Managerical reforms and developmental state capacity

Размещено на http://www.allbest.ru/

Managerical reforms and developmental state capacity

A. Nickson

This paper analyzes the impact that the `policy transfer' of managerial reforms associated with the New Public Management (NPM) has had upon enhancing the capacity of the broadly defined `developmental state'. The paper begins by outlining the key principles of the NPM-type managerial reforms proposed to governments as a response to the challenges of wider public sector reform. Section 2 analyzes the extent to which these managerial reforms have actually been implemented around the world, with special reference to the experience of countries where they have either been modified or abandoned. Section 3 addresses the issue of `policy transfer' by examining the extent to which aid- dependent countries have been more prone than other countries to adopt such managerial reforms. Section 4 outlines the major features of the developmental state, with a particular focus on the nature of the bureaucracy. Section 5 provides a critical assessment of the extent to which the introduction of managerial reforms can enhance the overall developmental capacity of the state. The final section offers some broader conclusions.

Keywords: New Public Management, policy transfer, developmental state.

Э. Никсон

РЕФОРМЫ СИСТЕМЫ УПРАВЛЕНИЯ И ПОТЕНЦИАЛ ГОСУДАРСТВЕННОГО РАЗВИТИЯ

В представленной работе анализируется влияние, которое оказали управленческие реформы, связанные с Новым Государственным Управлением (НГУ), на усиление потенциала так называемого `государственного развития'. В первой части статьи описываются основные принципы, лежащие в основе НГУ-управленческих реформ, предложенных правительствам стран в ответ на реформу бюджетной сферы. Во второй части анализируется степень реализации управленческих реформ в разных частях мира, с особым акцентом на опыт тех стран, где они были изменены или отвергнуты. Третья часть посвящена `политике воздействия'; автор исследует вопрос, в какой мере зависимые от помощи страны более склонны, чем другие страны, к проведению таких управленческих реформ. В четвертой части рассматриваются основные черты политики государственного развития, уделяется особое внимание природе бюрократии. Пятая часть посвящена критической оценке того, до какой степени проведение управленческих реформ может способствовать развитию государства. Последний раздел предлагает обобщающие выводы.

Ключевые слова: новое государственное управление, политика воздействия, государственное развитие.

Introduction

The last thirty years have witnessed a radical reshaping of the priority focus for public management reforms in LMICs, one which has largely been in response to a paradigm shift in the wider understanding of public sector reform. Prior to the 1980s, the stated focus of public management reform programmes was the establishment and strengthening of Weberian-style public management systems characteristic of many OECD nations. These tended to be characterised by meritocratic systems of recruitment, promotion and performance monitoring. Yet from the 1980s onwards, as the `neo- liberal' development paradigm gained ascendancy around the world, so the understanding of public management reform gradually turned its attention to the growing importance of the role of the private sector in the overall development process, including the delivery of basic public services. This encouraged the introduction of a package of managerial reforms of the public sector that became known collectively as the `New Public Management' (NPM). Although there is considerable controversy as to the extent to which neo-liberalism was the intellectual progenitor of the NPM, there is no doubt that the emergence and global spread of the latter coincided with the emergence of the former.

This paper seeks to analyze the impact that the `policy transfer' of these managerial reforms has had upon enhancing the capacity of the so-called `developmental state'. Before proceeding, it is necessary to address the conceptual confusion concerning the meaning of the `developmental state'. Until relatively recently this referred to the particular attributes of the state in a small number of East Asian countries (e.g. Japan, South Korea, Singapore, Hong Kong, Taiwan) that had experienced very rapid economic growth in the second half of the 20^th century. This sub-set was later widened to include Botswana [1; 48] and Mauritius [10], and more, recently Vietnam [53] and China. The core feature of the developmental state have been described as “... the intervention of the state in the economy, but in the form of policies that are credible and oriented towards growth, not in the form of the ownership and direct control by the state of large parts of the economy” [73. P. 617].

However, from the mid-1990s the term `developmental state' has been used in a much broader sense, as exemplified by the following quote: “A developmental state is now broadly understood as one that evinces a clear commitment to a national development agenda, that has solid capacity and reach, and that seeks to provide growth as well as poverty reduction and the provision of public services” [30 P. 4]. In the contemporary era of democratisation, it has even been suggested that the conception of the developmental state must be reconfigured to include a framework of competitive party politics [67]. Furthermore, a linkage has emerged between the developmental state and the wider good governance agenda. A central and distinguishing feature of the development state - `government effectiveness' - is now incorporated as one of the six criteria used for ranking countries under the World Bank (WB) Governance Indicators Dataset.

Two very different groups of countries can be identified within this wider characterisation of the developmental state. First, are those states whose governments have reacted against the `neo-liberal tide' in order to pursue a nationalistic development style in which the state plays a more direct role in the productive process and in the delivery of basic services (e.g. Bolivia, Ecuador, S Africa, Venezuela). Second, are those states whose governments have reacted against the predatory behaviour of their predecessors and now aspire to a new commitment to `development' in general and to democratisation, good governance and poverty reduction in particular. These are primarily aid-dependent HIPC countries, with governments pursuing a Poverty Reduction Strategy (PRSs) under strong donor monitoring. In both groups of countries, the challenge of strengthening the `developmental capacity of the state' is an important topical issue of debate. However, in neither case does the current concern for promoting the `developmental capacity of the state' necessarily imply that these countries are seeking to pursue a growth strategy akin to that of the classic `developmental state' model referred to above.

The paper begins by outlining the key principles of the NPM-type managerial reforms proposed to governments as a response to the challenges of wider public sector reform. Section 2 analyzes the extent to which these managerial reforms have actually been implemented around the world, with special reference to the experience of countries where they have either been modified or abandoned. Section 3 addresses the issue of `policy transfer' by examining the extent to which aid-dependent countries have been more prone than other countries to adopt such managerial reforms. Section 4 outlines the major features of the developmental state, with a particular focus on the nature of the bureaucracy. Section 5 provides a critical assessment of the extent to which the introduction of managerial reforms can enhance the overall developmental capacity of the state. The final section offers some broader conclusions.

The key principles of NPM-type managerial reforms

The NPM is a generic term used to describe a bundle of management techniques introduced to the public sector, many of which are borrowed from the private sector [29; 36; 82]. It emerged in New Zealand and the United Kingdom from the 1980s. Through the leverage of IMF structural adjustment programmes and the powerful intermediary of the World Bank and other IFIs, it began to be forcefully impressed on governments of LMICs around the world. The package of reform measures associated with the NPM is often referred to collectively as `managerialism'. Broadly speaking, this proposes a shift in the focus of higher levels of public management towards `steering' rather than `rowing', strategic thinking and targeting resources in order to achieve defined outputs and outcomes.

The IMF has promoted three main tenets of public sector reform in pursuit of its overarching concern for fiscal balance. The first two of these assert that i) in order to overcome chronic fiscal deficits, the overall role of the public sector in the economy should be reduced, and ii) within the framework of this reduced role, the public sector should retreat from involvement in productive sectors of the economy but strengthen its involvement in social sectors. The policies derived from these tenets - retrenchment/downsizing and privatisation - are commonly referred to as the `first generation' reforms of the Washington Consensus. The third tenet argues that in those sectors where the public sector retains a strong involvement, the central state should switch from a direct provider role to that of strategic `conductor', at one and the same time regulating and enabling the delivery of services by other providers, whether from sub-national levels of government (e.g. municipal or regional governments), semi- autonomous state bodies (e.g. executive agencies or deconcentrated administrative units) or the non-state sector (e.g. the private sector or not-for-profit sector). The NPM provides the intellectual basis for these so- called `second generation' reforms.

The NPM paradigm is underpinned by two key ideas that have shaped the public management reform process as outlined above, namely `public choice' and `principal-agent'[36]. Public choice theory offers a critique of the central assumptions of the Weberian model, arguing that bureaucrats are exclusively motivated by individual utility maximisation (i.e. rational self-interest) rather than the public good. It proposes a series of inter-organisational initiatives as a way of overcoming `rent-seeking' behaviour inside the state apparatus. These managerial reforms that advocated `flexibility' in staffing arrangements constituted a frontal attack on the homogeneity of the bureaucratic system that is a central tenet of Weberianism. They seek to open up markets in order to promoting efficiency in service delivery and giving `choice and voice' to users of public services. The introduction of market and quasi-market mechanisms is designed to stimulate competition between service providers in the belief that this will promote cost savings and customer responsiveness. The most common instruments are contracting-out, franchising, internal markets, vouchers, and user charges or fees. There is also an emphasis on the quality of service provision in an effort to make public services more `demand' rather than `supply' driven. This new responsiveness to customers is expressed through initiatives such as `citizen charters'. There is also an associated change in employment practices from traditional career tenure towards a preference for limited term contracts for senior staff, locally determined pay rather than uniform fixed salaries, the use of wholly monetised incentives, and the introduction of performance-related pay The second intellectual idea embedded in the NPM is principal-agent theory. This is central to the argument for the reassignment of roles to the different actors involved in the process of service delivery. It involves a split between the purchaser (the central ministry) and the providers (the rest, irrespective of whether in the public or private sector) that is driven primarily by the effort to capture gains in economic efficiency, equity and effectiveness.

Where service delivery is retained within the public sector, this entails the breaking up of large public sector bureaucracies via the disaggregation of functions. A clear distinction is made between the strategic policy core within a ministry (the principal) and the operational arms of ministries that are hived off to form separate executive agencies (the agents). Executive agencies are a tool for `unbundling the bureaucracy' and for `letting the managers manage' in accordance with PA theory. They have four major characteristics that distinguish them from a traditional unified bureaucracy - delegated management and associated financial autonomy, a specialisation on specific operational tasks, a focus on outputs, and performance-related contracts. Each agency is no longer related to the parent ministry through the traditional departmental hierarchy, but by an arms' length `performance agreement' of a contractual nature that has specified performance targets. In cases in which service delivery is transferred to the private sector, this involves a contractual agreement between the central ministry (the principal) and the private service deliverer (the agent).

The two above `drivers' - public choice and principal-agent theories - have given rise to five broad trends in public management. These are often overlapping and mutually reinforcing and collectively provide the `face' of NPM:the replacement of traditional `tall hierarchies' with flatter, flexible and more responsive organizational structures that are reformed around the specific operational tasks at hand rather than the traditional functions of finance and personnel.

1) the separation between funding, purchasing and the provision of services. This involves making a clearer organisational and financial separation between, on the one hand, defining the need for and paying for public services, and, on the other hand, the actual delivery of those services.

2) the delegation of management authority within public sector organizations. This involves giving top management the freedom to manage while at the same time defining more precisely their managerial responsibilities through performance targets linked to fixed-term contracts.

3) the delegation of budgets and financial responsibility to decentralised units through the creation of budget centres. This gives managers greater flexibility to consider the most cost-effective means to ensure the provision of required services to a pre- determined standard.

4) the shift from a focus on inputs and processes to outputs and outcomes in control and accountability mechanisms. This involves a greater emphasis on output controls in resource allocation, performance agreements, and performance-related pay. This requires managers to work to performance targets, indicators and output objectives.

The global spread of managerial reforms

After decades of inertia around the world, the NPM paradigm kicked off processes of public sector reform from the mid-1980s. Heavily promoted by the World Bank and other IFIs, it was a central feature of the so-called `second wave' of reforms associated with the Washington Consensus. The NPM `toolbox' thus became a major example of `policy transfer' in the field of public management. Although there is a wealth of literature reporting on country-specific examples, there is surprisingly little data that provides a comprehensive overview of this process of policy transfer of managerial reform. A recent study of the `public management' components of four global studies of state efficiency - the `government effectiveness' component of the World Bank Governance Indicators, the European Central Bank's public sector efficiency study, the `public institutions index' within the Global Competitiveness Report and the `government efficiency ranking' of the World Competitiveness Yearbook - concluded that they all suffered from serious data quality and questionable conceptual validity. Furthermore, none of them contained disaggregation that would enable comparative measurement of the specific impact of `market-driven' reforms.

Despite the absence of any such comprehensive overview, it is clear that the global spread of these managerial reforms of the public sector has been both uneven and patchy. Among OECD members, there has been a noticeable bias in take-up among the so-called `Anglo-American' countries, from the pioneering countries (New Zealand and United Kingdom) extending to Canada, Australia and, to a lesser extent, United States. A dramatic impact indicator is that by 2007 75 % of the UK civil service were employed in executive agencies.

By comparison, the public sectors of EU member nations have been more resistant to the introduction of such market-driven reforms. In France, there is continuing adherence to a body of administrative law regulating procedures and recruitment in the public sector. This dirigiste system is controlled by the grands corps, a politico-administrative elite. Consequently, attempts to introduce managerialist reforms have made little inroads, such that “... much of the machinery of a centralized civil service remains fundamentally unaltered” [62. P. 250]. In Sweden's highly decentralised system of public administration, local government has long held major responsibility for basic service delivery. Despite the introduction of NPM-type reforms, the `outsourcing' of local government activities and services has been limited to around 10 %, on average [74].

In Germany, the Neues Steuerungsmodell (New Steering Model) that emerged in the 1990s sought to bring greater flexibility to the Weberian bureaucracy by introducing certain managerial principles such as indicator-based performance management and cost-accounting. The resulting `amalgam' has been referred to as the classic example of the `Neo-Weberian State' (NWS) [9]. This term has been coined to express a growing trend among public bureaucracies in Continental Europe to adopt selected features of the NPM-type managerial reforms associated with `Anglo- American' countries. Under this `pick and mix' process, it is argued that these countries are seeking to enhance the efficiency and responsiveness of the state in resource allocation without buying into those aspects of the NPM paradigm that risk jettisoning the benefits of a unified rule-based bureaucracy.

The continuing commitment in the NWS to the Weberian ideal is expressed through a reaffirmation of the state as the main facilitator of solutions to emerging problems associated with technological change, globalisation, environmental threats and demographic changes; a reaffirmation of the role of representative democracy (at central, regional and municipal level) as the legitimising elements within the state apparatus; a reaffirmation of the role of administrative law in preserving the basic principles of relations between the citizen and the state, such as equality before the law and legal scrutiny of state activities; and the maintenance of the idea of `public service' imbued with a distinctive status, ethical culture and terms and conditions. The `neo' elements selected from managerialism are expressed in the NWS through a shift from an inward-looking concern for compliance with bureaucratic rules towards an outward looking orientation to meeting citizen's needs through the creation of a professional culture of quality; supplementation of the classic features of representative democracy with a range of consultative devices and, in some cases, the direct representation of citizens views; a shift in management towards greater focus on outputs rather than observance of procedures, with an accompanying shift from ex ante controls towards ex post evaluation; and a shift in the concept of bureaucratic professionalism away from a focus on expertise in administrative law towards citizen-oriented managers [62. P. 99-100].

South Korea, itself a leading exponent of the classic `developmental state' model, has also experimented since the 1997 financial crisis with the introduction of managerial reforms inspired by the NPM paradigm. Under the Kim Dae-Jung administration (1998-2003), the Civil Service Commission established an `open position system' to attract talented candidates from outside government to 20 % of the posts in the top three grades of the civil service. It also introduced performance-related pay for senior managers. Most government agencies and public enterprises adopted Service Charters and introduced a `Public Customer Satisfaction Index'. By December 2005, there were 23 executive agencies in operation reporting to 16 different ministries. However, there has been widespread informal resistance to the implementation of these reforms, considerably restricting managerial autonomy in executive agencies [19].

The picture is equally heterogeneous in LMICs. With the end of communist rule, reforms in Central and Eastern European countries concentrated initially on the establishment of democratic systems of government rather than public management reforms per se. However, as internal and external reform pressures increased in order to implement the acquis communitaire needed for closer ties to the European Union, the establishment of classic Weberian bureaucratic norms became the guiding principle - to the neglect of NPM-type institutional arrangements. “Classic continental career systems appear to be the main source of inspiration for Central and Eastern European states, as well as for those CIS countries where progress on civil service reform has been made. The German model has generally emerged as a dominant influence. [...] Even though the introduction of managerialist principles is often discussed, there generally appears to be a clear tendency to return to the `continental roots' of pre-1945” [78].

In Latin America, attempts to introduce NPM-type managerial reforms have proved to be unsustainable. Spearheaded by Kate Jenkins, previously Permanent Under-Secretary at the Cabinet Office during the Thatcher government and William Plowden, a leading advocate of NPM, a series of initiatives were carried out in Brazil, Colombia and Mexico. In Mexico, in the context of a corrupt bureaucracy riddled with clientelism and ghost workers, a 1996 attempt to introduce even greater flexibility in terms of recruitment and promotion was successfully rebuffed by the powerful finance ministry, which proposed moves towards a Weberian-type bureaucracy as a move to combat the proposals [61]. Perhaps unsurprisingly, a recent review of public administration reform in the region, which hardly made mention of the NPM, concluded that “After all these years, the progress of meritocracy is still the great challenge for public administration reform in Latin America” [25. P. 154].

In Sub-Saharan Africa, the take-up of managerial reforms has been greater among Anglophone than Francophone countries, perhaps reflecting the lower cultural barriers to policy transfer from its `Anglo' source. In Tanzania the 1997 Executive Agencies Act provided the framework for the creation of executive agencies on the UK model. Driven from the heart of government by the Civil Service Department, a branch of the presidency, it envisaged the creation of 46 EAs by 2004. By 2001 only nine agencies had been created. Ironically, the slow implementation of the programme was attributed to its `capture' by the bureaucracy, which viewed it primarily as a convenient mechanism for to responding to donor pressure, led by the IMF, to downsize the civil service while raising real incomes of senior staff. From 1992-2000 public sector employment fell by 27 % from 355,000 to 260,000. `Agencification' made a considerable contribution to that decline, with agencies reported a decline in their staff numbers from 20 % to 50 % [11].

The most persuasive evidence for `capture' is the fact that the agency head signs the `framework agreement' with the Permanent Secretary and not with the relevant minister, as in the UK. This marginalisation of the political elite from the `agencification' process largely explains the weak political will to drive the programme. Contrary to the legal requirement that the principal should determine the performance indicators, most agencies were allowed to develop their own targets with permanent secretaries abdicating their roles as `strategic managers'. As a result, performance monitoring, the key accountability mechanism in the agency arrangement, was missing. Consequently, agencies developed a degree of operational autonomy never envisaged by the PA model underpinning the agency concept. A study of the Tanzania experience of executive agencies concluded that “. they are treated in a far more autonomous way than either the textbooks or, indeed, the legislation prescribe” [11. P. 217]. The Tanzania experience highlights a wider issue of significance - that in spite of the rhetoric public sector reform in many LMICs is not necessarily solely about the `developmental' desire to improve the efficiency of service delivery, as presumed by the NPM paradigm. In quasi-predatory states, the institutional changes associated with managerial reforms that are imported under donor pressure, may be utilised in pursuit of other objectives. As the above study notes, “In Tanzania, the creation of executive agencies has been about more than efficiency and effectiveness reforms in the pursuit of better service delivery. It has also been about improving the salaries of senior civil servants and returning a profit to government” [11. P. 219].

Nevertheless, in much of SSA pay and employment measures, rather than the creation of executive agencies, have been at the centre of efforts to reform public management systems. These measures include: an increase in overall real wages and monetisation of benefits; decompression of pay scales to improve competitiveness of pay at higher levels; a simplification of grading systems, based on job evaluations; introduction of performance related pay; and improvement in general personnel management. Except for the introduction of performance related pay, it is difficult to categorise these reforms as primarily `managerial' in nature. Rather they seek to strengthen the typical attributes of a Weberian bureaucracy. In fact, even a fervent advocate of the benefits of NPM-type reforms for improving the performance of the state in SSA, recognised that, “The basic thrust of the [public service] reform process was, and continues to be, to build a professional, meritocratic and qualified public workforce to ensure effective and efficient delivery of public services and combat bureaucratic corruption” [37].

This review suggests that the global spread of NPM-type managerial reforms since the 1980s has been more limited than is commonly supposed. Instead, public management reforms in most LMICs, with the active support of aid donors, remain focussed primarily on the tortuous challenge of establishing the core features of a Weberian bureaucracy. As Manning as stated, “If there is relatively little NPM to be found in developing countries compared to earlier predictions, then there is even less evaluation... In fact there has been no systematic evaluation of outcomes “ [52. P. 305].

One of the few comparative overviews is a study of the impact of managerial reforms on the changing role of government in four core sectors (urban water, agricultural marketing, basic health and business development) and four countries (Ghana, India, Sri Lanka and Zimbabwe) [6] .The authors found that the new approaches to public management embodied in these reforms - regulating markets, enabling other providers, managing service delivery through decentralized structures, contracting out and charging users - “were inherently more complex and taxing of governmental capacity than previous arrangements” [6. P. 221]. The study concluded that this “raised questions about the appropriateness of the radical and comprehensive application of the new public management reform models to low-income countries” [6. P. 235].

Have aid-dependent countries been more prone to adopt managerial reforms than other low- and middle-income countries

No estimates exist of the total volume of foreign aid devoted to public sector management reform [7]. According to the OECD-DAC aid database, in 2006 $6,4bn was spent worldwide on strengthening `government and civil society' (excluding `conflict, peace and security'), equivalent to 6.6 % of global foreign aid in that year. Although no breakdown is available, a significant share of this amount was spent, as in previous years, on public management reforms and capacity-building. There has long been general agreement among major multilateral donors that the impact of this kind of aid has fallen well short of expectations [59; 89; 90; 26]. As Berg put it, “The substantial donor efforts to reform public sector management in low-income countries during the past 15-20 years can justifiably be called failures. And a significant share of responsibility for these dismal results has to be attributed to donor deficiencies as reformers” [7. P. 306].

To what extent could this poor impact of aid for public management reform be attributed to a misplaced emphasis on promoting NPM-type managerial reforms? During the heyday of neo-liberalism, IMF conditionality did act as a powerful mechanism through which market-oriented economic policy reforms in general were transferred to aid-dependent countries. These `first generation' reforms associated with structural adjustment also had a major impact on the bureaucracy in SSA and Latin America. Containment of the wage bill came to dominate civil service reform programmes because of its link to the overriding concern for overcoming the fiscal deficit. This `down-sizing' often had two negative impacts on the limited `Weberianness' of the bureaucracy by i) draining it of its “best and brightest” senior staff were often the first to take early retirement packages, and ii) creating a parallel corps of contractual (and sometimes re-recruited) personnel, leading to a dual salary structure.

However, in response to the widespread criticism associated with the staff retrenchment and privatisation that followed from structural adjustment lending, donors have committed themselves to new mechanisms to promote greater `ownership' of the state reform process. At the same time they have learnt from bitter experience that the state itself needs to become more effective in order to make aid more effective. These changing perceptions were embodied in the 2002 Paris Declaration on Aid Effectiveness, Ownership, Harmonisation, Alignment, Results and Mutual Accountability and the subsequent emphasis on General Budget Support (GBS) and Poverty Reduction Strategy Papers (PRSPs). However, given that IMF/World Bank approval of the PRSP is a prior condition for GBS, there is every reason to believe that its content is often modified in response to donor pressure, thereby undermining the very notion of country ownership. For this reason, it is plausible that PRSPs could operate in practice as a new form of donor conditionality. If this is the case, then a priori it is quite possible that aid-dependent countries could be particularly susceptible to the global `policy transfer' of NPM-type managerial reforms.

Yet there is little evidence that PRSPs and covert conditionality through GBS tend to incorporate explicit NPM type reforms. One reason for this could be that the highly aid- dependent HIPCs, for which PRSPs are mandatory, especially those in a post-conflict situation, tend to be characterised by very weak public administration systems. In this context, donors are likely to be reluctant to impose managerial reforms that require a higher level of administrative capacity than that found in such countries. On the other hand, the recent growth in aid and proliferation of donor agencies, both official and NGOs, by `poaching' competent and experienced staff, can have a very negative impact in weakening the meritocratic content of the bureaucracy [44].

The problem is that no systematic data is available to test the hypothesis that managerial reforms have indeed been more prominent in low-income aid dependent countries than elsewhere. One important source does provide partial data on the spread of NPM-type reforms. This is the Private Participation in Infrastructure (PPI) database, part of the Public-Private Infrastructure Advisory Facility (PPIAF) of the World Bank, which monitors growing private sector participation (PSP) in the delivery of basic services. The data base is restricted to projects in LMICs, covers four sectors (energy, transport, water and telecom) and classifies Private Sector Participation (PSP) into four categories (management and lease contracts, concessions, greenfield projects, and divestitures). Latin America accounted for 1,202 (32 %) of the total of 3,793 projects logged for the period 1900-2006, followed by East Asia (28 %), Europe and Central Asia (19 %), Sub- Saharan Africa (9 %), South Asia (9 %) and Middle East and North Africa (3 %). A cursory examination of the database suggests that there is no correlation between aid- dependence and take-up of PSP contractual relationship for service delivery. The top ten countries by project destination accounted for 63 % of the total investment committed, of around $1,100m and only two of these - Indonesia and Philippines - are aid dependent countries.⁸ (www.ppi.worldbank.org).

Information about cancelled projects on the PPI database also provides an indicator of the sustainability of this policy transfer. During the period 1990-2001 there were 48 cancelled private infrastructure projects, equivalent to only 1.9 % of the nearly 2,500 such projects. These cancelled projects represented 3.2 % ($24.2bn) of the total investment commitment over the period ($754bn) [33]. More than one-third of cancellations were toll road projects in Mexico. Only twelve cancellations were located in aid- dependent countries. The many reasons given for cancellation varied among sectors. In transport the main reason was that roads attract insufficient numbers of users to meet optimistic traffic forecasts. Controversies over tariff hikes in line with cost-recovery pricing were the main reason in the case of water and electricity. Interestingly, over half of all cancellations were involved projects for which competitive tendering had not taken place. In these cases cancellation resulted from political and citizen opposition that focussed on allegations of corruption because of the lack of transparency in the award. The number of cancelled or distressed projects for the whole period 1990-2006 rose to 234, accounting for 6 % of the total number and 8 % ($109bn) of the total investment. However, by the time of the latest PPI overview in November 2007, the sector failure rate for the total PSP investment over the period 1990-2006 has increased significantly, as follows: water (33 %), transport (11 %), energy(10 %) and telecom (4 %) (PPI 2007). These figures suggest a sharp increase in the overall failure rate from 2001-2006.

Anecdotal evidence can indeed point to some highly aid-dependent countries, such as Tanzania and Bangla Desh, where NPM-type managerial reforms of the public sector have been strongly adopted. But similarly, the uptake of such reforms has also been strong among other countries such as Singapore and Jamaica that are not aid-dependent. And in some highly aid-dependent countries there has been a strong adverse reaction, leading to the reversal/abandonment of such reforms. A comparison between the experience of two highly aid-dependent countries, Ghana and Bolivia, is instructive in this respect. In 1983 they became the first two countries to embark on a Structural Adjustment Programme (SAP) under strict IMF supervision. While Ghana undertook a range of NPM-type managerial reforms [46; 47], reforms efforts in Bolivia focused on establishing the legal basis for building a meritocratic bureaucracy though the 1999 promulgation of the first ever civil service statute and the creation of a Civil Service Commission in 2001 [80].

Two wider questions are pertinent to the relationship between aid dependence and the adoption of managerial reforms. First, the current thrust in foreign aid programmes towards improving `state capacity' raises the question of `capacity for what'? There is widespread recognition that a `capable' state is essential for the good functioning of a market-based economy. This places emphasis on the role of the state in addressing the delivery of goods and services with high social benefits such as basic health and education and targeted poverty reduction. But in the context of the development state the understanding of `capable' goes far beyond these functions that are more akin to a `welfare state'. Here `capable' is more about the skills needed to bring about the technological transformation of the productive base of the economy towards a growing insertion in the global economy through the export of high value-added goods and services.

Second, it may be asked whether a developmental state can be built at all by means of external assistance [73]. The classic developmental states were not highly dependent on foreign aid. Instead there was an emphasis on growing domestic resource mobilisation through the strengthening of fiscal institutions based on a mix of social trust and coercion. It can be argued that the external dependence of many LMICs, accentuated by the rapid increase in foreign aid to achieve the MDGs, is hindering the very transformation of the relationship between the state and its citizens (i.e. the `fiscal contract'), which is a central feature of the developmental state. If governments continue to rely for their revenue needs primarily on foreign aid, then they become accountable primarily to aid donors and not to their own domestic constituency. There is not then the same degree of pressure to build domestic legitimacy among their own citizens and to devise credible institutions and policies - which together provided the bedrock of the developmental state [57]. public management bureaucracy

The Developmental state and its bureaucracy

Following the ebb of the neo-liberal tide in the 1990s, there has been a renewed recognition of the crucial role of the state in the development process. This is exemplified by the 1997 World Development Report's re-affirmation that “the state is central to economic and social development” [86]. This has sparked a growing interest in the nature and behaviour of the state during the rapid growth experience of East Asian countries, starting with Japan, the so-called `Asian Tigers' (South Korea, Taiwan, Hong Kong and Singapore), and more recently, the PR China. This interest has spawned an extensive academic literature [23; 32; 81; 40; 85]. This literature nailed the coffin of the minimalist state, pointing to its theoretical flaws and its disastrous consequences for economic growth.

Although there is no single explanatory model for the emergence of the developmental state, a number of factors have been proposed that help to `explain' this process, such as strong political will and leadership as well as the uprooting of traditional elites and associated land reform. These factors have been contrasted with the prevalence of clientelism and patronage-based states in Latin America and neo-patrimonialism and predatory states in Sub-Saharan Africa, where `state capture' by rent-seeking elites has stymied efforts to erect a `developmental state'[51; 12].

This paper does not address this debate concerning the underlying causes of the emergence of the developmental state. Rather it is concerned with the wider policy implications of a central feature of the `developmental state' that distinguishes it from its counterparts in most LMICs. This is the so-called `embedded autonomy' of the state, which comprises two key ingredients [28]. First is a strong commitment to a professional bureaucracy of the Weberian kind, both meritocratic and career-based, that ensures a corresponding degree of autonomy from direct political interference. It is staffed by `principled agents', with a professional commitment based on a public service ethos, a strong sense of mission and esprit de corps. Second is the way in which this bureaucracy is embedded in the wider society through extensive interaction with selected parts of the private sector and engagement with civil society. This meritocratic civil service, insulated from clientelistic networks yet also responsive to society, has been the `transmission mechanism' of the development process spearheaded by strong political leadership and the will to succeed. The core tasks entrusted to the Weberian bureaucracy under this `governed market' development strategy involves responding flexibly to private sector needs without protecting vested interests, developing technology and high-quality basic education, state control of finance and the mobilisation domestic savings for development, identifying priority projects and giving them access to cheap credit and foreign exchange for imports[81].

The autonomous and embedded nature of the bureaucracy has been crucial for the implementation of the two `dynamic' and transformative roles of the developmental state [27]. First is the `midwifery' role of steering, assisting and inducing the private sector to innovate by providing a `green- house'-protected environment through conditional and limited tariff protection. Second is the `husbandry' role of prodding and supporting the private sector (through high quality education, research and development) so that it can compete internationally. These two roles are additional to the two conventional roles of the state - the `custodian' or `caretaker' role of the minimalist state that is limited to law and order, basic rules and regulations, and the `demiurge' or `producer' role of providing collective goods and basic infrastructure, often in competition with the private sector [27].

It is argued that the weak civil service systems in Latin America and SSA are ill- equipped to carry out even these conventional roles of the state, never mind the `dynamic' and transformative roles required of it by the developmental state. Here, the formal trappings of a Weberian bureaucracy are usually in place - with civil service laws throughout Latin America and Public Service Commissions in much of SSA - all designed to ensure meritocratic recruitment and promotion. However, these structures are manipulated by ruling elites who use public administrators as a source of patronage and a basis for building networks of clientelism. A most blatant expression of this subversion of the Weberian ideal is the common practice of using public administrators to `bring in the vote' at election time.

As mentioned in Section 1, the NPM-type managerial reforms derived from a theoretical critique (public choice and principal-agent) of the capacity of the traditional Weberian bureaucracy to promote growth and development. Yet there is evidence of a positive relationship between the degree of `Weberianness' of state bureaucracy and economic growth [28]. These writers measured the degree of `Weberianness' by focussing on the core attributes of such a state bureaucracy - meritocratic selection, promotion and reward. These variables had the virtue of highlighting the extent of autonomy of public administrators from interference in the form of patronage appointments and corruption by their political masters. While the NPM paradigm sought to address this principal-agent problem through the introduction of performance related contracting, the `Weberian' solution involved a more internalised form of control based on the professional ethics of staff recruited on the basis of merit. An analysis of 29 LMICs over the period 1970-1990 also found evidence of a strong positive relationship between `Weberianness' and poverty reduction [34], based on the view that the extent of poverty reduction is also an externality of a competent bureaucracy.

The effectiveness of managerial reforms in enhancing developmental state capacity

In this section we examine the extent to which the managerial reforms associated with the NPM are consistent with the particular aims and objectives of the developmental state. We do this by assessing the likely impact of these reforms on three of the distinguishing characteristics of the public management systems of developmental states - centralisation, high-level strategic planning and trust- based relations with the private sector.

Centralisation

Decentralisation is a central aspect of the NPM-type managerial reform agenda and features prominently on the aid donor agenda for public management reform in LMICs. Using Rondinelli's classic typology of decentralisation, we can say that in the NPM deconcentration is pursued in an extreme form through the establishment of executive agencies, delegation is pursued through the advocacy of private sector participation in service delivery and devolution is pursued by support for the transfer of service delivery responsibilities to sub-national tiers of government, and [68].

Yet it would seem that the highly decentralized governmental structures implicit in devolution are in sharp conflict with the centralist administrative structures that characterise successful developmental states. This is particularly so in the case of the devolutionary form of decentralisation, under which powers are transferred to lower tiers of government with a high degree of political, fiscal and administrative autonomy from central government. Devolution is promoted as a means to empower local citizens to participate in the development process, a style of governance at variance with the centralised and socially exclusion that characterised the developmental state. Without external control, decentralisation can also lead to a reduction in the meritocratic basis for civil service recruitment, especially in countries where the independence of the central Public Service Commission is already limited.

In fact both South Korea and Taiwan have relatively weak systems of local government while Hong Kong and Singapore function as city-states. In Korea governors and mayors for provincial and local governments were not elected until 1995. Until then local governments were no more than local administrative districts of central government. Local government heads were appointed by central government and their capacity for autonomous decision-making was virtually nonexistent.

Interestingly, Chile and Costa Rica, the two main contenders in Latin America for the title of `democratic development state' retain highly centralised systems of government, at variance with the regional trend towards devolutionary forms of decentralisation. Costa Rica has a long tradition of centralised government and was the last country in the region to introduce elected mayors, in 2002. Municipalities in Chile do have major service delivery responsibilities in education, basic health and social welfare, but carry these out under a strict `principal-agent' relationship with the respective sector ministries [58].

As the appeal of the developmental state increases, so it is possible that the recent donor vogue for decentralisation will wane. This will be assisted by increasing donor fatigue at the surprisingly poor results of donor-assisted decentralisation programmes [22] and the growing evidence of capture by local elites [73 P. 617]. The classic criticism of decentralisation focuses on the argument that it promotes increasing horizontal inequality [66]. At a time when LMIC governments are seeking to meet the MDGs, the attraction of a centralised approach to service delivery in order to reduce horizontal inequalities would seem to be considerable.

High-level strategic planning

The NPM subscribes to the view that the state should play an active role in promoting economic development. Yet the core objectives of the managerial reforms in the public sector have an innately `static' quality about them. They are concerned primarily with promoting competition in order to improve efficiency, equity and effectiveness by providing `the right incentives' in the allocation of public sector resources. There is no explicit concern for the `dynamic' considerations of providing `the right incentives' for expanding the productive base and technological level of the domestic economy. Concern for these issues is only indirect, based on the view that improvement in the allocation of public sector resources can underpin and encourage dynamism by the private sector, as measured by increased domestic and foreign investment. But ultimately the private sector is viewed as the engine of economic development without requiring any `steering' from the state.

This very `distanced' relationship between the actions of the state and the wider process of economic transformation is in sharp contrast to the interventionist nature of the developmental state, which has an explicit concern for `dynamic' change, namely how can the state best contribute to expanding the productive base of the economy and drive technological change in order to raise productivity levels and living standards.

Obviously the distinction between the two positions is more nuanced than the blunt picture presented above. However, in one major area of public expenditure - education - this sharp difference between the `static' NPM approach and the `dynamic' approach of the developmental state comes sharply into focus. Here, a notable feature of the developmental state has been the priority that it attaches to `qualitative' change in the direction of promoting technical and scientific education, which seeks to transfer and root technological advances from the global economy to the domestic arena. By contrast, the more `quantitative' approach underlying the NPM-type managerial reforms in the education sector has prioritised the use of blunt and general indicators, such as classroom size, teacher qualification levels, school autonomy and examination results.