief Executive, Richard Tilt, was able to comment in the 1997 that‘I am very pleased to be able to report that we achieved the majorityof our targets. Those targets we missed we missed narrowly’ (PrisonService, 1997, p. 4).

The ‘serious cause for concern’ category is the third and lowestlevel of performance and included the Child Support Agency. TheAgency achieved almost three-quarters of targets over the period butperformance was very uneven. Only five out of ten targets were met in 1995/96, rising to six out of eight in 1996/97 and eight out of eight in 1997/98. The Agency’s budget rose by 21 per cent over the three-year period, suggesting that improved performance waspurchased at great cost and amid public criticism from the Secretaryof State, who commented in 1998 that ‘many challenges still face theAgency’ (Chancellor of the Duchy of Lancaster, 1998, p. 192).



A broader survey of performance in these three bodies shows thatperformance against targets gives only a partial indication of pro-ductive efficiency and effectiveness. A NAO value for money studylooking at productive efficiency and effectiveness in ITSA foundimportant parts of the Agency’s work were beyond the scope of thebody’s performance measures (NAO, 1996b, pp. 1–2). The perfor-mance measures for HM Prison Service did not reflect the major dispute between the chief executive and the Home Secretary overresponsibility for prisoner escapes in 1995. Broader evidence aboutthe Child Support Agency’s poor performance was reflected in poorperformance against targets in 1995/96 but later years’ performanceagainst targets presented an overly rosy picture. There were a widerange of concerns about unfair treatment of clients, poor quality ser-vice and low levels of productive efficiency. For example, the PublicAccounts Committee (PAC) was concerned that the administrativecost of £224m exceeded the £206m in payments collected by thebody in the mid-1990s. The very slow rate of clearing claims, withjust over half cleared within twenty-six weeks, was also noted as amajor concern (Harlow, 1999, p. 169).

The case study of the Benefits Agency allows a broader assessmentof performance for a body that was in the ‘satisfactory or above cat-egory’ for performance against targets in Table 5.3. The Agencyachieved over 70 per cent of the Secretary of State’s targets in nineout of the ten years between 1991/92 and 2000/01, as summarised inTable 5.4. The Agency achieved over 80 per cent of its internal targetslinked to the Secretary of State’s targets (NAO, 1998b, p. 1). TheSecretary of State for Social Security, praised the Agency’s ‘efforts tomaintain continuing high standards of customer service’ in 1998(Chancellor of the Duchy of Lancaster, 1998, p. 192). However, therewas not a clear improvement over time, as might be expected by the

Table 5.4 Benefits Agency: Secretary of State’s targets achieved/targets setbetween 1991/92 and 2000/01

91/92 92/93 93/94 94/95 95/96 96/97 97/98 98/99 99/00 00/01

16/20 18/23 19/26 16/22 19/22 16/22 8/15 5/7 11/14 11/14

Source: Benefits Agency Annual Reports (1992a, 1993, 1994, 1995a, 1996, 1997, 1998,1999a, 2000; Department for Work and Pension (DWP), 2001).

public interest perspective. Instead, there was a dip in the percentageof targets achieved in 1996/97 and 1997/98. According to officials inthe Agency, this deterioration occurred because of an expansion of workload, which is consistent with figures for the volume ofclaims and relatively poor economic situation at the time (I10, SeniorOfficial, Benefits Agency).

Beyond the service provided to ministers, various measures of cus-tomer satisfaction give an indicator of performance as experiencedby users of the Agency’s services. Surveys of customer satisfactionfound that around 83 per cent of customers regarded the service assatisfactory or better in the mid-1990s (Benefits Agency, 1999a).However, this measure of customer service gave a far from compre-hensive assessment of all aspects of service (NAO, 1998b, pp. 18–19).The external system of complaint handling provides another per-spective on user satisfaction, based on complaints to the Parlia-mentary Commissioner for Administration (PCA). Whilst these caseswere not typical of transactions between the Agency and welfareclaimants they offer an indication of the level of major failings onindividual cases. In 1998, there were 468 complaints and seventyinvestigations about the Agency (PCA, 1998, figure 5; 1999, figure 5).This number was small relative to the many millions of transactionscarried out by the Agency each year. The performance of the Agencyunder the Citizen’s Charter offers an additional indication of perfor-mance for users. The Charter Standards were incorporated in a revamp of the Agency’s own customer charter in 1993 (BA, 1994, p. 7). Average performance of benefit work ran at about 80 per centof target standards (NAO, 1998b). However, performance variedbetween 56 per cent of Income Support claims made within the fivedays standard to 97.9 per cent of the target for crisis loans the sameday (NAO, 1998b, p. 44; BA, 1999a). Overall, the service for users was patchy.

Another indication of effectiveness was whether programme bud-get was given to the people eligible to receive it, which was part ofthe Agency’s mission statement. This aim was compromised by thehigh levels of fraud and error in benefit payments. In 1997/98 fraudand error was estimated at 7 per cent of total social security pay-ments, including expenditure of £1530 m on Income Support andincome based Jobseeker’s Allowance, £184 m on child benefit andfraudulent encashment of order-books and giro-cheques amounting


to £19 m (PAC, 2000, sections 1–3). There was further ineffectivenessin the lack of take-up of benefit; a substantial number of people whoshould have been receiving benefit did not get the welfare paymentsto which they were entitled. In 1996, the DSS estimated that only 76 per cent of those entitled to receive Income Support and only 70 per cent of those entitled to Family Credit received these benefits(NAO, 1998c, p. 25).

Comparing the effectiveness of the Agency with the pre-agencysystem gives an indication of change associated with executiveagency working. Such a comparison is hampered by the different mixof benefits delivered in the 1990s compared to the 1980s. However,an indicator of relative effectiveness is the level of ‘fraud and error’of benefit payments, summarised in Table 5.5 for SupplementaryBenefit (and, from April 1988, Income Support) and Family Credit.There was significant fraud and error in payments in the 1980s; inthe five-year period 1984/85 to 1988/89, error rates ranged between9.6 and 11.6 per cent with an average of 10.2 per cent. In the 1990s,the Agency was associated with roughly double the level of fraud anderror, ranging between 18.3 and 21.9 per cent with an average figureof 20.1 per cent. Similarly, figures for fraud and error on FamilyCredit, which was not delivered in the same form in the early 1980s,rose from 8.6 per cent in 1988/89 to 18.3 per cent on averagebetween 1995/96 and 1997/98. However, the deterioration in perfor-mance was not nearly as great as these figures might suggest. Themeasurement of fraud and error in the 1980s underestimated the truelevel by at least 30–40 per cent, and possibly more. The measuresused by the Agency in the 1990s were much more accurate. A further


Table 5.5 Percentage of fraud and error in payments before and after theestablishment of the Benefits Agency

Benefit 1984/85 to 1988/89 1995/96 to 1997/981988/89 (%) 1997/98 (%)

average (%) average (%)

Income Support/ 10.2 9.1 20.1 19.3SupplementaryBenefit

Family Credit — 8.6 18.3 17.3

Source: Carter et al., 1992, p. 93 and Benefits Agency, 1995a, 1996, 1997, 1998.

form of ineffectiveness was in ensuring that those entitled to claimbenefit took up this entitlement. In the 1980s, 89 per cent of thoseentitled to receive Supplementary Benefit received it (NAO, 1988, p. 19). An even larger potential client group continued to be ignoredunder executive agency working with only 76 per cent of those enti-tled to receive Income Support in the mid-1990s claiming this bene-fit. However, a substantial portion of this deterioration was the resultof improvements in measuring the take-up of benefit (NAO, 1998b,p. 25). On this basis, it seems best to conclude that the Agency wasassociated with a broadly similar, or possibly slightly worse, level ofeffectiveness.

Senior officials argued that there was a marked improvement inthe standard of service provided to claimants. There was better deco-ration of offices and a better ‘feel’ to the service, both of which wereespecially notable in the early years of the Agency (I10, Official,Benefits Agency; I13, Senior Official, Benefits Agency). A comparisonof the ‘quality’ of customer service before and after setting up theAgency illustrates this change. An NAO report on the quality of services to the public at local benefit offices highlighted serious dissatisfaction in the late 1980s. A customer survey found that 25 percent of claimants found the service poor with 75 per cent rating itfair or good (NAO, 1988, p. 1). By the mid-1990s the situation hadimproved with 83 per cent rating the service fair or good, althoughthis figure may have overstated performance. There was, however, anincrease in complaints made to the Parliamentary Commissioner forAdministration. In the late 1980s, there were, on average, 221 com-plaints and eighty-two investigations each year against the whole ofthe DSS (PCA, 1990, p. 28). The overwhelming majority of thesecomplaints referred to work later transferred to the Agency. By the1990s, the number of complaints had more than doubled althoughthere was a slight fall in investigations. This fall did not reflect areduction in complaints worthy of investigation because the PCAintroduced a system of settling cases informally without investiga-tion in the 1990s. There were, on average, a further forty-eight complaints of this type in the late 1990s, suggesting that the Agencymay have worsened rather than improved the occurrence of majorfailings on individual cases.

The complaints were about similar problems throughout the 1980sand 1990s, including delays and inaccurate payments (PCA, 1990,


1998, 1999). In the late 1980s the Commissioner lamented that ‘Formany years complaints about the administration of social securityhave loomed large in the Office’s investigative workload’ (PCA, 1990,p. 3). Almost a decade later, the Commissioner expressed a similarview that ‘complaints [against the Benefits Agency] accounted for thegreatest number of complaints which the Ombudsman receivedagainst any single body’ (PCA, 1999, section 3.4). Given the size ofthe Agency, the continued salience of complaints is unsurprising.The increase in complaints must be considered in the context of rising expectations of services and much better publicity about howto make complaints (I15, Senior Official, Benefits Agency). However,the Agency was not a panacea and long-standing problems of inef-fectiveness remained.

The overall productive efficiency of the Agency is given, in crudeterms, by the ratio of inputs to outputs. Resource inputs, includingstaff, increased each year, with a dip in the period 1995/96 to1997/98. However, workload fluctuated in a similar way to inputs.On the Agency’s own measure of productive efficiency, based onwork output per unit of staff, efficiency improved by 12 per centbetween the start of its life in 1991/92 and 1995/96. However theindex did not cover much of the work done in the central benefitdirectorates (NAO, 1998b, p. 32). A crude indicator of productive effi-ciency for part of the organisation’s activities was spending per unitof workload in Social Fund Crisis Loans. This workload figure wascontinuously available throughout the 1990s, at a time when mostother benefits changed, and reflected changes in workload related toeconomic cycles. The index of real term cost per case fell throughoutthe 1990s, except in 1998, and ended up in the late 1990s about 20 per cent lower than in the early 1990s. There was no similar indicator of efficiency to enable a reliable comparison between per-formance in the 1980s and 1990s. However, the findings for the1990s are consistent with interviewees’ claims that the increase inresources consumed by the Agency did not keep pace with increasesin workloads (I4, Official, Benefits Agency; I10, Senior Official,Benefits Agency). However, the substantial areas of ineffectiveness inpayment delivery, combined with a patchy quality of service, suggestthat productive efficiency did not dramatically improve in the firstdecade of the Agency’s operation.



Section 3: The effect of agency structures on the performance of the Benefits Agency

Tracing the effects of executive agency structures is difficult becausethe consequences of these structures are difficult to separate fromother factors, such as the environment in which the Agency wasworking. It is not possible to explore the counterfactual of how theactivities would have been handled if the old organisational arrange-ments had been maintained. However, the executive agency structures had a substantial influence on the behaviour of officialsand performance outcomes in the Benefits Agency. This experienceenables an evaluation of the public interest hypotheses that thestructures should bring about improved performance, and the hypo-theses of the bureau-shaping perspective which suggest a worseningof economy and unchanged productive efficiency.

The semi-detached organisational structure permitted the BenefitsAgency to develop its own corporate systems to focus on the task athand and to develop distinctive working patterns. Executive agencyworking enabled a new corporate identity of which refurbishedoffices and name badges for staff was just the most visible part (I4,Official, Benefits Agency; I13, Senior Official, Benefits Agency). Thechief executive set up a review of the Agency’s structure to explorehow better to organise its resources which was easier to set up, runand implement than under the old departmental structure (I13,Senior Official, Benefits Agency). The Agency was initially set up witha central core based in Leeds, a set of specialist benefit centres and alocal network structure of 159 Districts. The Districts were runthrough twenty Area Directorates in three territories (BenefitsAgency, 1993, p. 5). As a result of the review, the Agency reduced thenumber of territories from three to two and rationalised the areastructure from twenty to thirteen Area Directorates. These changesreduced the size of the relatively ‘unproductive’ middle managementlayer between the centre and the local network (I13, Senior Official,Benefits Agency). There was extensive development of Agency spe-cific management systems. The Agency exploited pay and gradingdelegations to develop systems tailored to its needs, whilst discussingmatters of common interest with other executive agencies in the DSS (Clifford, 2000, p. 38). The NAO found that significant cost savings had been achieved by contracting out work, for example, in

switchboard operator systems by negotiating with telephone serviceproviders, controlling private use of telephones by staff and checkingbills using improved inventories (NAO, 1996, pp. 5–6).

The accountability system had profound effects on the operationof the Agency. The budget limit was set by the Secretary of State andthe Agency operated within a gross running cost system of control.If the Agency felt that changes in workload affected its ability tocarry out its tasks it had to seek approval to change its business planand funding levels. The budgetary control mechanism was effectivewhen ministers decided to bring about reductions in administrativecosts in the mid-1990s. Staff in the Agency commented on thetoughness of the regime they faced. The chief executive described the request for a 25 per cent reduction in administrative costs in afour-year period as feeling like ‘somebody had three cups and a peaunder one of them’, the ‘pea’ being the budget he had hoped toobtain and the budgetary game being one he was unlikely to win(Social Security Committee, 1998, question 142). This experiencesuggests that, whilst there may have been a desire amongst officialsin the Agency for higher administrative budgets, consistent withattempted budget maximising behaviour, the outcome was notalways an increase in budget. Instead, administrative costs and programme expenditure varied according to the stage of the generaleconomic cycle and state of unemployment. Total economy was, inlarge part, linked to the level of demand from those claiming benefitaccording to the eligibility criteria set out in social security legislationand ministers were able to squeeze administrative costs when theychose to in the mid-1990s.

The performance target system was deeply embedded in the operation of the Agency. Internal management targets were tied tothe external reporting regimes. In 1997–98 the Agency’s manage-ment team monitored performance against all the 124 Secretary ofState’s and related internal management targets. It also led to moreformal and independent central checking to ensure accuracy in theagency’s own performance systems (I13, Senior Official, BenefitsAgency). A proportion of pay for people in the organisation, includ-ing the chief executive, was linked to achieving targets. This systemcreated incentives to direct considerable attention to activities thatimproved performance against targets (I13, Senior Official, BenefitsAgency).


Ministers and civil servants in the Department were concernedabout the their inability to act as an ‘intelligent customer’ for theAgency’s services using the target system (I14, Senior Official, DSS).The system led managers to pay insufficient attention to the perfor-mance of activities that were valued by ministers but not the mainfocus of targets. A major problem arose because the focus on achiev-ing the clearance time target hampered efforts to reduce fraud anderror, contributing to the poor performance in this latter area. Anofficial commented ‘We used to be under pressure to get all claimsdone within the clearance time [target] more than anything else’ (I4,Official, Benefits Agency). The focus on short-term performanceagainst targets and budget limits hindered attempts to redesign sys-tems. Area managers were responsible for the achievement of targetsseparately related to different benefits, so that each benefit wastreated largely as a separate activity within the organisation (NAO,1998b, p. 38). However, clients were often eligible for several benefitsand the same information was collected and stored by different partsof the Agency. A junior member of staff working in the local officescommented ‘Each of the systems we have is based on a different ben-efit rather than allowing us to deal with an individual claimant whomight be looking at several relevant benefits’ (I4, Official, BenefitsAgency). This pattern of working led to productive inefficiency anda poorer quality of service for claimants than if a client focused system was adopted, but these were not encouraged by the externaltarget system. The emphasis on achieving budget cuts hinderedinvestment that would have led to improvements in the longer run.Expenditure on capital investment fell year on year from 6 to 9 percent of administrative costs in the early 1990s to between 0.5 and 1 per cent in the late 1990s (Benefits Agency, 1993, 2001). Intervieweescommented on the pressure to cut administrative costs in the shortterm, especially in the mid to late 1990s (I13, Senior Official, BenefitsAgency).

Attempts to change the performance outcomes by changing theAgency’s targets were only partially successful. In 1998, the Depart-ment reduced the number of targets to concentrate on five accuracytargets and two financial recovery targets. The strategy was laid downin the document Safeguarding Social Security (DSS, 1999b). A key aimwas to reduce programme losses in Income Support and Jobseeker’sAllowance by 30 per cent by 2007, with at least a 10 per cent


reduction by 2002 (DSS, 1999b, p. 4). This change meant that evenmore areas of benefit work were not covered by the Secretary ofState’s targets.

The informal interpretation of targets by staff in the Agency limited some of the undesirable side effects of the system. Officials inthe Agency and Department did not think that it was possible to con-centrate attention on all Secretary of State’s targets at once, but theperformance target system did not rank their relative importance.Instead, senior officials in the Agency formed an impression of theimportance of different targets from informal signals by ministersabout political priorities, for example, in speeches and press releases(I13, Senior Official, Benefits Agency). Targets that Agency staff felthad been downgraded were put on an informal ‘care and mainte-nance’ basis, with staff attempting to avoid deterioration in perfor-mance but focusing more management attention on the prioritisedtargets (I13, Senior Official, Benefits Agency). Similarly, Agency stafffelt that not all activities of importance to ministers were expressedin the targets. For this reason, they were cautious about putting alltheir effort into achieving targets to the detriment of broader perfor-mance, although targets undoubtedly influenced behaviour to anextent (I10, Senior Official, Benefits Agency).

The second chief executive commented that staff did not believethat all of the Secretary of State’s targets were achievable. They alsothought that departmental officials shared this view so that theywould not be held to be at fault for failing to achieve the target. Hecited a target for accuracy of payments which was set at 87 per centin the mid-1990s, stating that ‘We have an overall target [for accu-racy] for the Agency and when I came in we were so far adrift fromit that people had given up [on achieving the target]’ (Social SecurityCommittee, 1998, question 152). Instead he called the target a goal,in the sense of something to move towards rather than necessarily toachieve in a single year as part of a project to get the Agency to raiseperformance from below 80 per cent towards 84 per cent. The figureof 87 per cent was only achievable in the longer term (Social SecurityCommittee, 1998, questions 188–90).

The targets reflected objectives of the Agency that were of priorityfor ministers rather than prioritisation of objectives by other groups. Inparticular, during the mid-1990s, customer service of particular inter-est to users and schemes to improve the take-up of benefit amongstpotential users were not prioritised relative to other objectives of the



Agency. During this period, ministers were most concerned about con-trolling costs, especially administrative costs, and there was a tensionbetween cost control and meeting other objectives. Whilst there werecustomer satisfaction surveys there was no extensive system of user consultation and potential users had virtually no voice in the target-setting process, instead the targets were set by ministers after consul-tation with senior Agency staff. The objectives valued most by usersand potential users lost out to ministers’ priorities.

The take-up of benefit was not prioritised in the performance tar-get system. The Agency’s chief executive, acknowledged that they didnot have an active national programme to boost the take-up of ben-efit (Social Security Committee, 1998, pp. 157–8). An intervieweecommented that, in the early and mid-1990s, ministers did not givethe impression in their informal statements or internal directivesthat take-up of benefit was a key aim of the organisation (I13, SeniorOfficial, Benefits Agency). Based on their experience with claimantsand potential claimants through their national network of centres,the Citizen’s Advice Bureau complained that ‘At present, advice and information about benefits appears to be a reduced priority’(Citizen’s Advice Bureau, 1998, p. 36).

The Agency’s concern with meeting budgetary targets made it lessresponsive to problems with the quality of service than it might havebeen. Whilst there were notable improvements in the décor of bene-fit offices and speed of clearance times, the target system led toarrangements that were not focused on the needs of clients. Agencystaff were highly specialised and focused on individual benefits, soclients had to be passed between several different members of staffcausing them inconvenience (Civil and Public Services Association,1997, p. 12). The Agency was very reluctant to give adequate redressto claimants when an error was made. The PCA noted that, evenwhen the Agency had been clearly at fault, it took pressure from himto persuade the Agency to give redress because of the Agency’s preoccupation with reducing costs (IA10, Senior Official, PCA; PCA,1999, section 3.4).

Conclusion

The public interest perspective’s hypothesis that executive agenciesimprove the economy of administrative costs by 5 per cent was par-tially consistent with the practice of agency working. There was a fall

of 4.6 per cent in real terms administrative costs for executive agen-cies as a group, with a median fall of 4.4 per cent, between 1995/96and 1997/98. However, there was a striking difference between trading and non-trading executive agencies. Whilst the public inter-est hypothesis is consistent with non-trading agencies, where themedian score was a fall of 5.3 per cent, the median increase of 8.2 percent for trading agencies is more consistent with the bureau-shapingperspective. The changes in the Benefits Agency were only partlyconsistent with the public interest perspective. Whilst real termsadministrative expenditure fell by 4.4 per cent between 1995/96 and1997/98 it was still 9 per cent higher in 2000/01 than in the first yearof the Agency in 1991/92. In the first decade of the Agency’s life, stafffigures rose 3 per cent and real terms total spending rose 13 per cent.

The assessment of productive efficiency and effectiveness, using abasket of indicators, produces results that are only partially consis-tent with public interest perspective. Executive agencies were broadlysuccessful in achieving their targets, 88 per cent of the seventy-twoexecutive agencies surveyed in the late 1990s were in the ‘satisfactoryor above’ category. On average, only 6 per cent of executive agencieshad accounts suggesting problems with probity or serious misuse ofresources. The Benefits Agency was able to improve some aspects ofcustomer service, including speed of clearance of some benefits, butthere were serious shortcomings, particularly fraud and error amount-ing to, on average, 7 per cent of social security payments. Some indicators suggest that there was greater fraud and error than beforethe Agency was established, though any deterioration is largely offsetby improvements in the accuracy of information. According to acrude measure of productive efficiency based on workload, there wereimprovements in the 1990s. The deterioration in economy in theearly 1990s was, in large part, correlated with rising demand for welfare services. But the problems in effectiveness meant that the pro-ductive efficiency of the Agency was compromised in key areas andlong-standing performance problems that pre-dated the Agency werenot ameliorated.

The semi-detached organisational structure of the Benefits Agencyallowed a better focus on the delivery of the Agency’s objectives,facilitating restructuring and more appropriate pay and grading sys-tems. Whilst budgets rose, ministers were able to use controls tosqueeze administrative costs in the mid-1990s. Performance targets


had some detrimental effects, especially leading officials to concen-trate on speed of clearance without sufficient regard to fraud anderror in payments in the early 1990s. There was a tension betweenachieving objectives prioritised by users and potential users of ser-vices and the objective of cost control prioritised by ministers in themid-1990s. However, ministerial priorities dominated those of usersand potential users of services in the target regime. Despite speedingup the time for processing some claims, there were deficiencies incustomer service in many areas. The Agency made little effort toincrease the take-up of benefits amongst those who were entitled butwho did not claim.


108

6Executive Agencies and Central Government SystemicPerformance

The hypotheses of the public interest and bureau-shaping perspectivesfor the systemic performance of central government are summarisedin Table 6.1. Section 1 assesses systemic economy. Section 2 assessessystemic productive efficiency and effectiveness. Section 3 assesses thecontribution of executive agency structures in the Benefits Agency tothe systemic performance of the social security system.

Table 6.1 Hypotheses about central government systemic performance


Economy 1. Economy improves with 1. Economy in administrativesavings in administrative and programme costs isand programme costs reduced through the

outcome of budget maximising behaviour

Effectiveness 2. Effectiveness improves 2. Effectiveness worsenswith better achievement because of public sectorof objectives externalities

Efficiency 3. Productive efficiency 3. Productive efficiencyimproves including worsens because of publicsavings in administrative sector externalitiesand programme costs

Central Government and Systemic Performance 109

Section 1: Systemic economy

According to the public interest perspective, executive agenciesimprove central government systemic administrative and programmecosts, whilst the bureau-shaping perspective expects a worsening inboth elements of systemic economy. Changes in real terms (in 1996/97prices) central government administrative costs and total spending,including administrative costs, for 1988/89 to 2001/02 are summarisedin Figure 6.1. Real terms administrative expenditure rose from£12,863 m in 1988/89 to a peak of £15,117 m in 1993/94 before fallingback in the late 1990s then rising to £15,180 m in 2000/01, a rise ofabout 17 per cent over the period as a whole. These figures suggest thatexecutive agencies were associated with periods of decreasing centralgovernment real terms administrative costs but, overall, that the levelof spending was higher at the end of the period than at the beginningwhich is more consistent with the bureau-shaping than the publicinterest hypotheses.

The executive agency system itself involved costs in running theperformance target and monitoring systems in the parent departments

300,000

200,000

100,000

0

Total costs Admin costs

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

£ m

199

6/97

pric

es

Figure 6.1 Economy of central government 1989–2002

Source: Chancellor of the Exchequer (1994, 1998, 2002).


and collecting the information required in executive agencies. Thecosts of running systems for the Benefits Agency were estimated at £7 m each year (NAO, 1998b, p. 20). If the costs were one-tenth of this figure for an ‘average’ executive agency, to reflect the large size of the Benefits Agency, then a conservative estimate of the totalcost across central government was £100 m each year in the late1990s. However, these costs were a small proportion of total centralgovernment administrative costs, forming only 0.6 per cent of thisfigure in 2001/02. By themselves, the systems were not a major contribution to the rise in central government administrative costssince 1988.

The executive agency model was associated with improvements in economy of civil service staff inputs. The total number of civil servants fell in the ten years after the reform from 580,000 in 1988to 463,000 in 1998, a fall of 20 per cent as shown in Table 3.2 inChapter 3. However, there was a similar proportionate fall in staffingin the same length of time prior to the reform (Efficiency Unit, 1988,p. 23). It was perhaps more of an achievement to have continuedreductions in the 1990s compared to the 1980s because the latterperiod reductions were from a lower base, all time civil service numbers having peaked in 1976 at 763,000 (Cabinet Office, 2000, p. 7). However, the reduction in staff was partly the result of con-tracting out of services to private firms in initiatives includingCompeting for Quality, so expenditure gives a better indication ofresources used by central government. After 1998, civil service num-bers began to rise, reaching 483,000 by 2001, so the period of execu-tive agency working was not exclusively a time of declining civilservice staff numbers.

Total real terms (in 1996/97 prices) central government spending,including not only administrative costs but also expenditure on pro-grammes, did not improve after the start of the executive agencyreform, as shown in Figure 6.1. Instead, real terms total expenditurerose from £164,700 m in 1988/89 to £217,000 m in 2001/02, a rise of32 per cent. The biggest increases occurred in the late 1980s/early1990s and since 2000. In the ten years after the reform started,spending rose 19 per cent, slightly higher than in the ten years beforethe reform when central government spending rose in real terms by13 per cent (Chancellor of the Exchequer, 1987, p. 34).


Section 2: Systemic productive efficiency and effectiveness

The public interest perspective’s hypotheses are for improvements tosystemic effectiveness, with better achievement of central govern-ment’s objectives, and improvements in productive efficiencyincluding improvement in administrative and programme costs. Incontrast, the bureau-shaping perspective suggests public sector exter-nalities compromise the achievement of systemic objectives and,combined with budget maximising, this worsens productive effi-ciency. Some of the negative consequences show up in individualexecutive agencies’ performance, through the undesirable side effectsof executive agencies on each other, whilst other problems emerge inthe performance of the rest of central government, including depart-ments. Systemic performance problems include actions with sharedbenefits not being taken and negative externalities as executive agencies are not concerned with the effects of their actions on otherbodies.

The systemic effects of executive agencies were increasingly recognised by Government towards the end of the 1990s, along withsimilar problems in other New Public Management (NPM) structuresin the UK public sector (James, 2002). These problems were collec-tively termed a lack of ‘joined-up government’ which was a keytheme of the Modernising Government White Paper (Minister for theCabinet Office, 1999), the Wiring it Up report by the Cabinet Office(Performance and Innovation Unit (PIU), 2000), the TreasuryComprehensive Spending Review of 1998 and Treasury SpendingReview of 2000 (Chief Secretary to the Treasury, 2000). In its broad-est sense, the ‘joined-up government’ agenda was a vague set of aspi-rations for performance suggesting that outcomes of value to citizensand their representatives should be delivered in a way that is not dictated by organisational boundaries. However, more specifically,‘joined-up government’ reflected concern about the performanceconsequences of boundaries between different departments, execu-tive agencies, non-departmental bodies, local government and other units. The authors of the Modernising Government White Papercommented that ‘Great gains in public sector management havecome from definition of task and delegation of management and the Government is determined that these are not lost. However, this


concentration on specific tasks has sometimes distracted attentionfrom the wider general objectives of government and people. TheGovernment wants to give more attention to the coherence of policyacross institutional boundaries … to operate in a joined-up way’(Minister for the Cabinet Office, 1999, p. 1). The traditional depart-mental division of central government, which pre-dated the execu-tive agency system, was reinforced by executive agencies reportingprimarily to their own parent department. In many cases, organisa-tional divides within departments were created or strengthened byfragmenting departments into executive agencies, including a splitbetween strategic policy making handled by departments withimplementation of policy by executive agencies.

The way executive agencies responded to new developments ininformation and communication technology illustrated the detri-mental effect of executive agency structures on systemic performance.Changing technology allowed the possibility of more ‘e-government’by providing services or information about services using web-basedinternet, intranet, email, electronic data transfer and associated sys-tems (NAO, 1999c, p. 11). Exploiting these opportunities is a keychallenge for the contemporary delivery of public services (Dunleavyand Margetts, 2000; Chadwick and May, 2003). A survey of seniorofficials in government, including senior executive agency officials,found that 85 per cent felt that websites were very important or important to the operation of their organisations ( NAO, 1999c, pp. 12, 17). The importance of e-government was noted in theModernising Government White Paper which suggested targets for elec-tronic capability in transactions between government and citizens of25 per cent by 2002, 50 per cent by 2005 and 100 per cent by 2008(NAO, 1999c, p. 53). The targets were subsequently tightened with atarget that all services should be able to be handled electronically by2005 (NAO, 2002b, p. 1).

The development of e-government was, in part, an issue of individ-ual executive agency performance. However, it raised significant systemic performance issues. In one sense, all progress towards use of web-based and similar technology is collaborative between organi-sations because they contribute to strengthening the overall network.Even more importantly for central government, the technologyoffered the prospect of new and improved services for citizens involv-ing contributions from several organisations, for sharing information


between bodies, and for the provision of common services used byseveral organisations in different parts of government.

Executive agencies made the greatest strides towards e-governmentin developing their own, narrowly focused systems. Some tradingagencies were able to use their commercial freedoms to develop ser-vices for their customers. A group of academics led by PatrickDunleavy and Helen Margetts worked with the National Audit Office(NAO) to survey developments in e-government. Trading agencies,notably HM Land Registry, Companies House, Driver StandardsAgency and the Patent Agency had websites that were among themost developed in the UK public sector. Progress was more modest innon-trading bodies. Although the Employment Service and CourtService were praised for having developed interactive websites, only53 per cent of sixty-six executive agencies surveyed had websites in1999, rising to 83 per cent in 2001 (NAO, 1999c, pp. 56–7). These figures were slightly lower than departments which had 86 per centcoverage in 1999 and 92 per cent in 2001 (NAO, 2002b, p. 59).

The potential of e-government technology to deliver systemic ben-efits was exploited much less than benefits for individual executiveagencies. Surveys revealed that users found central government sitesconfusing and that the sites were organised more around organisa-tional boundaries than the services or information wanted by users(NAO, 1999c, p. 50). The shared public sector portal, open.gov.uk,was launched in 1997 but was not used by all executive agencies. Inthe late 1990s, only one in twenty people accessing a central gov-ernment site did so via the portal (NAO, 1999c, p. 52). The site wasrelaunched as ukonline.gov.uk, but the progress in its developmentwas disappointing, with useage figures for mid-2001 little higherthan at the start of 1999. Another shared service, the GovernmentSecure Intranet, intended for communicating sensitive information,was subject to limited participation by executive agencies. Of theforty bodies on the system in the Autumn of 1999, the majority weredepartments rather than executive agencies (NAO, 1999c, p. 53). Thegrowth of facilities beyond email, such as document sharing, wasvery limited (NAO, 2002b, p. 53).

The development of co-ordinated e-government was limited, inpart, by the reduction of central co-ordination of IT systems in the1980s and early 1990s alongside the move to executive agency work-ing. Expenditure on IT was seen as primarily a matter for individual


executive agencies and departments. By 1996, only a small Central ITUnit was left in the Cabinet Office, primarily to provide advice to therest of government (NAO, 1999c, pp. 50–1). Executive agenciesdeveloped their own separate IT systems with their own standards,often with little thought to how they related to each other. In a sur-vey, sixteen out of twenty government departments noted the needfor partnership with other organisations as one of the top six issuesthat needed to be addressed if e-government was to be successfullyimplemented (NAO, 2002a, p. 2).

Section 3: The effect of executive agency structures on systemic performance in the social security system

The Benefits Agency was one of several central and local governmentbodies in the social security system, in total, spending around£90,000 m each year in the late 1990s (NAO, 1999a, p. 27). The usersof the system included 2.85 m sick or disabled, 1.3 m long-termunemployed, 0.75 m low paid and 0.94 m single parents on lowincome. Beyond these groups were around 12 m people includingpensioners, people with children and others who interacted with the system less frequently. The Benefits Agency was involved in sub-stantial joint working with the Department of Social SecurityHeadquarters (DSS HQ) in providing information for policy makingand running programmes on its behalf, the Employment Serviceexecutive agency in co-delivering the Jobseeker’s Allowance benefit,the War Pensions Agency in veterans welfare, the InformationTechnology Services Agency (ITSA) in IT systems and the Contribu-tions Agency in running the National Insurance system. Other bod-ies relied on the Benefits Agency to do their jobs properly, especiallylocal authorities in delivering Housing Benefit, which involved theAgency providing information about claimants’ eligibility for thisbenefit.

There were several significant public sector externalities in thesocial security system associated with the Benefits Agency, these aresummarised in Figure 6.2. The externalities arose from the organisa-tional identity of the Agency as a separate unit and the performancesystem which encouraged it to focus on its own aims. The externali-ties included poor information sharing between the Agency and DSSHeadquarters, negative externalities for other delivery bodies, and

opportunities forgone for developing shared services with otherorganisations.

Poor information exchange between the Agency and the DSS Headquarters

The separation of the Benefits Agency from the Department of Social Security (DSS) solidified a long-standing divide between policy making and operational parts of the social security system. The firstchief executive commented that the views of people delivering services were not taken sufficiently into account by those in policy-making positions (Bichard, 1999, pp. 7–8). The second chief execu-tive echoed this view, noting that the Agency attempted to identify‘weaknesses in the system which may essentially be down to somedetail of the policy design’ and ‘some of the rules and regulations’but that these were not always easy to communicate and rectify(Public Accounts Committee (PAC), 1999, question 165). An inter-viewee commented that executive agency working had not been ableto overcome the long-standing problem of ‘Grade 5s in the Adelphi[the location of DSS Headquarters] spending their careers developinghuge and complex rulebooks without much thought for the peoplewho have to work with them’ (I13, Senior Official, Benefits Agency).


Negative informationalexternalities(e.g. inaccurateinformation)

Poor information exchange(e.g. state pensions and problemsof new policies post 1997)

Charities,Post Office,other bodies

Unfulfilled potentialcollaboration (e.g. in e-government)

Departmentof Social Security

BenefitsAgency

Local Authorities’ Housing Benefit work

Figure 6.2 The Benefits Agency and public sector externalities in the socialsecurity system


Neither the Agency nor Department could cite many cases wherechanges had been made in response to comments from the Agency.One such case was the alterations made to legislation about mortgagecompensation which allowed for payment of arrears. Agency staffreported that levels of arrears were difficult to assess, leading to inac-curacies. They convinced DSS Headquarters staff that the legislationneeded to be changed to make the payment system easier to admin-ister. Within 12 months, accuracy had improved by 1 per cent, saving £13 m. But the Agency noted that this case was unusual andthat the mechanisms for involving delivery staff in policy makingwere limited (Benefits Agency, 1999b).

The communication problems were evident in many areas of thesocial security programme. However, a large policy failure occurredin failing to make potential claimants of the State Earnings RelatedPension aware of changes to the policy on eligibility. The handlingof state pensions required co-ordinating Benefits Agency staff inLeeds, Newcastle and in local offices with staff in the DepartmentHeadquarters in London responsible for supporting ministers in thedevelopment, maintenance and evaluation of pensions policy. TheDepartment originally made a change to the eligibility requirementsin 1986, before the Agency’s creation. However, the organisationalstructure inhibited the flow of information about this change toAgency staff. There was inadequate ‘end to end’ communicationfrom those making policy in the Headquarters to those in the localoffices providing the information to users (NAO, 2000c, pp. 10, 26).The Agency failed to incorporate information about the change ininformation they gave to clients. Leaflets and staff briefings toclaimants continued to give wrong information until 1999. On thisbasis, many users erroneously believed they were entitled to the pen-sions. In the end, it was a charity representing users of the servicesrather than Agency or DSS staff that brought the continued provisionof wrong information to the attention of ministers. The Departmentis likely to have to pay a large amount in compensation to thosegiven incorrect information (NAO, 2000c, p. 28).

The problems of developing policy in a system of executive agencyworking became more acute after 1997 when a new administrationcame into office. Their ‘New Labour’ agenda involved a change inemphasis towards encouraging employability and a social securitysystem more ‘responsive to, and providing a more direct service for

the public’ (Chancellor of the Duchy of Lancaster, 1997, p. 192). Thepolicy change further increased the need for inter-organisational collaboration in the social security system by requiring more inter-vention to improve the employability of clients rather than primarilyprocessing their benefit claims. However, the information collectedby the Agency was linked to a performance target regime focused onindividual benefits. This system did not provide policymakers withthe information about claimants that they wanted to develop the newstrategy for social welfare. The new policy environment required amore extensive and less crudely quantitative approach to assessingperformance and evaluating progress to enable fine-tuning of the policy during implementation (I15, Senior Official, Benefits Agency).Initiatives that might have improved communication between theHeadquarters and the Agency to facilitate this form of policy devel-opment were under-developed. The Benefits Agency had its ownintranet for communication within the organisation, separate fromthat in the DSS Headquarters and other parts of central government,which was less developed than the Headquarters arrangements (NAO, 1999c, p. 31). This uneven development limited the capacityof the system to pass complex information about delivery to thosemaking policy and to notify staff in the Agency about changes to policy.

Negative externalities affecting other bodies’ delivery of programmes

Beyond central government, there were significant negative exter-nalities created by the Benefits Agency for local government in theadministration of Housing Benefit. This programme cost about£11,100 m per year in 1996/97, or 12 per cent of the total spendingon social security (NAO, 1999a). Whilst local authorities were ineffi-cient in their administration of Housing Benefit because of problemsin their own operations, the Benefit Fraud Inspectorate (BFI) foundthat the Agency made a substantial contribution to poor perfor-mance (BFI, 1999d). About 66 per cent of people claiming HousingBenefit from a local authority made simultaneous claims for IncomeSupport or Jobseeker’s Allowance with the Agency (NAO, 1997b, p. 18). If a claimant was awarded Income Support or Jobseeker’sAllowance by the Agency then the information was passed on to localauthorities for their use in working out entitlement to Housing


Benefit. In particular, local authorities relied on the Agency for anassessment of claimants’ income or capital, a key consideration in theaward of Housing Benefit (BFI, 1999d, section 8.1). The Agency passedincorrect information about claimants’ eligibility resulting in fraudand error. Although the Agency had service level agreements withlocal authorities, which set out co-operation between the organisa-tions, a majority of local authorities felt that the agreements were notworking. Necessary improvements included better exchange of infor-mation, more cross agency working and better feedback on fraud cases(BFI, 1999d, section 1). In the mid-1990s, about 74 per cent of fraud-ulent claims on Housing Benefit were also fraudulent claims forIncome Support (NAO, 1997b, p. 21). About 7 per cent of the HousingBenefit budget was paid out in fraud and error, nearly 1 per cent of thetotal social security budget (NAO, 1997b, p. 1).

The Agency’s lack of concern with the incorrect information beingpassed on to local authorities arose, in part, because its own targetsdid not include this consequence of its activities. Although theAgency had a concern with its own performance in fraud and errorthis did not sufficiently reflect the importance of this area of work.The Agency did not take the service-level agreements seriouslyenough and there were internal targets for reducing fraud and errorwhich placed Agency staff in competition with those in local author-ities. The Agency scheme used the principle of ‘finders-keepers’ for detecting fraud with officials wanting to identify fraud to helpachieve their own targets for detection rather than working withlocal authorities to share information. The Agency’s ‘Spotlight’ anti-fraud initiatives did not involve local authorities as much asthey could have. In some cases there was an atmosphere of mutualsuspicion between the Agency and local authorities, although therelationship varied between different parts of the country (NAO1997b, pp. 62–71).

The Agency’s limited concern with the effect of its behaviour onother bodies was evident in working with other executive agencies inthe system. The Contributions Agency was responsible for protectingthe rights of contributors and the interests of taxpayers in the effi-cient payment and recording of National Insurance contributionsand partly relied on forms processed by the Benefits Agency.However, there were no high-level Benefits Agency targets for pass-ing forms to the Contributions Agency; 54 per cent of forms took


more than sixteen days resulting in out of date information beingused to create National Insurance accounts and damage to the col-lection of contributions (BFI, 1999c, sections 3.11–13). Some of theseperformance issues were ultimately resolved through the DSS man-agement structure, which supervised both executive agencies, andinformal pressure on the Benefits Agency (I9, Senior Official, DSS).However, the separate target regimes contributed to creating theproblem rather than contributing to its solution.

The Benefits Agency worked with the Employment Service andother bodies in delivering Jobseeker’s Allowance from 1996. The ben-efit was delivered from the Employment Service’s Jobcentres whichpaid the benefit on behalf of the Benefits Agency. However, much ofthe administrative work in support of the payments was handled bythe Benefits Agency. One senior official commented ‘Both agencieswould have preferred Jobseeker’s Allowance to go to one or another,rather than having to share it’ (I13, Senior Official, Benefits Agency).Staff in the Benefits Agency were under pressure to use resources topromote the achievement of their own targets rather than being con-cerned about the effects on the Employment Service. There was alack of information sharing, resulting in the same information beingrequested separately by each body, draining resources and inconve-niencing claimants (BFI, 1999b, Appendix B).

Attempts to smooth the delivery of the benefit were made by stafftransfers between the organisations. However, transfers were mademore difficult by different working conditions. Under the so-called‘Bichard Agreement’, named after the first head of the BenefitsAgency, staff were given the right not to be transferred to differentworking conditions when undertaking work related to Jobseeker’sAllowance. As a result, these staff continued to work behind screensto protect them from potentially hostile clients whereas theEmployment Service staff operated in a more open work environ-ment. This restriction maintained separate clusters of staff and con-tributed to difficulties in allocating staff to pressure points in thesystem (Social Security Committee, 1998, p. 176).

Despite these problems, the overall delivery of Jobseeker’sAllowance was much more successful then the experience withHousing Benefit. Most of the targets relating to the benefit were met.However, the working arrangements reflected structures that wereestablished between the Benefits Agency and the Employment


Service that went against the grain of the executive agency model.Structures included a joint management board. The second BenefitsAgency chief executive, Peter Mathison, and the chief executive of the Employment Service, Leigh Lewis, had quarterly meetings and prepared action plans to tackle difficulties. There was a JointOperations Team, with members from both bodies, to facilitate jointworking and local level meetings (Social Security Committee, 1998,question 163). In these respects, the successful joint working did notreflect successful operation of core features of the executive agencymodel but rather modifications to them in this area of these bodies’activities.

Unfulfilled potential shared services

The Benefits Agency did not develop co-operative approaches withother bodies to implement improvements to services that wouldhave promoted the Government’s objectives for the social securitysystem. Officials commented that the Benefits Agency was often poorat working with voluntary organisations to improve customer ser-vices. There was limited provision of information to claimantsthrough charities because of a concern to focus on existing targetsrather than developing new ways of delivering services (I4, Official,Benefits Agency).

The difficulties of developing the Single Work Focused Gateway,announced in September 1998 and later renamed ONE, illustratedsystemic performance problems. The project was intended to bringtogether the Benefits Agency, Employment Service, Child SupportAgency, local authorities and other welfare providers, including charities, to provide a more coherent service for claimants focused on their needs. This style of working initially covered only a smallpart of the Benefits Agency’s work, just £112 m for twelve pilots(Social Security Committee, 1999, paragraph 1). There were two prin-cipal variants of the pilots, the first using call centres and the secondusing partnerships between public and private sector bodies (SocialSecurity Committee, 1999, paragraph 10). The idea was to give aclaimant a personal adviser to take their details and assess employ-ment, benefit and training options from an initial ‘start-up meeting’.The personal adviser could then call on a team of experts to provideadvice to claimants, directing these organisations’ services to focuson the needs of the claimant.



Some aspects of the project involved a modification to executiveagency structures. The ONE projects were cited as examples of‘joined-up’ government in the 1999 Modernising Government WhitePaper (Prime Minister and Minister for the Cabinet Office, 1999, p. 31). The pilots were co-ordinated at a strategic level by a partner-ship of departments with a Project Board answerable to a joint min-isterial group. Below this level, the pilots were managed throughProject Management Groups at regional level bringing togetherBenefits Agency, Employment Service and local authority staff. Eachpilot was headed by a single manager who was drawn from one ofthese organisations (PIU, 2000, section 10.11).

Despite the modifications to executive agency working, several dif-ficulties of joint working between the Benefits Agency and otherorganisations emerged during the projects as a result of executiveagency structures. It was difficult to develop meaningful joint targetsto be shared by the different organisations that clearly allocatedresponsibility. The Social Security Committee expressed concern thatjob placement targets would be too crude to measure outcomes and that clearer assessment of the quality of advice about benefitsshould be undertaken (Social Security Committee, 1999, paragraphs68 and 70). In the end, no new targets were formally set, which made evaluation of performance very difficult. Management in thefield within the Benefits Agency continued to be done on execu-tive agency lines, through the district, area and regional offices.These offices were simultaneously trying to achieve the Agency’sown targets which, when these conflicted, generally took priorityover the joint projects (I15, Senior Official, Benefits Agency).

The styles of working and organisational arrangements differedbetween the bodies involved in ONE. Differences between the BenefitsAgency and other bodies’ personnel systems caused co-ordinationproblems. Each organisation had its own pay banding and promotionstructure which restricted the interchange of staff (Social SecurityCommittee, 1999, paragraphs 34 and 35). The separately developedcomputer systems hampered joint working. The Benefits Agency had its OpStrat systems (see Margetts, 1999) but operational sup-port for ONE was provided by the Labour Market System used by the Employment Service combined with a new Gateway Enquiry System. These systems were not completely compatible with eachother. Finally, the working styles of the organisations clashed. Local

authorities were felt to be more ‘can do’ and focused on solvingclients’ problems whereas the Benefits Agency style was more ‘pro-cedural’ (I14, Senior Official, DSS). The ‘procedural’ style in theAgency was partly because staff saw this approach as the best way to meet management targets (Social Security Committee, 1999, paragraph 41).

The Benefits Agency undertook an unsuccessful attempt to developa payment card for welfare claimants jointly with Post OfficeCounters Limited. The project sought to replace paper based meth-ods of paying social security benefits with a magnetic swipe paymentcard, and to automate the national network of post offices throughwhich many benefits were paid. A contractor, Pathway, wasemployed to provide the card system in May 1996. However, Agencyand Post Office faced pressure to achieve their separate aims and werenot able to agree on priorities for the project or trust each otherenough properly to share information about progress in its delivery(NAO, 2000c, p. 11). Their different reactions to the slow progress inrolling out the scheme illustrated these tensions. The Agency wasmost concerned about the impact of delays on administration costsand lost savings from reductions in fraud, in total running at £15 mper month in 1998. In contrast, Post Office Counters Limited wasmore concerned to achieve the indirect commercial benefits fromcontinuing payments from their offices. The payment card was even-tually discontinued in May 1999 with losses for both the Agency andthe Post Office (NAO, 2000c, pp. 30, 32).

Executive agency structures restricted the Benefits Agency’sinvolvement in collaborative e-government. The Agency launchedits website in October 1997 separately from other social security bod-ies with a team of its own staff to run it. There was massive under-investment in the site compared to the potential of the technology,especially given the potential for developing joint services with otherbodies. Officials commented that the site was not initially seen as akey priority (I10, Senior Official, Benefits Agency). The annual spendof £35,000 per year was just £14 per £ m of the Agency’s runningcosts. In Spring 1999, only about a dozen of the Agency’s staff hadpersonal computers fully suitable for internet access, so over 99 percent of officials could not view the Agency’s own site (NAO, 1999c,pp. 28–33). There was some co-ordination of IT across the Agencyand other bodies in the DSS through the ITSA. Links with bodies


outside the DSS were more limited, only 1 per cent of user sessionsfor all bodies within the DSS group involved a referral from the maingovernment website, open.gov.uk (NAO, 1999c, p. 30). The Agency’ssite was organised around different benefits delivered by the bodyrather than a user focused approach which might, for example, havefocused on different life events of service users, such as unemploy-ment, becoming retired or becoming sick. A redesigned websitewould have brought the services of other social security organisa-tions under one umbrella to provide a better service to claimants.

Conclusion

The use of executive agencies was not associated with improvementsin the systemic economy of central government as suggested by thepublic interest perspective over the period 1988 to 2001. Instead theexperience was more consistent with the bureau-shaping perspective.Whilst central government real terms administrative costs fell by 5.9 per cent between 1995/96 and 1997/98, these costs rose 17 percent, and real terms total expenditure rose 32 per cent, between 1988and 2001. Measuring economy in terms of staff, the number of civilservants fell by 17 per cent between 1988 and 2001. However, in part,this change occurred because of contracting out of services ratherthan cuts in resources, as reflected in the figures for administrativeexpenditure.

Consistent with the bureau-shaping perspective, executive agen-cies were associated with public sector externalities which, in somecases, detrimentally affected systemic productive efficiency andeffectiveness. Executive agency working limited the development ofsystemic e-government services. In the social security system, theBenefits Agency’s structures contributed to substantial public sectorexternalities and ineffectiveness and productive inefficiency in otherbodies. These problems were not new, the pre-agency system wasfragmented between local and central government and betweenparts of the DSS and other departments. However, executive agencyworking increased fragmentation and strengthened incentives forstaff to focus on narrow organisational goals.

There was a lack of communication between policymakers in the DSS HQ and staff in the Agency. Welfare policy was designedwithout sufficient thought for implementation and communication


problems led to inadequate sharing of information, contributing tothe failure to pass on correct information about state pensions eligi-bility to claimants. This problem became more significant under theLabour Government elected in 1997 because they sought to pursue amore active and complex policy agenda and found it difficult todevelop this policy by relying on information produced through theAgency’s reporting systems. The Benefits Agency created a negativeexternality in the poor quality information that it produced and thatlocal authorities used in administering Housing Benefit. The Agencyfailed to develop joint services with other bodies. Whilst joint work-ing with the Employment Service was relatively successful this was,in large part, because of ad hoc additions to executive agency struc-tures including joint boards and joint local working arrangements.The Agency developed e-government focused on its narrow concernsrather than working with other bodies to exploit the potential forsystemic cost savings and better services.


Part III

Conclusions

7The Perspectives, Findings about Practice and the Future Use of Executive Agencies

The findings from Part II suggest answers to the three questions setout in Chapter 1 and allow an overall assessment of the relative mer-its of the two perspectives. These answers and assessments are set outin Section 1. Section 2 draws on the perspectives and the findings tosuggest ways in which the use of executive agencies may develop inthe future. Section 3 assesses the potential relevance of the perspec-tives and the findings for understanding executive agencies in othercountries.

Section 1: Answers to the three questions about executive agencies and the relative merits of the perspectives

Drawing together the findings suggests that, overall, bureau-shapinghypotheses are more consistent with the practice set out in Part IIthan public interest hypotheses. In explaining why the executiveagency reform occurred, both perspectives are consistent with thedominant role of actors within, rather than outside, central govern-ment, as set out in Chapter 3. However, consistent with bureau-shaping rather than public interest hypotheses, politicians did nothave a plan for executive agency reform. The Next Steps proposal wasinfluenced by senior officials’ preferences for the organisation oftheir departments which limited the range of options. Pressure frompoliticians for more management attention to be paid to executive

127

activity triggered bureau-shaping strategies to pass executive activityto executive agencies rather than the alternative of becoming hands-on managers of integrated departments (a move from A to B ratherthan from A to C in Figure 2.3). Senior officials did not seem to bepursuing exclusively self-regarding motivations in this context but amixture of interest in their work and a view that this was the bestway to make policy. Their actions were generally consistent withmaintaining their position as the key policy workers in central gov-ernment. Both perspectives underemphasise the role of officials in the Efficiency Unit with an interest in developing reforms thatcould be implemented, making the initiation of Next Steps indirectbureau-shaping. They also neglect the role of staff in some execu-tive agency candidates who wanted greater autonomy, which wassignificant in setting up ‘pull’ pressures for change in some bodieswith non-mainstream civil service staff, especially specialist researchbodies.

The interaction between politicians and senior officials in the insti-tutional environment of central government departments was a keyinfluence on the reform. In this light, Dunleavy’s original bureau-shaping model places too much emphasis on the unconstrainedaction of officials and neglects the role of politicians. In contrast, the alternative account of Next Steps by Marsh et al. (2000, 2001, pp. 159–61) places too great an emphasis on politicians command-ing the reform and pushing through change. The difference betweenthe findings of this study and the work of Marsh et al., in part,reflects a difference in methodological approach. They gave greatestimportance to interview material, noting ‘Certainly, evidence fromour interviews indicated the role of politicians was crucial in the evo-lution and development of Next Steps. This, of course, contradictsthe argument that the process was driven by bureaucrats and, morebroadly, it fundamentally questions the thesis that it is civil servantswho are able to shape events or issues in their, as opposed to the min-ister’s, favour’ (Marsh et al., 2001, p. 159). In contrast, the methodadopted in this study uses interviews, other statements by actors andthe choices of actors to identify misrepresentation or partial presen-tation of events. This approach reveals that politicians triggered theprocess of change through their dissatisfaction with the status quobut senior civil servants adopted the specific executive agency reformas a bureau-shaping strategy.


The executive agency reform was a revolutionary change to theorganisational structure of central government. Around 83 per centof changes were mitigated or pure executive agency creation. Theprincipal exception was the 37 per cent of trading agencies that arosefrom nominal change. In 2001, 57 per cent of civil servants workedin executive agencies, broadly consistent with the half to three-quarters suggested in the bureau-shaping perspective but less thanthe 95 per cent suggested by the public interest perspective. Consis-tent with bureau-shaping, on average, 70 per cent of senior officialsin each department ended up working in the parent departmentrather than in executive agencies. Whilst many, relatively small,executive agencies handled non-routine work, this work was oftenspecialist, scientific research work, rather than policy work tradition-ally handled by senior civil servants.

Executive agencies all had the two principle features of the model,semi-detached organisation with a chief executive and specificaccountability framework including performance targets and an ele-ment of personal accountability by the chief executive. However, as Chapter 4 sets out, the operation of these systems was varied. Inparticular, trading agencies and executive agencies that were periph-eral to their departments’ activities were better able to develop theirown systems than non-trading bodies that were mainstream to theirdepartments. The Benefits Agency was an example of the latter kindof body but shared the principal features of the model with otherexecutive agencies.

The chief executives personified the delegation of managementand corporate accountability in executive agencies and were a dis-tinct new group of officials in central government. They did not havethe long tenure of traditional civil servants and there was some shiftin responsibility, and potential praise or blame, from departmentsand ministers to them for operational performance. However, themore politically salient issues of interest to those in Parliament,users, citizens and the media typically involved some form of jointdepartmental and executive agency responsibility for performance.Although these issues were generally the exception rather than therule, chief executives trod a fine line between using the frameworksin which they operated to focus on improving performance againsttargets, and, at the same time, exercising sensitivity to informal sig-nals from ministers and officials in their parent departments about

Future Use of Executive Agencies 129

how to handle matters of political sensitivity. Chief executives with knowledge of the Whitehall village as ‘insiders’ or those from‘outside’ who were able and willing to adopt this method of workingwere able to avoid highly public disputes. Similarly, departments did not generally use the ‘harder edged’ aspects of the accountabilitysystem, including non-renewal of contract, because they realised the difficulty of fully separating responsibility for strategic and operational performance.

The highest profile dispute about chief executive responsibilityoccurred when both these norms were broken and the ‘outsider’chief executive of the Prison Service clashed with a Home Secretaryseeking to avoid blame for poor performance. The dispute ended inthe chief executive being removed and some damage to the HomeSecretary’s reputation for competence. This case demonstrated thedifficulty of shifting blame for performance in areas of politicalsaliency. However, the Home Secretary did not resign and no otherSecretary of State has resigned for inadequacies in setting the strate-gic frameworks of executive agencies. These findings are consistentwith broader work on ministerial resignations which suggest thatministers did not, in general, resign because of poor departmen-tal performance (Dowding and Kang, 1998). In this context, execu-tive agency working was not a way for ministers to shift blame to chief executives of to executive agency enable them to resist pres-sure to resign. Such a reform was, in most cases, not necessary toavoid the risk of pressure to resign because of poor administrativeperformance.

Individual executive agencies’ economy, over the period as awhole, is more consistent with bureau-shaping than public interesthypotheses, as shown in Chapter 5. There was an improvement ineconomy during the late 1990s with real terms administrative costsfalling by a mean score of 4.4 per cent between 1995/96 and1997/98. However, the picture was mixed and, whilst non-tradingagencies’ administrative costs fell by 5.3 per cent, trading agenciescosts rose by 8.2 per cent. By 2001 in the Benefits Agency, despitefalls in the late 1990s, real terms administrative costs were 9 per centhigher, and real terms programme costs were 13 per cent higher, thanthey were at the start of executive agency working. There wereimprovements in some aspects of the effectiveness of individualexecutive agencies, although there was substantial variation between


bodies. According to their performance against targets, effectivenesswas satisfactory or above in 88 per cent of cases in the late 1990s.However, broader measures of effectiveness provide a less positivecharacterisation of the Benefits Agency’s work in paying the rightmoney to the right people at the right time. The Agency was notassociated with large improvements over the full range of its activi-ties; long-standing problems that pre-dated the Agency remained.There was fraud and error amounting to about 7 per cent of the socialsecurity programme in the mid-1990s and around a quarter of thoseentitled to the main benefits handled by the Agency did not receivethem. There was no demonstrable step change improvement in productive efficiency, contrary to the public interest expectation.

The semi-detached organisational structure enabled the BenefitsAgency to develop its own pay and grading systems and workingpractices to make better use of some resources, improving aspects of service to users, notably the speed of processing claims. The exter-nal performance target regime affected the Agency’s method of work-ing, with Secretary of State’s targets being reflected and developedthrough a system of internal management targets linked to perfor-mance pay for staff. Performance was improved by targeted effort insome areas but work that was not the principal focus of targets,including fraud and error in the early 1990s and the take-up of ben-efit, suffered as a result. The Department attempted to change theemphasis of the target regime to ameliorate poor performance infraud and error but these efforts were only partly successful. Not allactivities could be included in the target regime to inform manage-ment action, because such a regime would have been too broad touse as the basis for allocating effort. The tendency for ‘unbalanced’performance focused narrowly on targets was partially reduced bycivil servants in the Agency picking up on informal signals and notpursuing the logic of the target regime to its full extent. Ministersand staff in the Department co-operated in this practice, their assessment of the Agency’s performance was broader than simpleperformance against targets. However, throughout the period, thetargets were influential and principally reflected the objectives of the Agency that were prioritised by ministers rather than users orpotential users of the service. In part, this situation led to continuedinadequacy of some services and a very limited programme to boostthe take up of benefit.


The consequences of executive agencies for central governmentsystemic performance were, in general, more consistent with bureau-shaping than public interest expectations, as discussed in Chapter 6.Although there was an improvement in central government realterms administrative costs between 1995/96 and 1997/98, between1988 and 2001 overall real terms administrative expenditure rose 17 per cent and real terms total expenditure rose 32 per cent. Therewere substantial problems with public sector externalities, consistentwith bureau-shaping hypotheses. Executive agencies’ concern withtheir own activities led them to concentrate on e-government solu-tions with a narrow focus on their own users or customers. Thisapproach limited the development of e-government with systemicproperties to exploit the potential of the technology to achieve collective benefits across government. In the social security system,the Benefits Agency’s focus on its own targets and distinctive workingpractices contributed to significant negative public sector externali-ties. These externalities included harm to policymaking and imple-mentation because of poor information exchange between theAgency and the DSS Headquarters, which was particularly evident inthe state pensions failure. Inaccurate information generated by theAgency was used by other bodies in the system and there were unful-filled potential shared services, including the limited development ofsystemic e-government. The problems of systemic performancebecame even more salient after 1997 when the incoming New LabourGovernment changed the emphasis of policy towards employabilitywhich required more exchange of information in the system and aclient focused delivery system rather than one based on processingindividual benefits.

The Benefits Agency shared the two characteristic features oforganisational separation and an accountability regime with all otherexecutive agencies. In this respect, despite different working practicesand contexts, findings related to the Agency are relevant to othercases. The organisational barriers to e-government suggest that sev-eral of the systemic performance issues were common to executiveagencies. However, the problems of systemic performance were lessimportant in executive agencies which were less integrated into pro-grammes developed or delivered in co-operation with other parts ofthe public sector. These executive agencies were often peripheral,specialist, or had a focus on a clear external customer group and


formed about half the total number of bodies. Within this group,trading agencies, which were a relatively small percentage of the totalnumber of executive agencies, had substantial interaction with distinct customer groups outside of government, charging directlyfor their services. The operation of these structures and their effecton performance particularly merits further research.

The origin of the Next Steps reform, in part, as a ‘bureau-shaping’strategy explains why central government overlooked its previousexperience with similar arm’s-length forms which pointed to some ofthe performance problems that eventually emerged. The experiencewith nationalised, public corporations suggested difficulties of set-ting a framework which provided these bodies with challenging targets to improve performance and which prevented interventionby ministers and officials contrary to the management freedomsgranted to these bodies (Prosser, 1986, pp. 19–21). Similarly, the sys-temic performance problems were suggested by the long runningproblem of departmentalism in Whitehall. Departmentalisminvolved departments focusing too much attention on their ownconcerns to the exclusion of the broader consequences of their activ-ities. This feature of central government was widely recognised as aproblem by senior officials in the 1980s and 1990s (Marsh et al.,2001, pp. 124–30). Executive agencies intensified departmentalismby strengthening incentives for these bodies to respond narrowly to their parent departments’ concerns and further fragmented thesystem by splitting up departments between policy sections and different executive agencies. However, the Next Steps report did notdirectly tackle the issue of the trade-off between improving individ-ual and systemic performance.

The findings about executive agencies contribute to an emergingbody of knowledge about New Public Management (NPM) that sug-gests some common themes across different components of reformin the UK. The importance of senior officials and politicians in devel-oping and directing reform was a characteristic shared with NPMreforms in health, education, housing and social care. The reforms toset up corporate units with more performance measures and targetsinvolved a minimum of consultation with actors outside of central government, except a few policy advisers (Bartlett et al., 1998,pp. 1–16; Pollitt et al., 1998, pp. 48–53). As with executive agencies,centrally set performance measures and ministerial appointment of


members to run units reflected accountability systems dominated byexecutive politicians rather than service users or citizens (Ferlie et al.,1998, pp. 204–10). Some of the side and perverse effects on perfor-mance were shared with other NPM structures (James, 2002). The corporate hospital and school units often appeared to have improvedtheir own performance but created problems of co-ordination and created difficulties for each other including harming the provision ofjoint services (Pollitt et al., 1998, pp. 74–5, 117–19). The unintendedconsequences of performance targets in focusing attention on whatwas targeted to the detriment of other valued outcomes was anincreasingly salient issue across the public sector (Smith, 1995; Benyonand Edwards, 1999). Executive agencies brought these performanceissues to the centre of government in Whitehall.

Section 2: Prospects for the future use of executive agencies

A complication in assessing the future use of executive agencies inthe UK is the devolution of some responsibilities to Scotland andWales in 1999 and, in an on/off manner since that time, to NorthernIreland. Of the 127 executive agencies at the start of 2002, ninety-two bodies reported to departments in England, twenty-one to theNorthern Ireland Executive, twelve to the Scottish Executive and twoto the Welsh Assembly (Office of Public Services Reform and HMTreasury, 2002, p. 60). The officials in the Scottish Executive and theWelsh Assembly remained civil servants after devolution, like thosein England, but it is possible that, if more substantially differentpolitical and administrative arrangements develop, bodies in theseareas will follow their own trajectory in the future. Northern Irelandexecutive agencies were staffed by the Northern Ireland Civil Serviceprior to 1999 and, in this sense, were always distinctive. However, ifdevolution has a chance to bed down, these bodies may develop fur-ther distinctive characteristics and a separate trajectory as a group.

The public interest perspective suggests that executive agencies arelikely to continue to be politicians’ preferred form for delivering cen-tral government services. The bureau-shaping perspective suggeststhat the forms will be stable as long as politicians do not reverse theirconcern with the management of executive activity, which couldtrigger bureau-shaping in the opposite direction. For example, theconstraint could shift if politicians loose interest in managerialism


relative to other priorities for government that compete for theirattention. Whilst many of the findings in Part II were consistent withthe bureau-shaping hypotheses, some factors affecting the course ofreform were beyond those suggested by either perspective. These fac-tors included the career incentives of officials in central units tocome up with successful reform initiatives and ‘pull’ factors fromstaff in some specialist executive agencies who desired greater free-dom. These pressures suggest that the executive agency model maybe unstable in some, limited, contexts. More broadly, the systemicperformance problems offer an opportunity for officials in centralunits to enhance their careers by suggesting additional structures toameliorate the situation.

Two possible fundamental changes to executive agency workingare to move away from the semi-detached organisational form byeither rolling executive agencies back into departments or increasingtheir autonomy from departments. The full reintegration of execu-tive agencies into their departments seems unlikely except in a fewcases. There were a few examples of such a change in the late 1990swith the Securities and Facilities Executive and the Civil ServiceCollege being integrated into the Central Management and PolicyStudies section within the Cabinet Office in 1998 and 1999 respec-tively. In the latter case, the reorganisation was intended to improvethe links between training courses provided by the College andstrategic personnel and other issues (Centre for Management andPolicy Studies, 2001, pp. 2–5).

The pressure for more autonomy appears strongest in trading agen-cies which raise much of their revenue from distinct customer groupsand in specialist research facilities staffed by non-mainstream offi-cials. Demands for more autonomy by staff to respond to clientgroups were important causes of passing executive agencies to pri-vate ownership during the 1990s (Gains, 1999, p. 713). The changesin bodies that were privatised nearly always involved considerableinput from the staff in the executive agency rather than externalpressure (I1, Senior Official, Cabinet Office). Of the sixteen cases ofprivatisation, twelve were sold to private companies, two cases weremanagement buy-outs, and two bodies had their managementwholly contracted out to the private sector, as shown in Table 7.1.

In a central government system with a broadly stable set of execu-tive agencies, issues of systemic performance are likely to be tackledby forms of co-ordination or central steering of activities, as outlined


in Figure 7.1. The vertical dimension is the degree of central steeringof the system by central bodies, ranging from strong steering by a cen-tral regulatory authority to weak central steering leaving individualexecutive agencies to their own devices. The horizontal dimensionrelates to the strength of arrangements between individual organisa-tions for co-ordinating their activity, ranging from strong and endur-ing relationships in a network to weaker, sporadic, links. The fouroptions in Figure 7.1 are combinations of positions at opposite endsof the two dimensions: sporadic links, networks, regulated sporadiclinks and regulated networks. All four forms were adopted to someextent in the late 1990s, although a trend is emerging towards moreuse of regulated networks relative to other forms, especially in thesocial security sector.

Sporadic links

In this arrangement, co-ordination between executive agencies is amatter for voluntary agreement between bodies to exploit particularopportunities rather than being part of longer term relationships.The links rest on mutually beneficial exchange and there is no use ofauthority from higher level units although these units may set up incentives to encourage executive agencies to co-ordinate theiractivities. The Treasury established new measures to ‘join-up’ central


Table 7.1 Privatisation of executive agencies

Executive agency (with date of change)

Twelve sold to Accounts Services Agency (1995), Agricultural private Development and Advisory Service (1997), companies Chessington Computer Centre (1996), DVOIT (1993),

HMSO (1996), National Engineering Laboratory (1995), Natural Resources Institute (1996), Occupational Health and Safety Agency (1996), Paymaster (1997), Recruitment and Assessment Services (1996), TransportResearch Laboratory (1996), Warren Spring Laboratory (1994)

Two management Building Research Establishment (1997), Laboratory ofbuy-outs the Government Chemist (1996)

Two contracted National Physical Laboratory (1995), Teachers’ Pensionout management Agency (1996)

Source: Cabinet Office (1997b, 1998a, 2002b).

government by providing budgetary incentives to finance projectsinvolving joint working in the late 1990s. The £230 m per year Investto Save budget funded innovative projects involving two or moredepartments and the £2500 m Capital Modernisation Fund wasawarded, in part, to joint projects (PIU, 2000, box 9.1). In theBenefits Agency, an Invest to Save project developed electronic kiosksto present information about a range of services in a variety of loca-tions outside of the Agency’s premises (I13, Senior Official, BenefitsAgency).

The Cabinet Office encouraged executive agencies to participate ina voluntary scheme to benchmark their performance against eachother and private sector organisations from 1995. Executive agenciesundertook self assessment rather than formal, external, assessmentbut it was hoped they might learn from best practice in other organ-isations and that comparisons between bodies might create peer pres-sure to drive up performance (Samuels, 1997, p. 3). The first phaseinvolved thirty executive agencies in 1995–96, a second phase ranfrom April 1997 to January 1998 and included seventy bodies. Athird phase was introduced to run for three years from 1998 involv-ing further executive agencies (Samuels, 1998, p. 12). The BenefitsAgency participated from 1996 and the commitment of senior staffwas reflected in the chief executive’s membership of the governing


Strong

Centralsteering

Weak

Weak Horizontal links Strong

Networks e.g. Woking Ageagency (merged BenefitsAgency and EmploymentService) running theJobcentre Plus business

Regulatedsporadic linkse.g. departmental PublicService Agreements,Benefit Fraud Inspectorate

Sporadiclinks e.g. benchmarking,occasional co-operationbetween BenefitsAgency and other bodies

Regulatednetworks e.g. cross-cuttingPublic Service Agreements andshared targets, Department forWork and Pensionsoversight of its ‘businesses’

Figure 7.1 Structures for co-ordinating and steering executive agencies

board of the European Quality Foundation which ran the scheme(Social Security Committee, 1998, question 153).

The contribution of benchmarking to solving problems of systemicperformance was, however, not as great as its role in raising perfor-mance within the organisations. An analysis of improvement plansresulting from the process (Samuels, 1997, Annex C) reveals thattwo-thirds of the improvements were focused on changes to internalprocesses with only one-third focused on external relations with cus-tomers or other parts of government. The focus on external clientswas most apparent in trading agencies. The voluntary nature ofscheme made it difficult to use to resolve public sector externalitiesbetween bodies if they could not agree amongst themselves on thebest course of action.

A possible way to facilitate voluntary co-ordination that has have not yet been tried is to encourage bargaining to remove publicsector externalities (James, 2000, p. 341). This solution suggests thataffected bodies could pay the producers of externalities to stop pro-ducing negative externalities as an alternative to regulation from ahigher level body. The scheme is analogous to the general solution tothe problem of externalities in markets when bargaining is withoutsignificant cost and there are clearly defined property rights (Coase,1960). In this general case, those suffering the effects of an external-ity band together to try pay the people causing the externality tochange their behaviour. For example, people living near a factorywhich pollutes their air offer to pay the factory to use an alternative,non-polluting, production technique. The allocation of propertyrights determines who compensates whom. If local residents do nothave the right to clean air then they pay the factory to stop pollut-ing. In the case of the Benefits Agency, the organisation workedtowards its own targets and, as a side effect, pushed up fraud in localauthorities’ administration of Housing Benefit by producing faultyinformation which was then used by them in their processing ofcases. A hypothetical bargaining solution to this negative public sec-tor externality would involve the department deciding whether localauthorities have the right to accurate information or whether theymust pay the Benefits Agency to get it. The drawing up of contractsinvolving exchanges of budget for information could then be left tothe local Benefits Agency offices and local authorities on the ground.These bodies are likely to be in a better position to know the value of


the accurate information to their work with, for example, the valuediffering depending on levels of attempted fraud in the locality.Provided the costs of arranging these contracts are not too great, theycould be used to deal with the externality in a potentially less costlyway than a regulatory solution entailing a higher level body havingto intervene to insist on the Agency improving the provision ofinformation.

Networks

Networks are an alternative to sporadic links consisting of moreenduring partnerships between organisations including professionalties between individuals in different organisations, staff exchangeand arrangements for managing common interests. There are a vari-ety of perspectives on how networks and partnerships operate in thepublic sector (Kickert et al., 1997; Bardach, 1998). UK central gov-ernment increasingly used several different forms of networks in thecontext of executive agency working.

The strongest form of network organisation was the horizontalmerger of executive agencies, at the same time maintaining the vertical separation between departments and these bodies. Abouttwenty executive agencies were merged with other bodies since the reform began, although the picture is complicated by occasionalde-mergers. The executive agencies involved in horizontal mergerswere the Central Computer and Telecommunications Agency,Central Science Laboratory, Central Statistical Office, Coastguard,Food Science Laboratories, Marine Safety Agency, Property Advisersto the Civil Estate, The Buying Agency and thirteen bodies in theMinistry of Defence (Cabinet Office, 1998a; Minister for the CabinetOffice, 2000, p. 9; Cabinet Office, 2002b). The rationales for thesemergers included avoiding duplication, achieving economies of scaleor better co-ordination. The reasoning is illustrated by the case of the creation of the Office of Government Commerce, which broughttogether bodies including The Buying Agency and the PropertyAdvisers to the Civil Estate. The main justification for the change wasto create a ‘one-stop shop’ for procurement in central governmentand to counter ineffectiveness and inefficiency from fragmented procurement systems, especially in IT. Individual departments andexecutive agencies did not develop this expertise themselves becausethey only occasionally needed to engage in particular forms of


procurement. The Office was further intended to ameliorate theproblem of individual departments and executive agencies commis-sioning IT and other systems without sufficient regard of the need tohave compatibility with systems in other bodies which was a barrierto the development of systemic e-government.

There were attempts to improve the integration of careers across dif-ferent departments and executive agencies. In general, civil servantsdid not think that moving to an executive agency would assist theirpromotion to the most senior grades (Trosa, 1994, pp. 48–9). TheSenior Civil Service, set up in 1996, was intended to encourage civilservants to take a broader perspective than the body in which theyworked and to promote the movement of staff across central gov-ernment bodies. The Civil Service Management Committee, laterrenamed the Management Board, actively managed the top 600 post-ings. The Cabinet Secretary suggested that, to reach the upper ranks of the Service, staff needed experience in working in frontline deliveryor operational management and of working in more than one organ-isational culture (Wilson, 1999, section 16; PIU, 2000, section 4.11).However, in most departments, there was no ‘critical mass’ of civil servants with extensive experience of working in executive agencies.The percentage of staff moving on loan outside their own departmentwas just under 7 per cent each year in the 1990s (PIU 2000, figure 8.2).

Regulated sporadic links and regulated networks

In a broad definition, regulation is often used as a synonym for control. However, in this section, regulation is used to mean a bodyusing some form of authority to steer executive agencies strategicallyat arms-length. Regulation of sporadic links involves steering execu-tive agencies that occasionally co-ordinate activity amongst them-selves and contrasts with the regulation of networks, where there are more extensive and enduring links between executive agencies.In the UK public sector as a whole, there was an increase in theresources devoted to the use of regulation of public bodies along-side New Public Management changes in the 1980s and 1990s (Hood et al., 1998, 1999, 2000). In central government, whilst centralregulation of pay, grading and management practices were relaxedalongside the move to executive agency working, new or reformedregulatory systems emerged. Indeed, the executive agency modelitself incorporates a form of regulation of sporadic links or networks


in the relationship between a parent department and the executiveagencies within its direct purview.

The departmental network form of regulation was used to sort outsome disputes between executive agencies in the DSS, where the executive agencies and the Headquarters were all represented on theDepartmental Management Board. Matters of common interest were discussed by the Board, and Headquarters staff and ministersoccasionally intervened to put pressure on executive agencies to work more co-operatively (I14, Senior Official, Department of SocialSecurity). Departmentally led initiatives to encourage longer term co-operation and network working were also increasingly pursued inother departments. One of the most developed forms of regulatednetwork was the Driver Vehicle Operator group established in 1999 tolink the four executive agencies handling matters related to vehicletransportation in the Department for Transport, Local Governmentand the Regions. The group was chaired by a departmental officialand included all four chief executives. According to officials involvedin the initiative, executive agencies’ independence of action was curtailed by requiring them to discuss and, if at all possible, to takeinto account the effects of their action on other bodies (Kutlu, 2001,pp. 167–8). The group assisted the development of common servicesincluding a joint website for interacting with service users (Office ofPublic Services Reform and HM Tresury, 2002, p. 25).

The central units of the Cabinet Office and Treasury intervened to improve co-ordination between executive agencies, but this wasnormally to facilitate sporadic links rather than to set up and influ-ence networks of bodies. The Cabinet Office did not generally inter-vene using formal authority and, where more formal systems wereused, they were predominantly focused on improving individualbodies’ performance rather than systemic performance. The CabinetOffice required executive agencies to gain accreditation under theInvestors in People scheme, demonstrating that they had systems for training employees and developing skills. By 1998, thirty-twoexecutive agencies had been fully accredited and all were expected to undergo the process by 2000 (Cabinet Office, 1998a, pp. 8–9). By April 1999, all Benefits Agency staff worked in units recognised by Investors in People (DSS, 1999a, p. 57). However, the Investors in People scheme was largely focused on individual executive agen-cies’ performance rather than systemic performance. Similarly, the


Citizen’s Charter scheme, later renamed Service First, required exec-utive agencies to develop their own standards for services, monitorperformance against them and raise performance if it was unaccept-able. The standards were mostly focused on individual bodies ratherthan standards for services involving the work of several organisa-tions (IA7, Senior Official, Cabinet Office). In the Benefits Agency,these central standards influenced the Agency’s regime and the Unitwas able to block proposals by the Agency if it felt they were not set-ting the right standards (I13, Senior Official, Benefits Agency). Forexample, some targets were initially set to cover 60–70 per cent ofcustomers. However, the Citizen’s Charter Unit recommended thatthe quality of service should be achievable for 100 per cent of clients.To fulfil this requirement, the Agency had to lower the standard setto have a realistic chance of achieving the higher percentage ofclients covered (NAO, 1998b, p. 44).

The Cabinet Office produced annual public reports collating theperformance of executive agencies as a group, including performanceagainst targets. However, reporting concentrated on individual per-formance rather than developing measures of performance in sectorsof government involving the activities of more than one executiveagency. Whilst the Cabinet Office attempted to set up comparativemeasures of performance these were not fully developed because ofdifficulties in creating measures that were fully comparable acrossdifferent bodies. Instead, voluntary co-operation with the bench-marking scheme was encouraged and there was no requirement thatexecutive agencies should participate (I3, Senior Official, CabinetOffice).

The Treasury kept strategic oversight and control centred on thepublic expenditure system although much of the responsibility fordealing with executive agencies was handled by departments ratherthan involving Treasury intervention (IA1, Senior Official, HMTreasury). A more systematic approach to regulating central govern-ment as a network, including executive agencies, emerged in the late1990s with the Public Service Agreements (PSAs) regime. PSAs arosefrom the Comprehensive Spending Review of 1998 and wereintended to provide an overall umbrella for setting and monitoringdepartmental and executive agency objectives, targets, outputs andoutcomes (Chief Secretary to the Treasury, 1998, pp. 1–3). PSAs arepublic documents and departments have to report progress against


targets to the Ministerial Committee on Public Services and PublicExpenditure (PSX) (PIU, 2000, section 7.18).

Whilst most PSAs relate to individual departments, there was anattempt to improve the treatment of policies and services straddlingorganisational boundaries. Departmental PSAs include a small pro-portion of objectives and targets that are shared by more than onedepartment and its executive agencies. There are ‘cross-cutting PSAs’containing sets of joint objectives and targets for programmes in afew areas. There were initially twenty-eight PSAs based on depart-mental boundaries with three cross-cutting PSAs in criminal justice,drugs and help for young families. In total, there were fifty targetsheld jointly by two or more departments out of several hundred and PSAs predominantly reflected separate departmental processes(PIU, 2000, box 9.1). However, the PSA regime better integrated theBenefits Agency’s performance plans with broader systems. TheAgency’s targets were incorporated in the broader Department ofSocial Security PSA and the body instituted a new PerformanceManagement Regime from April 2000 to take account of this devel-opment. The shared objectives and targets resembled the arrange-ments already in place between the Agency and the EmploymentService for delivering Jobseeker’s Allowance. But there was a limit tohow far targets could capture all the elements of working with otherbodies which, combined with multiple and conflicting targets, lim-ited the extent to which this approach facilitated better co-operativeworking (I14, Senior Official, Department of Social Security).

There were attempts to increase the regulation of executive agen-cies by bodies other than departments and central units. TheNational Audit Office (NAO) was important in highlighting some ofthe systemic problems of performance that affected the ‘value formoney’ of government. Whilst the resources used in financial auditwere largely static, the NAO directed more resources to ‘value formoney’ studies in the 1990s (Hood et al., 1999, pp. 87–9). The NAOrevealed the effect of inaccurate information supplied by the BenefitsAgency’s on the administration of Housing Benefit (NAO, 1997b).However, it was difficult for NAO to get bodies in central governmentto tackle these problems because it had no direct authority to man-date action. Instead the organisation operated primarily through amixture of informal pressure and publicity (I5, Senior Official,National Audit Office; I12, Senior Official, National Audit Office).


The Benefits Agency was subject to more regulation than mostexecutive agencies and had a specific regulator, the Benefit FraudInspectorate (BFI), set up in 1997 to monitor the Agency and bodiesworking with it in the social security system. The BFI had a mandateto examine individual bodies’ performance and to take a systemicperspective, in particular on the level and causes of fraud and error in the system (Secretary of State for Social Security and ChiefSecretary to the Treasury, 1999, p. 22). The Inspectorate cost £6.5 mto run in 1998/99 and employed 133 staff, a substantial investmentin regulation (BFI, 1999b). The organisation was successful in uncov-ering the negative externalities affecting local authorities and theContributions Agency (BFI, 1999d). However, it proved more diffi-cult for BFI staff to bring about improvements, not having a linemanagement relationship with the Agency or other bodies in the system and being unable to impose more extensive network basedforms of working on the system (I11, Senior Official, Benefit FraudInspectorate).

A more substantial change towards the regulated network form oforganisation in social security was developed from the late 1990s.The form was developed to ameliorate the problems encountered inthe ONE pilot projects and better to fulfil the developing policyagenda of Welfare to Work which called for a more client focuseddelivery structure (I15, Senior Official, Benefits Agency). The pressurefor the change came in part from entrepreneurial officials in theTreasury (I15, Senior Official, Benefits Agency). There was a cross-cutting review of Welfare to Work and ONE in the 2000 SpendingReview which was led by a Treasury official and reported to theEconomic Affairs, Welfare to Work (EA (WW)) and Public Spending(PSX) cabinet committees, which were both chaired by theChancellor. Following the review, there was a major reorganisation tocreate the structures summarised in Figure 7.2. The Department forWork and Pensions (DWP) replaced the Department for SocialSecurity (DSS) in 2001. The Benefits Agency and the EmploymentService were merged to create a ‘working age’ agency operating theJobcentre Plus ‘business’ (Benefits Agency, 2002, pp. 5–9). JobcentrePlus integrated the delivery structures for providing benefit pay-ments and job services for the ‘working age’ client group whilst thePension Service took over responsibility for these aspects of BenefitsAgency work. The other businesses in the Department were the


Disability Carer Service, the Child Support Agency and the AppealsService, the latter two being formed from existing executive agencies.

The changes involved reallocation of a few, relatively small, strate-gic tasks, from the executive agencies to the Department’s CorporateStrategy and Shared Services section but maintained a semi-detachedarrangement between policy development and the implementationof programmes. This reform continued a trend established in 2000,when the relatively small strategic parts of the Information TechnologyServices Agency were reintegrated into the DSS Headquarters at thesame time as the bulk of IT work was let to a private firm, ElectronicData Systems, under contract. The individual executive agencies’Framework Documents were replaced with an overall managementframework for the Department and its new ‘businesses’. The ‘busi-nesses’ remained semi-detached organisations, each headed by achief executive and separate from those primarily responsible for pol-icy work and ministerial support located in the Corporate Strategyand Shared Services section. However, the systems were intended toimprove communication between these organisations, to encouragethe provision of shared services and to bring about greater staffmovement between the ‘businesses’ and the centre (I14, SeniorOfficial, Department of Social Security; Lomax, 2002).

The recent changes marked a softening of the harder edges of theexecutive agency performance regime, with talk of ‘linked commu-nities’ and a ‘shared agenda’ in the Department rather than a more


Targets and budgetPerformanceinformation

Servicesand

payments

Department of Work and PensionsCorporate Strategy and Shared Services:

human resources, IT, estates,communications, legal, analytical

Client Groups:Working age and

childrenPensions and disability

Businesses:Jobcentre Plus

The Child Support AgencyThe Pension Service

Disability Carer ServiceThe Appeals Service

Figure 7.2 Department for Work and Pensions

simple divide between setting/monitoring targets and delivery(Lomax, 2002). In the Corporate Strategy and Shared Services sec-tion, new Client Director posts were established to focus on clientneeds and specify service priorities and targets. As under the BenefitsAgency arrangements, there was regime of targets for the differentbusinesses set by the Secretary of State linked to the PSA for theDepartment. Each business chief executive continued to be responsi-ble for achieving the targets for their own business. However, thereform recognised more explicitly that a broader consideration ofperformance beyond quantifiable performance against targets wasrequired to steer the system (I9, Senior Official, Department of SocialSecurity).

It is too early to judge the effect of these structures on workingpractices but they involve a softening of the executive agency modelrather than its replacement. The systemic performance consequencesof the less ‘hard edged’ performance target regime, which furtherdeveloped the informal practices operated during the time of theBenefits Agency, might reduce the incentive for staff to concentrateoverly narrowly on targets to the detriment of other bodies’ perfor-mance. Similarly, shared services may be more easily developed andmaintained than in the more fragmented system. However, thearrangements do not seem to alter the dominance of ministers in set-ting priorities. Whilst the rhetoric is of ‘client focused’ services, thesystem appears to leave ministers rather than clients in the positionof judging what clients want.

Section 3: Relevance of the perspectives and the findings for countries emulating the UK reform

The hypotheses drawn from the two perspectives offer a set of expec-tations about executive agencies that are of relevance to other coun-tries’ central government systems. To the extent that the hypothesesreceiving empirical support from the UK case receive further supportin these other cases they provide the basis for generalisation ofknowledge about executive agencies. A feature of executive agencyreform, consistent with broader NPM reform, was the internationalinfluence of organisational models and experiences, with countriesattempting to learn from reform ideas and practices in other juris-dictions (Pollitt et al., 2000; James, 2001). The public management


section of the OECD devoted considerable attention to executiveagency reform in briefings and study reports circulated to membercountries. The UK was just one of several countries’ experiences thatwere noted and other countries, particularly Sweden, have a long andwell known history of the widespread operation of similar structures(OECD, 1995, 1998, 2002).

A good starting point for exploring the perspectives’ hypotheses in different countries would be to examine OECD countries with, inbroad terms, similar levels of governmental development, thatembarked on reforms directly influenced by the UK experience. TheUK Cabinet Office hosted delegations of officials and politicians fromseveral countries who expressed an interest in the Next Steps reform(Goldsworthy, 1991, pp. i–ii). The Canadian Government explicitlydrew on the UK experience in creating Special Operating Agenciesfrom 1989, with seventeen bodies up by the end of 1996 (Thomas,1996, pp. 1–2). In US Federal Government, proposals for PerformanceBased Organisations (PBO) were modelled, in part, on the UK initia-tive (NPR, 1996) and the Netherlands set up over twenty departmen-tal agencies between 1991 and 1998, in part drawing on the UKreform (Pollitt et al., 2000). Executive agencies were adopted in Koreafrom 1999 in a reform influenced by the changes in the UK, althoughthe bodies that emerged were substantially altered to fit local circum-stances (Yoon, 2001). Japan emulated the UK initiative through amajor reform to create Independent Administrative Institutions (IAI)from 2000 (Kaneko, 1999; Matsuda, 1999; Hirose, 2000; Hori, 2002).

The consistency of the hypotheses with developments in thesecountries is likely to be limited by differences in the context ofreform, including the extent to which central government providespublic services in a broadly similar way to the UK. Their potential relevance can be illustrated in the context of the US and Japan,although a full evaluation would require much more extensiveresearch. The executive agency model offers a benchmark for com-paring reforms in different countries. The potential for executiveagency reform in US Federal Government was lower than in the UKbecause many services were not performed at this level of govern-ment including, for example, driver and vehicle licensing, vehiclecertification and fire service training which were all handled by exec-utive agencies in the UK. The potential for use of executive agencieswas further reduced by the existence of similar structures pre-dating


the PBO proposals. The system was already organisationally fragmented and the Government Performance and Results Act of1993 had already required Federal bodies to develop strategic plans,performance measures, annual performance plans, and performancereporting (NPR, 1999). However, PBO were very similar to the execu-tive agency model, involving separating service operation functionsfrom their policy components and placing them in separate organi-sations reporting to departments. The organisations were given free-dom to manage personnel, procurement and other services and wereheaded by chief executives, hired on a fixed term contract through acompetitive search. Each organisation was given a three- to five-yearframework with measurable goals and targets for improvement andthe chief executives were held accountable for performance. Initially,seven functions were proposed for PBO status in 1997. However, the reform did not look like becoming such a significant change asthe executive agency reform in the UK and only a few bodies were setup by the late 1990s (Friel, 1999).

In contrast to the US, the reform to create IndependentAdministrative Institutions (IAIs) in Japan looks like a major change. The activities handled by central government were moresimilar to those in the UK than in the US and there was less use offormal performance measures, so the potential for an executiveagency reform was greater. Setting up IAIs involved separating strate-gic ‘policy’ functions from ‘policy execution’, with better corporatefinancial management and more explicit systems of performanceplanning and evaluation. Each IAI had a degree of management free-dom with an arms length relationship with ministries but, unlike UKexecutive agencies, they were not formally part of central govern-ment departments and in this respect they resembled Non-Departmental Public Bodies. Progress was rapid; fifty-six bodies wereset up in April 2000 with functions transferred including researchactivities, mint and printing operations and operational activitiessuch as motor vehicle inspection. Further IAIs were planned inhigher education (Hori, 2002, p. 10). Using the executive agencymodel as a benchmark, a study of Japan could explore whether thesechanges are ‘pure’, ‘mitigated’, or ‘nominal’ to assess the significanceof the reform.

The public interest and bureau-shaping perspectives offer contrast-ing hypotheses explaining the reforms, although both perspectives


hypothesise that actors within the administrative and political executive are important influences on the course of reform. Researchcould examine the hypotheses in the US and Japanese contexts. Thepublic interest perspective’s hypothesis that politicians lead execu-tive agency reform appears to have some support in the US where thePresident, Vice-president and presidential appointees had particularprominence in developing reforms to the executive in the 1990s. TheNational Performance Review (NPR) run by the Vice-president pro-posed a wide ranging reform programme on the basis of a number ofstudies including observation of the UK experience (NPR, 1996).However, contrary to both perspectives, legislative as well as execu-tive politicians were influential in decision making about reform.Congress formally created all executive departments including PBOs,determined the funding of programmes and had the power to reor-ganise government bodies. Any bureau-shaping influences on reformseem likely to have operated differently in the US to the UK. Whilstthe US Senior Executive Service was a set of officials pursuing careersin Federal Government, their influence was reduced by the role ofpresidential appointees who constituted about 10 per cent of 7000senior jobs, especially the most senior posts (Fesler and Kettl, 1996,pp. 179–213). Research could examine the institutional arrange-ment of legislative and executive politicians, political appointees andcareer officials, the effect on bureau-shaping strategies and the limited adoption of PBOs.

In Japan, both perspectives appear consistent with the dominanceof actors within central government in initiating and developing thereform. The public interest perspective appears consistent with min-isters’ contribution to the development of IAIs through theAdministrative Reform Council, which reported in December 1997.The reform got strong central direction in 1998 with the establish-ment of the Headquarters for Central Government reform, headed bythe Prime Minister and composed of all the Ministers of State. Whilstthere was legislation to create the new bodies, the role of legislativepoliticians appeared very different from that in the US because ofparty based control in a Parliamentary system (Kaneko, 1999, p. 1).The potential for senior officials to embark on bureau-shaping strate-gies to bring about reform appeared to be high, like the UK ratherthan the US. There was a permanent elite civil service that was tradi-tionally very influential in executive decision making (Muramatsu


and Krauss, 1996, p. 241). Further research could assess how politiciansand officials interacted to bring about a major reform in this context.

Finally, the public interest and bureau-shaping perspectives offercontrasting hypotheses about the likely consequences of adoptingexecutive agencies. The bureau-shaping perspective’s hypothesesreceive support in the UK and, whilst the different contexts are likely to have a substantial intervening effect on performance out-comes, the findings suggest that a large majority of individual exec-utive agencies are likely to be reasonably effective. However, ademonstrable, step change, improvement in effectiveness or produc-tive efficiency over the old systems is unlikely and executive agenciesare likely to be associated with worsening economy, except for shortperiods of time. Adopting executive agencies is likely to contribute tosubstantial systemic performance problems through public sectorexternalities in cases similar to the Benefits Agency involving working closely with other bodies. However, the net balances intrade-offs between changes to systemic performance and possibleimprovements in individual executive agencies’ performance arelikely to be better in bodies that are more peripheral to broader government activity.


Appendices

Appendix 1: The Bureau-ShapingModel

The bureau-shaping perspective on executive agencies draws extensively onthe bureau-shaping model (Dunleavy, 1985, 1986, 1989a,b, 1991) by incor-porating the insight that officials have a preference for policy work ratherthan routine work and the concept of bureau-shaping strategies. Dunleavydeveloped the bureau-shaping model in two parts. First, the bureau-shapingtypology of bureaus, based on differences in the mix of budgets in a bureau.Organisations with a high proportion of ‘core budget’ spent on their ownactivities are ‘delivery’ bureau whilst organisations passing on budget for others to spend are ‘transfer’, ‘contracts’ or ‘control’ bureau, depending onwho receives the money (Dunleavy, 1991, pp. 183–8). The typology wasapplied to describe the ‘architecture’ of the British state in the late 1980s(Dunleavy, 1989a,b).

The second part of the bureau-shaping model is a general theory of admin-istrative reform which suggests that rational officials value work related util-ity and undertake ‘bureau-shaping’ strategies in pursuit of these benefits(Dunleavy, 1985, 1986, 1991, pp. 200–9). Several bureau-shaping strategiesare suggested including major internal reorganisations to promote policywork over routine activities, transformations of internal work practices, redefinition of relations with external partners to enhance policy contacts,competition with other bureaux to protect the scope of interesting work, loadshedding, hiving off and contracting out functions which are seen as unde-sirable (Dunleavy, 1991, pp. 203–4). The bureau-shaping model has beendeveloped as an explanation of privatisation and de-institutionalisation inthe public sector (Dunleavy, 1986, 1991, pp. 236–48) to examine reform of afinance ministry (Wallis and Dollery, 2001), changes in the Australian Federalbudget sector (Dollery and Hamburger, 1986), changes in local governmentincluding contracting out (Biggs and Dunleavy, 1995; Cope, 1995; Aulich,1999) and privatisation of government owned oil industries (Hoopes, 1997).

153

Appendix 2: The Formal Bureau-Shaping Perspective onExecutive Agency Reform

The preferences of an individual senior official for the organisation of his orher department can be expressed formally and is analogous to the simple economic model of consumer preference subject to a budget constraint. Forthe sake of clarity, the figures in Chapter 2 are drawn schematically. The pref-erences of an individual senior official are represented by the utility function

U � X1–2 � Y

1–2

where U is level of utility, X is budget per senior official and Y is proportionof policy work time in total work time.

The constraint on the official imposed by politicians is

Y � 1 � aX

where 1 is the limited work time available to the senior official, a is the pro-portion of management time in total work time which politicians insist anofficial must spend supervising each unit of budget per official and 0 � a � 1.

The slope of each indifference curve is given by the marginal utility of Xdivided by the marginal utility of Y

dY—dX

��1–2 X�

1–2�/�1–2 Y

�1–2�� Y

1–2/X

1–2

The slope of the constraint is given by

dY—dX

��a

At the interior optimum, the slopes of the constraint and indifference curveare equal

Y1/2/X1/2 � a, or Y � a2X, which can be solved with the constraint to yieldthe optimum values of X and Y,

Y* �

X* �

11�(1/a)

1a (1�a)

154

The function used in the formal perspective assumes independence of themarginal utilities of X and Y implying that policy work time and budget persenior official affect utility only by the sum of their separate contributionsand not by forming particularly desirable or undesirable combinations. Thelevel of policy work is largely independent of the level of budget per seniorofficial. A negligible level of budget per senior official is required to undertakepolicy work compared to that involved in executive activity because policywork tends to be more labour intensive in terms of senior officials and lessintensive in other resources than executive activity. The perspective assumesa stable relationship between changes in executive activity and budget persenior official and between changes in the level of policy work and propor-tion of policy work time. Stability requires that the number of senior officialsand efficiency in the use of budget and policy work time are constant.Without these assumptions budget per senior official, for example, could be reduced by increased efficiency leaving the level of executive activityunchanged. The tightening of politicians’ constraint on senior officialsshown in Figure 2.3 is expressed formally by a rise in a. The equations foroptimum X and Y imply that a rise in a will cause a fall in X and a rise in Y.The rise in a causes a rise in Y because the substitution effect of a rise in a outweigh the income effect which reduces the available ‘income’ for bothX and Y.

Appendix 2 155

Appendix 3: InterviewsConducted for the Study

The interviews were conducted on a Chatham House basis such that interviewees’ comments are not attributed to them by name. The first set ofinterviews are coded ‘Ix’ and were conducted exclusively for this study. Thesecond set of interviews were conducted jointly with the ESRC funded‘Bureaucratic Gamekeeping’ project and are coded IAx. For this latter set ofinterviews, the interview number in brackets indicates the ‘BureaucraticGamekeeping’ project code (Hood et al., 1999, pp. 228–30).

Interviews

I1 9/8/97 Senior Official, Cabinet OfficeI2 9/8/97 Senior Official, Cabinet OfficeI3 9/8/97 Senior Official, Cabinet OfficeI4 30/3/00 Official, Benefits AgencyI5 30/3/00 Official, National Audit OfficeI6 4/4/00 Senior Official, National Audit OfficeI7 4/4/00 Senior Official, National Audit OfficeI8 4/4/00 Senior Official, National Audit OfficeI9 14/4/00 Senior Official, Department of Social SecurityI10 10/5/00 Senior Official, Benefits AgencyI11 16/5/00 Senior Official, Benefit Fraud InspectorateI12 17/5/00 Senior Official, National Audit OfficeI13 1/6/00 Senior Official, Benefits AgencyI14 15/6/00 Senior Official, Department of Social SecurityI15 20/7/00 Senior Official, Benefits AgencyI16 24/10/01 Senior Official, Cabinet Office

Additional interviews conducted jointly with the ESRC funded ‘Bureaucratic Gamekeeping’ project

IA1 11/1/96 (I7) Senior Official, HM TreasuryIA2 16/1/96 (I8) Senior Official, HM TreasuryIA3 18/1/96 (I9) Senior Official, HM TreasuryIA4 12/2/96 (I13) Senior Official, HM TreasuryIA5 13/2/96 (I14) Senior Official, HM TreasuryIA6 20/2/96 (I15) Senior Official, National Audit OfficeIA7 22/2/96 (I16) Senior Official, Cabinet OfficeIA8 26/2/96 (I17) Senior Official, HM Treasury

156

IA9 1/3/96 (I18) Senior Official, Cabinet OfficeIA10 25/6/96 (I23) Senior Official, Parliamentary Commissioner for

AdministrationIA11 27/6/96 (I24) Senior Official, Home OfficeIA12 1/6/96 (I25) Senior Official, National Audit OfficeIA13 11/10/96 (I33) Senior Official, Northern Ireland Audit OfficeIA14 7/4/97 (I36) Senior Official, Department of Finance and Personnel

Northern IrelandIA15 8/4/97 (I37) Senior Official, Northern Ireland OmbudsmanIA16 11/9/96 (IEE7) Senior Official, HM Prison Service

Appendix 3 157

Appendix 4: Census of 173 Executive Agencies Createdbetween 1988 and 2001

The executive agencies are classified by the department in which the memberof the government to which they reported, generally the Secretary of State,was located at the time of their establishment. The date of establishment isgiven and * indicates a trading agency. The information for the classificationwas principally gathered from the Next Steps annual reviews from 1990onwards (Prime Minister and Minister for the Civil Service and the Ministerof State, Privy Council Office, 1990, 1991; Chancellor of the Duchy ofLancaster, 1992, 1993, 1994, 1995, 1996, 1997, 1998; Minister for the CabinetOffice, 1999, 2000; Office of Public Service Reform and HM Treasury, 2002).The departments with executive agencies are listed below, in order of thenumber of bodies created. There were, in addition, departments that were des-ignated as operating on ‘agency lines’. These included Customs and Excisefrom 1991, Inland Revenue from 1992, the Crown Prosecution Service andSerious Fraud Office from 1997. In 2001, these bodies employed 90,440 staffmaking 367,000 staff working in agencies or bodies working on ‘agency lines’,or 76 per cent of all civil servants (Cabinet Office, 2002, p. 18). However, thereforms in these departments were principally alterations to internal man-agement arrangements rather than a full application of the executive agencymodel and are not counted as a full use of the model.

Executive agencies Year

Ministry of Defence (fifty-two executive agencies)Army Base Repair Organisation 1993Armed Forces Personnel Administration 1997Army Base Storage and Distribution 1995Army Individual Training Organisation 1996Army Personnel Centre 1996Army Technical Support 1995Chemical and Biological Defence 1991Defence Accounts 1991Defence Analytical Services 1992Defence Animal Centre 1993

158


Defence Aviation Repair Agency* 2001Defence Bills Agency 1995Defence Clothing & Textile Agency 1994Defence Codification 1996Defence Communications Service Agency 1998Defence Dental Agency 1996Defence Estates 1997Defence Evaluation and Research Agency 1995Defence Geographic and Imagery Intelligence Agency 2000Defence Housing Executive 1999Defence Intelligence and Security Centre 1996Defence Operational Analysis 1992Defence Postal and Courier Service 1992Defence Research 1991Defence Science and Technology Laboratory* 2001Defence Storage and Distribution Agency 1999Defence Transport and Movements 1999Defence Transport and Movements Executive 1995Disposal Sales 1994Duke of York’s Military School 1992Hydrographic Office* 1990Joint Air Reconnaissance Intelligence 1996Logistic Information Systems Agency 1994Maintenance Group Defence 1994Medical Supplies Agency 1996Military Survey 1991Ministry of Defence Police 1996Meteorological Office* 1990Naval Aircraft Repair 1992Naval Base and Supply Agency 1996Naval Manning Agency 1996Naval Recruitment and Training Agency 1995Pay and Personnel Agency 1996Queen Victoria School 1992Royal Air Force Logistics Support Service 1996Royal Air Force Maintenance Group Defence Agency 1991Royal Air Force Signals Engineering Establishment 1994Royal Air Force Training Group Defence Agency 1994Service Children’s Education 1996Service Children Schools 1991Ships Support Agency 1996Warship Support Agency 2001

Appendix 4 159


Northern Ireland (NI) Departments (twenty-four executive agencies)Business Development Service 1996Compensation 1996Driver and Vehicle Licensing (NI) 1993Driver and Vehicle Testing Agency* 1992Environment and Heritage Service 1996Forensic Science Agency (NI) 1995Forest Service 1998Government Purchasing Agency 1996Health Estates (NI) 1995Industrial Research and Technology Unit 1995Land Registers of (NI) 1996NI Child Support Agency 1993NI Prison Service 1995NI Statistics and Research Agency 1996Ordnance Survey (NI) 1992Planning Service (NI) 1996Public Record Office (NI) 1995Rate Collection Agency 1991Rivers Agency (NI) 1996Roads Service 1996Social Security (NI) 1991Training and Employment (NI) 1990Valuation and Lands (NI) 1993Water Service 1996

Department of Trade and Industry (eleven executive agencies)Accounts Services 1991Companies House* 1988Laboratory of the Government Chemist 1989National Engineering Laboratory 1990National Physical Laboratory 1990National Weights & Measures Laboratory 1989Patent Office* 1990Radiocommunications Agency 1990Small Business Service 2000The Insolvency Service 1990Warren Spring Laboratory 1989

Scottish Office Departments and Scottish Executive (eleven executive agencies)Communities Scotland 2001HM Inspectorate of Education 2001Historic Scotland 1991

160 Appendix 4


Registers of Scotland* 1990Scottish Agricultural Science Agency 1992Scottish Court Service 1995Scottish Fisheries Protection 1991Scottish Office Pensions 1993Scottish Prison Service 1993Scottish Record Office 1993Student Awards for Scotland 1994

Ministry of Agriculture, Fisheries and Food (MAFF) and subsequent Department for Environment, Food and Rural Affairs (eleven executive agencies)ADAS 1992Central Science Laboratory 1994Central Veterinary Laboratory 1990Forest Enterprise 1996Forest Research 1997Intervention Board 1990Meat Hygiene Service 1995Pesticides Safety Directorate 1993Rural Payments Agency 2001Veterinary Laboratories Agency 1990Veterinary Medicines Directorate 1990

Cabinet Office (ten executive agencies)Buying Agency 1991Central Computer and Telecommunications Agency 1996Central Office of Information* 1990Chessington Computer Centre 1993Civil Service College 1989Her Majesty’s Stationary Office 1988Occupational Health Service 1990Property Advisers to the Civil Estate 1996Recruitment and Assessment Services 1991Security Facilities Executive 1993

Department of Transport (merged to form the Department of the Environment, Transport and the Regions) (ten executive agencies)Coastguard 1994Driver and Vehicle Licensing Agency 1990Driving Standards Agency* 1990DVOIT 1992

Appendix 4 161


Highways Agency 1994Marine Safety 1994Maritime and Coastguard Agency 1998Transport Research Laboratory 1992Vehicle Certification Agency 1990Vehicle Inspectorate* 1988

Department of Social Security and subsequent Department for Work and Pensions (eight executive agencies)Appeals Service Agency 2000Information Technology Services Agency 1990Jobcentre Plus 2002Social Security Benefits Agency 1991Social Security Child Support Agency 1993Social Security Contributions Agency 1991Social Security: Resettlement Agency 1989Social Security War Pensions Agency 1994

HM Treasury (seven executive agencies)Central Statistical Office 1991Debt Management Office 1998National Savings 1996Office for National Statistics 1996Paymaster 1993Royal Mint* 1990Valuation Office 1991

Department of the Environment (merged to form the Department of the Environment, Transport and the Regions) (six executive agencies)Building Research Establishment 1990Historic Royal Palaces 1989Ordnance Survey* 1990Planning Inspectorate 1992QEII Conference Centre* 1989The Rent Service 1999

Department of Health (five executive agencies)Medical Devices Agency 1994Medicines Control Agency* 1991National Health Service Estates* 1991National Health Service Pensions 1992National Health Service Purchasing and Supply Agency 2000

162 Appendix 4


Lord Chancellor’s Department (five executive agencies)Court Service 1995Her Majesty’s Land Registry* 1990Public Guardianship Office 2001Public Record Office 1992Public Trust Office 1994

Home Office (four executive agencies)Fire Service College* 1992Forensic Science Service* 1991Her Majesty’s Prison Service 1993UK Passport 1991

Treasury Solicitors Department (two executive agencies)Government Property Lawyers 1993Treasury Solicitors Department itself became an 1996executive agency of the Attorney General

Welsh Office and subsequently Welsh Assembly (two executive agencies)CADW: Welsh Historic Monuments 1991Welsh European Funding Office 2000

Department of Employment (merged to form Department for Education and Employment) (one executive agency)Employment Service 1990

Department of National Heritage (merged with Culture Media and Sport) (one executive agency)Royal Parks 1993

Department of Education (merged to form Department for Education and Employment) (one executive agency)Teachers’ Pensions Agency 1992

Foreign and Commonwealth Office (one executive agency)Wilton Park 1991

Overseas Development Administration (one executive agency)Natural Resources Institute 1990

Appendix 4 163

A few executive agencies were nominally departments in their own right,although they still reported to ministers in ‘full’ departments. These bodieswere the Central Office of Information, Central Statistical Office, HM LandRegistry, Intervention Board, National Savings, Ordnance Survey, Paymaster,Public Record Office, Registers of Scotland, Royal Mint, Scottish Record Office,Treasury Solicitor’s Department. A few executive agencies had multiple report-ing lines, ADAS reported to MAFF and the Welsh Office, Forest Enterprisereported MAFF and the Welsh Office, the Intervention Board reported to MAFFand the Scottish, Welsh and Northern Ireland Offices, the Meat HygieneService reported to MAFF and Scottish and Welsh Offices, and the PlanningInspectorate reported to the Department of the Environment and WelshOffice.

The proportion of staff who ended up in executive agencies, in generalterms, reflected variation in the significance of executive activity handled bydepartments and is summarised in Table A4.1. The reform was most signifi-cant in the Department of Social Security with the first wave of reformputting 97 per cent of staff in six executive agencies. The Lord Chancellor’sDepartment, Home Office and the Department of Employment, before thelatter Department’s merger with Education in 1995, all had 80 per cent oftheir staff in executive agencies (Cabinet Office, 1991, 1992). The majority ofdepartments had between 20 and 50 per cent of staff in executive agencies. Incontrast, the sole executive agency of the Foreign and Commonwealth Officecontained less than 1 per cent of the Department’s total staff.

164 Appendix 4

Appendix 4 165

Table A4.1 Percentage of departments’ staff in executive agencies andstaffing figures

Per cent of All Civil Civilstaff in Servants Servants in

agencies agencies

Social Security 97 89,000 86,500Lord Chancellor’s Department 90 12,000 10,500Department for Employment 80 50,000 40,000Home Office 80 50,000 40,000Department for the 74 19,000 14,000Environment Transport andthe Regions

Northern Ireland 50 16,000 8,000Trade and Industry 50 10,000 5,000Scottish Office 46 13,000 6,000Cabinet Office/Office of 40 2,500 1,000Public Service

Ministry of Agriculture 35 11,500 4,000Fisheries and Food

Overseas Development 33 1,500 500Administration

Ministry of Defence 26 122,000 32,000(including military staff)

Department for Education 20 2,500 500Wales 20 2,500 500Treasury 13 15,000 2,000Foreign and Commonwealth � 1 6,000 30Office

Source: Civil Service Statistics (Cabinet Office, 1991, 1992, 1995b).

Appendix 5: Central Government Accounts Audited

Table A5.1 summarises the results of a survey of NAO financial audit of allcentral government accounts including executive agencies’ accounts between1993 and 1999.

Table A5.1 Central government accounts audited 1993/94 to 1998/99

Average 1993/ 1994/ 1995/ 1996/ 1997/ 1998/1993–99 94 95 96 97 98 99

Agency 78 66 72 78 76 81 96accounts

Agency 2 2 2 1 1 2 2qualified

Agency 3 4 5 4 1 6 0reported

Per cent 6 9 10 6 3 10 2qualified/reported

Central govt 491 446 514 516 465 473 530accounts

Central 36 40 37 34 46 33 26qualified/reported

Per cent 7 9 7 7 9 7 5qualified/reported

Source: Compiled from NAO, 1996c, 1997, 1998d, 1999b, 2000b.

166

167

Bibliography

Aulich, C. (1999) ‘Bureaucratic Limits to Markets: The Case of LocalGovernment in Victoria, Australia’ Public Money and Management Vol. 19,No. 4, pp. 37–43.

Barberis, P. (1996) The Elite of the Elite: Permanent Secretaries in the British HigherCivil Service Aldershot, Darmouth.

Bardach, E. (1998) Getting Agencies to Work Together Washington D.C.,Brookings Institute.

Barker, A. (1998) ‘Political Responsibility for UK Prison Security – MinistersEscape Again’ Public Administration Vol. 76, No. 1, pp. 1–23.

Barry, B. (1990) Political Argument : A Re-issue with a New Introduction HemelHempstead, Harvester Wheatsheaf.

Bartlett, W., Roberts, A. and Le Grand, J. (1998) ‘The Development of Quasi-market in the 1990s’ in Bartlett, W., Roberts, A. and Le Grand, J. (eds) ARevolution in Social Policy University of Bristol, The Policy Press.

Barzelay, M. (2000) The New Public Management: Improving Research and PolicyDialogue Berkley, University of California Press.

Benefits Agency (1991) Framework Document London, HMSO.Benefits Agency (1992a) Interim Annual Report 1991–92 London, Benefits

Agency.Benefits Agency (1992b) Benefits Agency Business Plan Leeds, Benefits

Agency.Benefits Agency (1993) Interim Annual Report 1992–93 Leeds, Benefits Agency.Benefits Agency (1994) Annual Report 1993–94 HC 474, Session 1993–94

London, HMSO.Benefits Agency (1995a) Annual Report 1994–95 HC 621, Session 1994–95

London, HMSO.Benefits Agency (1995b) Framework Document London, Department of Social

Security.Benefits Agency (1996) Annual Report 1994–95 HC 489, Session 1994–95

London, HMSO.Benefits Agency (1997) Benefits Agency: Annual Report and Accounts 1996/97

HC 78, Session 1996–97 London, The Stationery Office.Benefits Agency (1998) Annual Report and Accounts 1997/98 HC 901 London,

The Stationery Office.Benefits Agency (1999a) Annual Report and Accounts 1998–99 HC 580 London,

The Stationery Office.Benefits Agency (1999b) Supplementary Memorandum Submitted by the DSS

Benefits Agency to the Public Accounts Committee Report on Vote 1 AppropriationAccount 1996–97: Class XII (Administered Social Security Benefits and OtherPayments) HC 419 I, Session 1998–99 London, The Stationery Office.

Benefits Agency (2000) Annual Report and Accounts 1999–2000 HC 640London, The Stationery Office.

Benefits Agency (2001) Annual Report and Accounts 2000–01 HC 153 Session2000–01 London, The Stationery Office.

Benefits Agency (2002) Annual Report and Accounts 2001–02 HC 929, Session2001–02 London, The Stationery Office.

Benefit Fraud Inspectorate (1999a) Report on London South Directorate (AD3)Benefits Agency London, Benefit Fraud Inspectorate.

Benefit Fraud Inspectorate (1999b) Report on East Devon Devon District CouncilLondon, Benefit Fraud Inspectorate.

Benefit Fraud Inspectorate (1999c) Report on the Social Security ContributionsAgency London, Benefit Fraud Inspectorate.

Benefit Fraud Inspectorate (1999d) Annual Report: Securing the System London,Benefit Fraud Inspectorate.

Bichard, M. (1999) Modernising the Policy Process London, Public Managementand Policy Association.

Biggs, S. and Dunleavy, P. (1995) ‘Changing Organisational Patterns in LocalGovernment: a Bureau-shaping Analysis’ in Lovenduski, J. and Stanyer, J.(eds) Contemporary Political Studies 1995: Volume II, Belfast, Political StudiesAssociation.

Benyon, J. and Edwards, A. (1999) ‘Community Governance of CrimeControl’ in Stoker, G. (ed.) The New Management of British Local GovernanceLondon, Macmillan, pp. 145–67.

Blais, A. and Dion, S. (1991) The Budget Maximising Bureaucrat: Appraisals andEvidence Pittsburgh, University of Pittsburgh Press.

Butler, D. and Kavanagh, D. (1992) The British General Election of 1992London, Macmillan.

Butler, D. and Kavanagh, D. (1997) The British General Election of 1997London, Macmillan.

Butler, R. (1991) ‘Foreward’ in Goldsworthy, D. Setting Up Next Steps London,HMSO, pp. i–ii.

Butler, R. (1994) ‘Reinventing Government: A Symposium’ PublicAdministration Vol. 72, No. 2, pp. 263–70.

Cabinet Office (1988) The Civil Service Yearbook London, HMSO.Cabinet Office (1989) The Civil Service Yearbook London, HMSO.Cabinet Office (1990) The Civil Service Yearbook London, HMSO.Cabinet Office (1991a) Civil Service Statistics London, HMSO.Cabinet Office (1991b) The Civil Service Yearbook London, HMSO.Cabinet Office (1992) Civil Service Statistics London, HMSO.Cabinet Office (1992b) The Civil Service Yearbook London, HMSO.Cabinet Office (1993) The Civil Service Yearbook London, HMSO.Cabinet Office (1994) The Civil Service Yearbook London, HMSO.Cabinet Office (1995a) Guidance on Agency Reviews London, Cabinet Office.Cabinet Office (1995b) Civil Service Statistics 1995 London, HMSO.Cabinet Office (1995c) Public Bodies 1995, London, HMSO.Cabinet Office (1995d) The Civil Service Yearbook London, HMSO.

168 Bibliography

Cabinet Office (1996) The Civil Service Yearbook London, The Stationery Office.Cabinet Office (1997a) The Civil Service Yearbook London, The Stationery

Office.Cabinet Office (1997b) Next Steps Briefing Note March 1997, London, Cabinet

Office.Cabinet Office (1998a) Next Steps Briefing Note September 1998, London,

Cabinet Office.Cabinet Office (1998b) The Civil Service Yearbook London, The Stationery

Office.Cabinet Office (1999) Civil Service Statistics 1999 London, The Stationery Office.Cabinet Office (2000) Civil Service Statistics 1999 London, Government

Statistical Service.Cabinet Office (2001a) Agency Policy Review Team Presentation to Chief

Executives’ Conference October 2001 http://www.civilservice.gov.uk/agencies/ace/con_intro.htm (last accessed 17 June 2001).

Cabinet Office (2001b) Civil Service Statistics 2000 London, HMSO.Cabinet Office (2002a) Civil Service Statistics 2001 London The Stationery

Office.Cabinet Office (2002b) Former Agencies Information www.civilservice.gov.uk/

agencies (last accessed 1st December).Caiden, G.E. (1991) Administrative Reform Comes of Age London, Walter de

Gruyter.Campbell, C. and Wilson, G. (1995) The End of Whitehall: Death of a Paradigm?

Oxford, Blackwell.Carter, N. and Greer, P. (1993) ‘Evaluating Agencies: Next Steps and

Performance Indicators’ Public Administration Vol. 71, No. 4, pp. 407–16.Carter, N., Klein, R. and Day, P. (1992) How Organisations Measure Success: The

Use of Performance Indicators in Government London, Routledge.Centre for Management and Policy Studies (2001) Centre for Management and

Policy Studies Portfolio Including the Civil Service College Directory 2001–2002London, Cabinet Office.

Chadwick, A. and May, C. (2003) ‘Interaction between States and Citizens inthe Age of the Internet: E-government in the US, Britain and the EuropeanUnion’ Governance Vol. 16, No. 2.

Chancellor of the Duchy of Lancaster (1992) The Next Steps Agencies Review1992 Cm 2111 London, HMSO.

Chancellor of the Duchy of Lancaster (1993) Next Steps: Agencies inGovernment Review 1993 Cm 2430 London, HMSO.

Chancellor of the Duchy of Lancaster (1994) Next Steps Review 1994 Cm 2750London, HMSO.

Chancellor of the Duchy of Lancaster (1995) Next Steps Review 1995 Cm 3164London, HMSO.

Chancellor of the Duchy of Lancaster (1997) Next Steps Agencies in GovernmentReview 1996 Cm 3579 London, The Stationery Office.

Chancellor of the Duchy of Lancaster (1998) Next Steps Report Cm 3889London, The Stationery Office.

Bibliography 169

Chancellor of the Exchequer (1987) The Government’s Expenditure Plans1987–88 to 1989–90 Vol. II Cm 56-II London, HMSO.

Chancellor of the Exchequer (1991) Competing for Quality – Buying Better PublicServices Cm 1730 London, HMSO.

Chancellor of the Exchequer (1994) Public Expenditure Statistical Supplement tothe Financial Statement and Budget Report 1994–95 Cm 2519 London, HMSO.

Chancellor of the Exchequer (1998) Public Expenditure Statistical Analyses1998–99 Cm 3901 London, The Stationery Office.



Charkham, J.P. (1994) Keeping Good Company: A Study of Corporate Governancein Five Countries Oxford, Clarendon Press.

Chief Secretary to the Treasury (1998) Public Services for the Future:Modernisation, Reform, Accountability: Comprehensive Spending Review: PublicService Agreements 1999–2002 Cm 4181 London, The Stationery Office.

Chief Secretary to the Treasury (2000) 2000 Spending Review: Public ServiceAgreements Cm 4808 London, The Stationery Office.

Christensen, T. and Laegreid, P. (2001) New Public Management: TheTransformation of Ideas and Practice Aldershot, Ashgate.

Citizen’s Advice Bureau (1998) Short Changed London, Citizen’s Advice Bureau.Civil and Public Services Association (1997) Better Government London, Civil

and Public Services Association.Clifford, C. (2000) ‘Transforming Civil Service Employee Relations: Executive

Agencies, Managerial Freedom and the New Public Management’ Paper presented to the British Universities Industrial Relations Association 50thConference 7–9th July 2000, University of Warwick.

Coase, R. (1960) ‘The Problem of Social Costs’ Journal of Law and EconomicsVol. 3, No. 1, pp. 1–44.

Conservative Party (1987) The Next Moves Forward: The Conservative Manifesto1987 London, Conservative Central Office.

Conservative Party (1992) The Best Future for Britain: The ConservativeManifesto 1992 London, Conservative Central Office.

Cope, S. (1995) Contracting-out in Local Government: Cutting by Privatising’Public Policy and Administration Vol. 10, No. 3, pp. 29–44.

Council of Civil Service Unions (1996) Evidence to the House of Lords SelectCommittee on the Public Services London, Council of Civil Service Unions.

Crouch, C. and Streeck, W. (1997) Political Economy of Modern CapitalismLondon, Sage.

Deakin, N. and Parry, R. (2000) The Treasury and Social Policy Basingstoke,Palgrave.

Department of Social Security (1993) Social Security Departmental Report: TheGovernment’s Expenditure Plans 1993–94 to 1995–96 Cm 2213 London, HMSO.

Department of Social Security (1995) Benefits Agency Framework Document1995–2000 London, DSS.

170 Bibliography

Department of Social Security (1999a) Social Security Departmental Report: TheGovernment’s Expenditure Plans 1999/2000 Cm 4214 London, The StationeryOffice.

Department of Social Security (1999b) A New Contract for Welfare:SafeguardingSocial Security Cm 4276 London, The Stationery Office.

Department of Work and Pensions (2001) Departmental Report: TheGovernment’s Expenditure Plans 2002/3 to 2003/4, Cm 5424 London,The Stationery Office.

Dollery, B. and Hamburger, P. (1996) ‘Modelling Bureaucracy: The Case of theAustralian Federal Budget Sector 1982–1992’ Public Administration Vol. 74,No. 3, pp. 477–507.

Domberger, S. (1998) The Contracting Organisation: A Strategic Guide toOutsourcing Oxford, Oxford University Press.

Dopson, S. (1993) ‘Are Agencies an Act of Faith? The Experience of HMSO’Public Money and Management Vol. 13, No. 3, pp. 17–23.

Dowding, K.M. (1995) The Civil Service London, Routledge.Dowding, K.M. and Kang, W-T. (1998) ‘Ministerial Resignations 1945–97’

Public Administration Vol. 76, No. 3, pp. 411–29.Dudley, G. (1993) ‘Next Steps Agencies, Political Salience and the Arms-length

Principle: Barbara Castle at the Ministry of Transport 1965–68’ PublicAdministration Vol. 72, No. 2, pp. 219–40.

Dunleavy, P. (1985) ‘Bureaucrats, Budgets and the Growth of the State:Reconstructing an Instrumental Model’ British Journal of Political ScienceVol. 15, No. 3, pp. 299–328.

Dunleavy, P. (1986) ‘Explaining the Privatization Boom: Public Choice VersusRadical Approaches’ Public Administration Vol. 64, pp. 13–34.

Dunleavy, P. (1989a) ‘The Architecture of the British State, Part One:Framework for Analysis’, Public Administration Vol. 67, pp. 249–75.

Dunleavy, P. (1989b) ‘The Architecture of the British State, Part Two:Empirical Findings’ Public Administration Vol. 68, pp. 3–28.

Dunleavy, P. (1991) Democracy, Bureaucracy and Public Choice HemelHempstead, Harvester Wheatsheaf.

Dunleavy, P. (1994) ‘The Globalisation of Public Services Production: Can Government be “Best in World?” ’ Public Policy and AdministrationVol. 9, No. 2, pp. 36–64.

Dunleavy, P. and Margetts, H. (2000) ‘The Advent of Digital Government: PublicBureaucracy and the State in the Internet Age’ Paper presented to the AnnualConference of the American Political Science Association Washington D.C., 4th September.

Dunsire, A. and Hood, C.C. (1989) Cutback Management in Public BureaucraciesCambridge, Cambridge University Press.

Efficiency Unit (1988) Improving Management in Government: The Next StepsLondon, HMSO.

Efficiency Unit (1991) Making the Most of Next Steps London, HMSO.Efficiency Unit (1993) Career Management and Succession Planning Study

London, HMSO.

Bibliography 171

Efficiency Unit (1996) Competing for Quality Policy Review: An Efficiency UnitScrutiny London, HMSO.

Evans, P. (1994) ‘Members of Parliament and Agencies: ParliamentaryQuestions’ in P. Giddings (ed.) Parliamentary Accountability: A Study ofParliament and Executive Agencies London, Macmillan.

Ferlie, E., Ashburner, L., Fitzgerald, L. and Pettigrew, A. (1996) The New PublicManagement in Action Oxford, Oxford University Press.

Fesler, J.W. and Kettl, D.F. (1996) The Politics of the Administrative ProcessChatham, New Jersey, Chatham House.

Flinders, M.V. and Smith, M.J. (1999) Quangos, Accountability and Reform: thePolitics of Quasi-government London, Macmillan Press.

Fogden, M.E.G. (1993) ‘Management Change in the Employment Service’Public Money and Management Vol. 13, No. 3, pp. 9–16.

Friel, B. (1999) Perfromance Based Organisations Back in Style http://www.govexec.com/dailyfed/0299/020399b1.htm (last accessed 15 August)

Fulton Comittee (1968) The Civil Service Report of the Committee 1968–88 Vol. I Cm 3638 London, HMSO.

Gains, F. (1999) ‘Implementing Privatisation Policies in “Next Steps”Agencies’ Public Administration Vol. 77, No. 4, pp. 713–30.

Goldsworthy, D. (1991) Setting Up Next Steps London, HMSO.Greer, P. (1994) Transforming Central Government: the Next Steps Initiative

Buckingham, Open University Press.Greve, C., Flinders, M.V. and Van Thiel, S. (1999). Quangos: What’s in a

Name? Defining Quangos from a Comparative Perspective. GovernanceVol. 2, No. 2, pp. 29–46.

Hansard (1988) Official Report, Session 1987–88 London, HMSO.Harlow, C. (1999) ‘Accountability, New Public Management, and the

Problems of the Child Support Agency’ Journal of Law and Society Vol. 26,No. 2, pp. 150–74.

Heclo, H. and Wildavsky, A. (1981) The Private Government of Public MoneyLondon, Macmillan.

Hennessy, P. (1989) Whitehall London, Secker and Warburg.Heseltine, M. (1987) Where There’s a Will London: Hutchinson.Hirose, K. (2000) ‘Regulating Government in Japan’ Paper presented to

workshop Comparing Regulation of Public Sector Bodies, London School ofEconomics, October.

Hoggett, P. (1991) ‘A New Management in the Public Sector’ Policy and PoliticsVol. 19. No. 4, pp. 243–56.

Hogwood, B.W. (1995) ‘Whitehall Families: Core Departments and AgencyForms in Britain’ International Review of Administrative Sciences Vol. 61, pp. 511–30.

Hogwood, B.W. (1997) ‘The Machinery of Government 1979–97’ PoliticalStudies Vol. 45, No. 4, pp. 704–15.

Hogwood, B., Judge, D. and McVicar, M. (1997) ‘The “Pondlife” of ExecutiveAgencies: Parliament and “Informatory Accountability” ’ Public Policy andAdministration Vol. 12, No. 2, pp. 95–115.

Hood, C.C. (1991) ‘A Public Management for All Seasons?’ PublicAdministration Vol. 69, No. 1, pp. 3–19.

172 Bibliography

Hood, C.C. (1994) Explaining Economic Policy Reversals Buckingham, OpenUniversity Press.

Hood, C.C. and James, O. (1997) ‘The Central Executive’ in Dunleavy, P. et al.(eds) Developments in British Politics 5 London, Macmillan, pp. 177–204.

Hood, C.C. and Jones, G.W. (1990) ‘Progress in the Government’s Next StepsInitiative’ Memorandum Submitted to the Treasury and Civil Service Committeeincluded in Appendices to the Minutes of Evidence of the Eighth Report Progressin the Next Steps Initiative HC 481 Session 1989–90 London, HMSO.

Hood, C.C., James, O. and Scott, C.D. (2000) ‘Regulation in Government; Hasit Increased, Is it Increasing, Should it be Diminished?’ Public AdministrationVol. 78, No. 2, pp. 283–304.

Hood, C.C., James, O., Jones, G.W., Scott, C. and Travers, T. (1998) ‘RegulationInside Government: Where New Public Management Meets the AuditExplosion’ Public Money and Management Vol. 18, No. 2, pp. 61–8.

Hood, C.C., Scott, C.W., James, O., Jones, G.W., Travers, T. (1999) RegulationInside Government: Waste-watchers, Quality Police and Sleaze-busters Oxford,Oxford University Press.

Hoopes, S. (1997) Oil Privatization, Public Choice and International ForcesLondon, Macmillan.

Hori, M. (2002) ‘Japanese Public Administration and its Adaptation to Newpublic Management from an Optimist’s View’ Paper presented to AnnualMeeting of the American Political Science Association Conference, Boston M.A.,Sept 1st.

Horton, S. and Jones, J. (1996) ‘Who are the New Public Managers? An InitialAnalysis of Next Steps Chief Executives and their Managerial Role’ PublicPolicy and Administration Vol. 11, No. 4, pp. 30–46.

Howlett, M. and Ramesh, M. (1995) Studying Public Policy: Policy Cycles andPolicy Subsystems Oxford, Oxford University Press.

Hyndman, N.S. and Anderson, R. (1998) ‘Performance Information,Accountability and Executive Agencies’ Public Money and ManagementVol. 18, No. 3, pp. 23–30.

Information Technology Services Agency (1996) Annual Report and Accounts,HC 443, Session 1995–96 London, The Stationery Office.



James, O. (1995a) ‘The Agency “Revolution” in Whitehall: A Bureau-shapingAnalysis’ in Lovenduski, J. and Stanyer, J. (eds) Contemporary Political Studies1995: Proceedings of the Political Studies Association’s 1995 Annual ConferenceQueen’s University, Belfast, UK Political Studies Association.

James, O. (1995b) ‘Explaining the Next Steps in the Department of SocialSecurity: The Bureau-shaping Model of Central State Reorganisation’Political Studies, Vol. 43, No. 4, pp. 614–29.

James, O. (2000) ‘Regulation Inside Government: Public Interest Justificationsand Regulatory Failures’ Public Administration Vol. 78, No. 2, pp. 327–44.

James, O. (2001) ‘Business Models and the Transfer of Business-like CentralGovernment Agencies’ Governance Vol. 14, No. 2., pp. 233–52.

Bibliography 173

James, O. (2002) ‘Beyond the “New Public Management”: RegulatedPartnerships in Britain’ in Ahn, B., Halligan, J. and Wilks, S. (eds) ReformingPublic and Corporate Governance Cheltenham, Edward Elgar, pp. 61–82.

James, O. and Manning, N. (1996) ‘Public Management Reform: A GlobalPerspective’ Politics Vol. 16, No. 3, pp. 143–50.

John, P. (1998) Analysing Public Policy London, Pinter.Jordan, G. and O’Toole, B. (1995) ‘The Next Steps: Origins and Destinations’

in O’Toole, B. and Jordan, G. (eds) (1995) Next Steps: Improving Managementin Government? Aldershot, Dartmouth, pp. 3–16.

Kaneko, Y. (1999) Government Reform in Japan Management and Co-ordination Agency, Government of Japan.

Kemp, P. (1990) ‘Next Steps for the Civil Service’ Governance, Vol. 3, No. 2, pp. 186–96.

Kickert, W.J.M., Klijn, E-H. and Koppenjan, J.F.M. (1997) Managing ComplexNetworks: Strategies for the Public Sector London, Sage.

Kutlu, O. (2001) Administrative Reform in the UK: the Next Steps Agencies Reformand its Applicability to Turkey Exeter, University of Exeter PhD Thesis.

Lane, J.E. (1993) The Public Sector: Concepts, Models and Approaches London,Sage.

Lawson, N. (1993) The View from Number 11 London, Bantam Press.Lewis, D. (1997) Hidden Agendas: Politics, Law and Disorder London, Hamish

Hamilton.Lomax, R. (2002) New Services, New Department Presentation to the Chief

Executives’ Spring Conference, Edinburgh, 15th March.Lord President of the Council (1981) Efficiency in the Civil Service, Cm 8293

London, HMSO.Margetts, H. (1999) Information Technology in Government: Britain and America

London, Routledge.Marsh, D., Smith, M.J. and Richards, D. (2000) ‘Bureaucrats, Politicians in

Whitehall: Analysing the Bureau-shaping Model’ British Journal of PoliticalScience Vol. 30, No. 3, pp. 461–82.

Marsh, D., Richards, D. and Smith, M.J. (2001) Changing Patterns of Goverance:Reinventing Whitehall Basingstoke, Palgrave.

Massey, A. (1995) After Next Steps London, Cabinet Office.Matsuda, T. (1999) Some Points for Independent Administrative Corporation

System Management and Co-ordination Agency, Government of Japan.Mellon, E. (1993) ‘Executive Agencies: Leading Change from the Outside-in’

Public Money and Management Vol. 13, No. 2, pp. 25–31.Milgrom, P. and Roberts, J. (1992) Economics, Organisation and Management

London, Prentice Hall.Minister for the Cabinet Office (1999) Next Steps Report 1998 Cm 4273

London, The Stationery Office.Minister for the Cabinet Office (2000) Executive Agencies 1999 Report Cm 4658

London, The Stationery Office.Moodie, M., Mizan, N. Heron, R. and Mackay, B. (1988) The Business of Service

London, DHSS.

174 Bibliography

Mueller, A.E. (1989) Public Choice II Cambridge, Cambridge University Press.Muramatsu, M. and Krauss, E.S. (1996) ‘Japan: the Paradox of Success’ in

Olsen, J.P. and Peters, B.G. (eds) (1996) Lessons from Experience: ExperientialLearning in Administrative Reforms in Eight Democracies Oslo, ScandinavianUniversity Press, pp. 214–43.

National Audit Office (1988) Department of Health and Social Security: Quality ofService to the Public at Local Offices HC 451, Session 1987–88 London, HMSO.

National Audit Office (1989) The Next Steps Initiative HC 410, Session 1988–89London, HMSO.

National Audit Office (1992) Department of Social Security: Resettlement AgencyHC 22, Session 1991–92 London, HMSO.

National Audit Office (1995a) Protecting and Presenting Scotland’s HeritageProperties HC 430, Session 1994–95 London, HMSO.

National Audit Office (1995b) The Banking Services Provided by the Office of HMPaymaster General HC 513, Session 1994–95 London, HMSO.

National Audit Office (1995c) The Meteorological Office Executive Agency HC693, Session 1994–95 London, HMSO.

National Audit Office (1996a) Management of Telephones in the Benefits AgencyHC 126, Session 1995–96 London, HMSO.

National Audit Office (1996b) Information Technology Services Agency:Outsourcing the Service Delivery Options HC 255, Session 1995–96 London,HMSO.

National Audit Office (1996c) General Report of the Comptroller and AuditorGeneral for 1994–95 on Appropriation and Other Accounts 15-XIII London, TheStationery Office.

National Audit Office (1997a) Financial Auditing and Reporting 1995–96 GeneralReport of the Comptroller and Auditor General 11-XII London, The StationeryOffice.

National Audit Office (1997b) Measures to Combat Housing Benefit Fraud HC164, Session 1997–98 London, The Stationery Office.

National Audit Office (1998a) Underpayments to Public Service Pensioners onInvalidity Benefit HC 681, Session 1997–98, London, The Stationery Office.

National Audit Office (1998b) Benefits Agency: Performance Measurement HC952, Session 1997–98 London, The Stationery Office.

National Audit Office (1998c) Department of Social Security: Progress on Measuresto Combat Housing Benefit Fraud HC 31, Session 1998–99, London, TheStationery Office.

National Audit Office (1998d) Financial Auditing and Reporting 1996–97General Report of the Comptroller and Auditor General 251-XIX London, TheStationery Office.

National Audit Office (1999a) Department of Social Security: progress on Measuresto Combat Housing Benefit Fraud HC 319, Session 1998–99 London, TheStationery Office.

National Audit Office (1999b) Financial Auditing and Reporting 1997–98General Report of the Comptroller and Auditor General 251-XIX London, TheStationery Office.

Bibliography 175

National Audit Office (1999c) Government on the Web HC 87, Session1999–2000 London, The Stationery Office.

National Audit Office (2000a) State Earnings-related Pension Scheme: The Failureto Inform the Public of Reduced Pension Rights for Widows and Widowers HC320, Session 1999–2000 London, The Stationery Office.

National Audit Office (2000b) Financial Auditing and Reporting 1998–99General Report of the Comptroller and Auditor General 11-XIX London, TheStationery Office.

National Audit Office (2000c) The Cancellation of the Benefits Payment CardProject HC 857 Session 1999–2000 London, The Stationery Office.

National Audit Office (2001) Measuring the Performance of GovernmentDepartments, HC 301, Session 2000–01 London, The Stationery Office.

National Audit Office (2002a) Better Public Services Through E-Government HC704-I, Session 2001–2002 London, The Stationery Office.

National Audit Office (2002b) Government on the Web II HC 764, Session2001–2002 London, The Stationery Office.

National Performance Review (1996) Reinvention’s Next Steps: Governing in aBalanced Budget World http://www.npr.gov/library/papers//bkgrd/balbud.htm(last accessed 10 August 2001).

National Performance Review (1998) Creating the Government’s FirstPerformance Based Organisation to Modernise Student Aid Deliveryhttp://www.npr.gov/library/news/pbofact.html (Last accessed 10/8/01).

National Performance Review (1999) Summary of Next Steps: the UK CentralGovernment Management Reform http://www.npr.gov/21cent/nextstep.html (lastaccessed 10 August 2001).

Natzler, D. and Silk, P. (1994) ‘Departmental Select Committees and the NextSteps Programme’ in Giddings, P. (ed.) Parliamentary Accountability: A Studyof Parliament and Executive Agencies London, Macmillan, pp. 71–94.

Next Steps Project Team (1995a) The Strategic Management of Agencies: Modelsfor Management London, Office of Public Service.

Next Steps Project Team (1995b) The Strategic Management of Agencies: FullReport and Case Studies London, Office of Public Service.

Next Steps Team (1996) Next Steps Briefing Note: Feb. 1996, Cabinet Office,London.

Next Steps Team (1997) Next Steps Briefing Note: Oct. 1997, Cabinet Office,London.

Next Steps Team (1998) Next Steps Briefing Note, Sept. 1998 London, CabinetOffice.

Niskanen, W.A. (1971) Bureaucracy and Representative Government Chicago,Aldine-Atherton.

Niskanen, W.A. (1991) ‘A Reflection on Bureaucrats and RepresentativeGovernment’ in Blais, A. and Dion, S. (eds) The Budget-MaximisingBureaucrat: Appraisals and Evidence Pittsburgh, University of PittsburghPress, pp. 13–31.

OECD (1995) Public Management Developments Update 1995 Paris, OECD.OECD (1998) Survey of Public Management Developments Paris, OECD.

176 Bibliography

OECD (2002) Distributed Public Governance: Agencies, Authorities and otherAutonomous Bodies Paris, OECD.

Office of the Minister for the Civil Service (1989) Developments in the NextSteps Programme: The Government Reply to the Fifth Report from the Treasuryand Civil Service Committee, HC 348, Session 1988–89 London, HMSO.

Office of Public Services Reform and HM Treasury (2002) Better Govern-ment Services: Executive Agencies in the 21st Century London, The CabinetOffice.

Orzechowski, W. (1977) ‘Economic Models of Bureaucracy: Survey, Extensionsand Evidence’ in Borcherding, T.E. (ed.) Budgets and Bureaucrats: The Sourcesof Government Growth Durham, North Carolina, Duke University Press, pp. 229–59.

O’Toole, B. and Jordan, G. (1995) Next Steps: Improving Management inGovernment? Aldershot, Dartmouth.

Parliamentary Commissioner for Administration (1990) ParliamentaryCommissioner for Administration: Third Report Annual Report for 1989 HC 353,Session 1989–90 London, The Stationery Office.

Parliamentary Commissioner for Administration (1998) ParliamentaryCommissioner for Administration: Summary of Annual Report for 1997–98London, Parliamentary Commissioner for Administration.

Parliamentary Commissioner for Administration (1999) ParliamentaryCommissioner for Administration: Seventh Report Annual Report for 1998–99 HC572, Session 1998–99 London, The Stationery Office.

Parry, R., Hood, C.C. and James, O. (1997) ‘Reinventing the Treasury:Economic Rationalism or an Econocrat’s Fallacy of Control’ PublicAdministration Vol. 75, No. 3, pp. 395–415.

Parsons, W. (1995) Public Policy: An Introduction to the Theory and Practice ofPolicy Analysis Aldershot, Edward Elgar.

Pendlebury, M., Jones, R. and Karbharti, S. (1992) ‘Accounting for ExecutiveAgencies in the UK Government’ Financial Accountability and ManagementVol. 8, No. 1, pp. 35–48.

Performance and Innovation Unit (2000) Wiring It Up: Whitehall’sManagement of Cross-cutting Policies and Services London, The StationeryOffice.

Pliatzky, L. (1980) Report on Non-Departmental Public Bodies Cm 7797 London,HMSO.

Pliatzky, L. (1992) ‘Quangos and Agencies’ Public Administration Vol. 70, No. 4, pp. 555–63.

Polidano, C. (1999) ‘The Bureaucrat Who Fell Under a Bus: MinisterialResponsibility, Executive Agencies and the Derek Lewis Affair’ GovernanceVol. 12, No. 2, pp. 201–29.

Pollitt, C. (1993) Managerialism and the Public Services Oxford, Blackwell.Pollitt, C. (1995) ‘Justification by Works or Faith? Evaluating the New Public

Management’ Evaluation Vol. 1, No. 2, pp. 133–54.Pollitt, C. and Bouckaert, G. (2000) Public Management Reform Oxford, Oxford

University Press.

Bibliography 177

Pollitt, C., Birchall, J. and Putnam, K. (1998) Decentralising Public ServiceManagement London, Macmillan.

Pollitt, C., Bathgate, K., Smullen, A. and Talbot, C. (2000) ‘Agencies: A TestCase for Convergence?’ Paper presented to the Fourth InternationalSymposium on Public Management Erasmus University Rotterdam, April.

Price Waterhouse (1993a) Agencies in Government London, Price Watehouse.Price Waterhouse (1993b) Executive Agencies: Facts and Trends London, Price

Waterhouse.Prime Minister (1990) Progress in the Next Steps Initiative, Cm 1263 London,

HMSO.Prime Minister (1991) The Citizen’s Charter, Cm 1599 London, HMSO.Prime Minister and Minister for the Civil Service (1970) The Reorganisation of

Central Government Cm 4506 London, HMSO.Prime Minister and Chancellor of the Exchequer (1982) Efficiency and

Effectiveness in the Civil Service: Government Observations on the Third Reportfrom the Treasury and Civil Service Select Committee, Session 1981–82, Cm 8616London, HMSO.

Prime Minister and Minister for the Cabinet Office (1999) ModernisingGovernment Cm 4310 London, The Stationery Office.

Prime Minister and Minister for the Civil Service and the Minister of State,Privy Council Office (1990a) Progress in the Next Steps Initiative Cm 1263London HMSO.

Prime Minister and Minister for the Civil Service and the Minister of State,Privy Council Office (1990b) Improving Management in Government: The NextSteps Agencies Review 1990 Cm 1261 London, HMSO.

Prime Minister and Minister for the Civil Service and the Minister of State,Privy Council Office (1991) Improving Management in Government – The NextSteps Agencies Review 1991 Cm 1760 London, HMSO.

Prime Minister, Chancellor of the Exchequer and Chancellor of the Duchy ofLancaster (1994) The Civil Service: Continuity and Change Cm 2627 London,HMSO.

Prison Service (HM) (1996) Prison Service Annual Report and Accounts HC 247,Session 1996–97 London, The Stationery Office.




Prosser, T. (1986) Nationalised Industries and Public Control: Legal, Constitutionaland Political Issues Oxford, Blackwell.

Public Accounts Committee (1989) 36th Report The Next Steps Initiative HC420, Session 1988–89 London, HMSO.

Public Accounts Committee (1999) Minutes of Evidence Wednesday 28th April1999 HC 419 I, Session 1998–99 London, The Stationery Office.

Public Accounts Committee (2000) Appropriation Accounts 1997–98, Class XII, Vote 1 (Central Government Administered Social Security Benefits

178 Bibliography

and Other Payments) HC 103, Session 1999–2000 London, The StationeryOffice.

Public Information Office, Department of the Library, House of Commons(1997) House of Commons: Sessional Information Digest 1995–96 London, TheStationery Office.

Public Service Committee (1996) 2nd Report Ministerial Accountability andResponsibility HC 313-I, Session 1995–96 London, HMSO.

Rhodes, R.A.W. (1997) Understanding Governance: Policy Networks, Governance,Reflexivity and Accountability Buckingham, Open University Press.

Richards, D. and Smith, M. (2002) Governance and Public Policy in the UKOxford, Oxford University Press.

Samuels, M. (1997) Benchmarking Next Steps Executive Agencies London,Cabinet Office.

Samuels, M. (1998) Towards Best Practice: An Evaluation of the First Two Years ofthe Public Sector Benchmarking Project 1996–8 London, Cabinet Office.

Secretary of State for Social Security and Chief Secretary to the Treasury (1999)Social Security Departmental Report: the Government’s Expenditure Plans1999/2000 Cm 4214, London, The Stationery Office.

Sen, A. (1973) ‘Behaviour and the Concept of Preference’ Economica Vol. 40,No. 3, pp. 241–59.

Smith, P. (1995) ‘On the Unintended Consequences of Publishing PerformanceData in the Public Sector’ International Journal of Public AdministrationVol. 18, pp. 277–310.

Social and Community Planning Research (1991) British Social Attitudes SurveyCompilation 1983–1991 University of Essex, UK Data Archive.

Social Security Committee (1993) 1st Report The Operation of the Child SupportAct HC 69, Session 1993–94 London, HMSO.

Social Security Committee (1994) 5th Report The Operation of the Child SupportAct: Proposals for Change HC 470, Session 1993–94 London, HMSO.

Social Security Committee (1995) 2nd Report The Performance and Operation ofthe Child Support Agency HC 50, Session 1995–96 London, HMSO.

Social Security Committee (1998) Social Security Committee: Minutes ofEvidence: Thursday 21st May 1998 HC 587iii, Session 1997–98 London, TheStationery Office.

Social Security Committee (1999) Social Security Committee: Seventh Report The ONE Service Pilots HC 412, Session 1998–99 London, The StationeryOffice.

Talbot, C. (1996) ‘Ministers and Agencies: Responsibilities and Performance’in The Public Service Committee Second Report Ministerial Accountability andResponsibility Volume II Minutes of Evidence HC 313, Session 1995–96 London,HMSO, pp. 39–55.

Talbot, C. (1998) ‘Small Step or Giant Leap? Evaluating Ten Years of the “NextSteps” Agencies in the UK (1988–98)’ Research Paper University of Glamorgan.

Talbot, C. (2002) So Who Needs Agencies: A Handbook Exploring the Issues of Using‘Agencies’ as a Reform Device in Central Government University of Glamorgan.

Talbot, C., Pollitt, C., Bathgate, K., Caulfield, J., Reilly, A. and Smullen, A.(2000) ‘The Idea of Agency: Researching the Agencification of the (Public

Bibliography 179

Service) World’ Paper presented to the American Political Studies AssociationAnnual Conference Washington D.C., August.

Thain, C. and Wright, M. (1995) The Treasury and Whitehall : the Planning andControl of Public Expenditure, 1976–1993 Oxford, Clarendon Press.

Thatcher, M. (1993) The Downing Street Years London, Harper-Collins.Theakston, K. (1995) The Civil Service Since 1945 Oxford, Blackwell.Thomas, M. (1996) Management Practices No. 16: Special Operating Agencies:

Audit and Evaluation Canadian Centre for Management Development.Treasury (HM) (1988a) Civil Service Statistics 1988 London, HMSO.Treasury (HM) (1988b) Civil Service Staff in Post 1988 Manpower Statistics

Division, HM Treasury.Treasury (HM) (1989) The Financing and Accountability of Next Steps Agencies

Cm 914 London, HMSO.Treasury (HM) (1990) Civil Service Statistics 1989–90 London, HMSO.Treasury (HM) (1991) Competing for Quality Cm 1730 London, HMSO.Treasury (HM) (1992a) Executive Agencies; a Guide to Setting Targets and

Monitoring Performance London, HMSO.Treasury (HM) (1992b) Private Finance Guidance London, HMSOTreasury (HM) (1996) Public Expenditure Statistical Analyses 1996–97 Cm 3201

London, HMSO.Treasury (HM) (2000a) The Government’s Expenditure Plans 2000–2001 to

2001–2002: Chancellor of the Exchequer’s Department Part of Cm 4601–4621London, The Stationery Office.

Treasury (HM) (2002) Public Expenditure Statistical Analyses 2002–3, Cm 5401London, The Stationery Office.

Treasury and Civil Service Committee (1982) Treasury and Civil ServiceCommittee, Third Report, Efficiency and Effectiveness in the Civil Service, HC 236Vol. II- Minutes of Evidence, Session 1981–82 London, HMSO.

Treasury and Civil Service Committee (1990) Eighth Report: Progress in the NextSteps Initiative Session HC 481 Session 1989–90 London, HMSO.

Treasury and Civil Service Committee (1993) 5th Report The Role of the CivilService Volume II Minutes of Evidence HC 27-I, Session 1993–94 London, HMSO.

Trosa, S. (1994) Next Steps: Moving On London, Office of Public Service.Wallis, J. and Dollery, B. (2001) ‘Understanding Cultural Changes in an

Economic Control Agency: The New Zealand Treasury’ Journal of PublicPolicy Vol. 21, No. 2, pp. 191–212.

Weimer, D.L. and Vining, A.R. (1996) ‘Economics’ in Kettl, D.F. and Milward, H.B. (eds) The State of Public Management Baltimore, John HopkinsUniversity Press, pp. 92–117.

Wilson, D. and Game, C. (1994) Local Government in the United KingdomBasingstoke, Macmillan.

Wilson, R. (1999) Report to the Prime Minister: Civil Service Reform London,Cabinet Office.

Yoon, B-S. (2001) The New Public Management: Its Applicability to AdministrativeReform in Korea Exeter, University of Exeter PhD Thesis.

Zifcak, S. (1994) Administrative Reform in Whitehall and Canberra Buckingham,Open University Press.

180 Bibliography

Index

181

accountability, see Benefit FraudInspectorate, Citizen’sCharter/Service First, FinancialManagement Initiative,Parliamentary Commissionerfor Administration, PublicAccounts Committee, see underBenefits Agency, executiveagencies, ministers, NationalAudit Office, Parliament

administrative reform, see bureau-shaping, Next Steps reform,public interest perspective

Barzelay, Michael, 1benchmarking, 136–7, see also

Business Excellence ModelBenefit Fraud Inspectorate, 117–18,

144Benefits Agency

accountability, 79–82Citizen’s Charter standards for the

Benefits Agency, 83, 97, 142creation as part of Next Steps

reform, 61–4, 129economy, 90–2, 130effectiveness and efficiency,

96–100, 131effect of executive agency

structures on performance,101–5, 114–23, 131

employment flexibilities, 82, 102,121

information technology, 122–3reform to create Working Age

agency and Pension Service aspart of Department of Workand Pensions, 144–6

relationship with Department forSocial Security, 74–6, 115–17,162

relationship with Parliament, 67, 85relationship with users, 86,

115–16, 119–33systemic performance, 114–23,

132, 138–9Bichard, Michael, 63, 115, 119Bouckaert, Geert, 2budget-maximising

civil servants’ behaviour andbudget outcomes, 48, 88–92,102, 109–10, 130–2

Niskanen budget-outputmaximising model, 10, 32–5

budget systems, see under BenefitsAgency, executive agencies,Treasury (HM)

bureau-shapingbureau-shaping perspective on

executive agencies andhypotheses, 10–11, 23–37,154–5; hypotheses aboutreform and consistency withreform, 41–68; hypothesesabout performance andconsistency with outcomes,88–107, 109–24; relativemerits of bureau-shaping andpublic interest perspectives,127–133; relevance of theperspective for reforms inother countries, 148–9

Dunleavy bureau-shaping model,10, 23–5, 153

Business Excellence Model, 93–4Butler, Sir Robin, 12, 51

cabinet committees, see EconomicAffairs (Welfare to Work)Committee, Public ExpenditureCommittee

Cabinet Office, 43–5, 54–5, 83, 135,137, 141–2, 147

central government, see departments,executive agencies, economy,effectiveness, efficiency

Charter Mark, 85–6chief executives, see under executive

agenciesChild Support Agency, 85, 95–6,

145Citizen’s Charter/Service First, 83,

97, 141–2civil servants

numbers of and trends in numberof over time, 7, 43, 56–8, 110,158

traditional conceptions of, 1see also departments, Senior Civil

Service, see also under budgetmaximising, bureau-shaping

Civil Service ManagementCommittee, 140

Clifford, Christopher, 101Competing for Quality Initiative,

53–4, 82–3consumers, 34–5, see also Citizen’s

Chartercontracts, see Competing for Quality

Initiative, see under executiveagencies

Contributions Agency, 118, 162

Department for Work and Pensions,144–6

Department of Social Securityand Next Steps reform, 61–4relationship with Benefits Agency,

74–6, 115–17departments, see Department of

Social Security, Department forWork and Pensions, Treasury(HM)

devolution, 134, see also NorthernIreland, Scotland, Wales

Dowding, Keith, 130Dunleavy, Patrick, 113, see also under

bureau-shaping

Economic Affairs (Welfare to Work)Committee, 144

economydefined, 8–9of individual executive agencies,

88–92systemic economy, 109–10

effectivenessdefined, 9of individual executive agencies,

92–100systemic effectiveness, 111–23

efficiencyallocative efficiency, 9efficiency of individual executive

agencies, 92–100productive efficiency, 9systemic efficiency, 111–23

Efficiency Unit, 17–18, 48–50, 54–5,128

e-government, 7, 112–14, 121–3Employment Service, 73, 90, 114,

119–21, 143–4employment systems, see under

Benefits Agency, executiveagencies

executive agenciesaccountability, 3–5, 76–80, 83–6,

102–5, 118–19, 122–3budget flexibilities, 3–4, 80–2census of bodies created, 158–64chief executives, 3–4, 19–20, 67,

72–6, 78–9, 81–2, 84, 129–30employment flexibilities, 3–4, 82executive agency model, 2–7;

influence of new institutionaleconomics includingprincipal–agent models, 51–2;influence of Anglo-Americanprivate sector businessmodels, 50–1

Framework Documents, 70–1,74–5, 83, 145

future prospects for use in UK,134–46

growth in number, 56–8individual performance, 88–100

182 Index

182

executive agencies – continuedinternational experience with

similar forms of organisation,1–2, 146–50

mainstream agencies, 71, 73, 129management flexibilities, see

budget flexibilities,employment flexibilities,semi-detached organisation

mergers between agencies, 139,144–6

Ministerial Advisory Boards, 77non-trading agencies, 3–4, 34,

56–8, 60–1, 72–3, 87, 89–90,113, 129–30

pattern of reform, 58–60, 129,158–66

performance contracts, seeexecutive agency model,Framework Documents,performance targets

performance targets, 3–4, 38,76–81, 83, 87, 129

policy/administration split, 3–4,10, 19–20, 61, 63–4, 66–8, 71,74–5, 112, 115–17, 129, 133,135, 145

privatisations, 135–6regulation of, 140–6, see also

accountabilityreintegration with departments,

135relationships with Treasury, 72, 81semi-detached organisation, 3–4,

70–6, see also budgetflexibilities, employmentflexibilities, policy/administration split

systemic performance, 109–23,132

trading agencies, 3–5, 20–1, 35,56–8, 60–1, 70–3, 81, 85,89–90, 106, 113, 129–30, 133,135, 138, 158–64

Financial Management Initiative,45–6, 50

Fraser Report, 76Fulton Report, 6

Gains, Francesca, 135governance, see departments, see

under accountability, executiveagencies, ministers, Parliament

government expenditure, see budgetsystems, Public ExpenditureCommittee, see under economy,executive agencies

Government Performance andResults Act (US), 148

Greer, Patricia, 54

Heseltine, Michael, 43–5Hood, Christopher, 1, 140, 156–7Housing Benefit, 114–15, 117–18,

124, 138, 143human resources, see under Benefits

Agency, civil servants, executiveagencies

Ibbs, Sir Robin, 48–9implementation, see under Benefits

Agency, executive agencies,Next Steps reform

incentives, see under accountability,executive agencies, Next Stepsreform

Income Support, 97–9, 103, 117–18Independent Administrative

Institutions (in Japan), 148–50information technology, 7, 121–2,

see also e-government,Information TechnologyServices Agency

Information Technology ServicesAgency, 64, 145, 162

international issues, seeOrganisation for EconomicCooperation and Development,see under executive agencies

Investors in People, 141

Jobcentre Plus, 144–5Jobseeker’s Allowance, 90, 103, 114,

117, 119

Index 183

joined-up government, 111–12, 121

Kemp, Sir Peter, 12, 54, 56

Lewis, Derek, 79local government, 6, 41, 53, 63,

111, 114–15, 117–18, 120–1,138–9, 144, 153

managerialism, 45–6, 50, 134–5Margetts, Helen, 113market testing, see Competing for

Quality InitiativeMarsh, David, 46–7, 128Mathison, Peter, 104, 120methodology, 11–14, 156–7ministers

ministerial resignations, 130role in Next Steps reform, 43–6,

127–8satisfaction with executive

agencies’ performance, 94–5Modernising Government Report, 7,

111–12, 121

National Audit Officefinancial audit of executive

agencies, 92–3, 166value for money studies, 84, 113,

143National Performance Review (US),

149networks, 139–46new institutional economics, 50–2New Labour, 111–12, 116–17New Public Management, 1–2, 4–6,

50–2, 111–12, 133–4, 146–7Next Steps

implementation of reform, 55–64,see also under BenefitsAgency

initiation of reform, 41–55Marsh, Richards and Smith’s

study, 46–7, 128Prime Minister’s announcement

of reform, 17–18Project Team, 54–6, 69

report of 1988; and executiveagency model, 2–3; andpublic interest perspective,9–10, 18–23

Niskanen, William, 10, 26, 32–4Non-departmental public bodies, 6,

52, 111, 148Northern Ireland, 59, 134

ONE/Single Work Focused Gateway,120–1

organisation, see under departments,executive agencies, New PublicManagement

Organisation for Economic Co-operation and Development,146–7

outcome and output measures, seeperformance, performancetargets

Parliamentcommittees and executive

agencies, 44, 81–6Parliamentary questions about

executive agencies, 66–7see also Public Accounts

Committee, Social SecurityCommittee, ParliamentaryCommissioner forAdministration

Parliamentary Commissioner forAdministration, 97, 99

pay systems, see under executiveagencies, New PublicManagement

Pensions Service, 144–5performance, see under Benefits

Agency, economy, effectiveness,efficiency, executive agencies

Performance and Innovation Unit,111

Performance Based Organisations (inUS), 147–9

performance targets, see PublicService Agreements, see underexecutive agencies

184 Index

personnel systems, see underexecutive agencies, New PublicManagement

politicians, see under ministers,Parliament, preferences

Pollitt, Christopher, 1–2, 7, 11preferences

of ministers, 28of senior officials for the

organisation of theirdepartments, 24, 27–36

Prison Service (HM), 67, 71, 79,95–6, 130

privatisationof executive agencies, 135–6of public utilities, 52–3see also under New Public

Managementproductivity, see under efficiency,

executive agenciesPublic Accounts Committee, 81–2,

96Public Expenditure Committee,

142–4public interest perspective

defined, 9–10, 17–23hypotheses about performance

and consistency withoutcomes, 88–107, 109–24

hypotheses about reform andconsistency with reform,41–68

relative merits of bureau-shapingand public interestperspectives, 127–33

public sector externalities, 35–6,138–9

Public Service Agreements, 142–3

quasi-markets, 5

rational choice, 10, see alsoNiskanen, Dunleavy, see alsounder budget-maximising,bureau-shaping

reform, see under bureau-shaping,Next Steps reform, publicinterest perspective

regulationinside government, 140systemic regulation of executive

agencies in centralgovernment, 140–6

reorganisation, see bureau-shaping,Next Steps reform, publicinterest model

Rhodes, Rod, 10Richards, David, 46–7, 128

Scotland, 59, 134Senior Civil Service, 67–8, 140Smith, Martin, 46–7, 128Social Security Committee, 119,

121–2Special Operating Agencies (in

Canada), 147strategic planning, see Public Service

Agreements, see also underexecutive agencies

Talbot, Colin, 7targets, see performance targetsTheakston, Kevin, 54trading funds/agencies, see under

executive agenciesTreasury (HM)

co-ordination of government,including Public ServiceAgreements, 111, 136, 141–3

role in reforms including NextSteps, 44–5, 54–5, 144

unions, 54, 63–4

Wales, 59, 134Whitehall, 7, see also under Benefits

Agency, civil servants,departments, executiveagencies, ministers

Zifcak, Spencer, 50

Index 185