Introduction
For several decades now, Western democracies have been engaged in an ongoing struggle with public finance problems, such as excessive public deficits, accumulating sovereign debts and a lack of resources for development, leading many countries to cut local spending and raise taxes. (The Fiscal Stability Treaty approved by a majority of EU member states in March 2012, which requires ratifying states to enact balanced budget laws, is one of the latest salvos in this struggle). The new public management (NPM) approach, first promulgated in the 1980s, aimed to address these problems by encouraging the adoption of ideas, values and practices from the private to the public sector (Martin, Reference Martin2002; Kingdom, Reference Kingdom2003; Beeri, Reference Beeri2012a). The NPM ethos holds that by modernizing public-sector institutions and processes in line with private-sector practices, governments can enhance their efficiency while still improving the quality of their services. The public debate on this issue has created high expectations regarding the quality of public goods and, particularly, local services.
During the 1990s and 2000s, underperforming local authorities have been in the spotlight and have continued to keep policy makers and managers busy (Cornforth & Paton, Reference Cornforth and Paton2004). National governments across different countries with different political regimes – including the United States (Hambleton & Sweeting, Reference Hambleton and Sweeting2004), the United Kingdom (Boyne, Reference Boyne2006), Germany (Kuhlmann, Bogumil, & Grohs, Reference Kuhlmann, Bogumil and Grohs2008) and Japan (Kudo, Reference Kudo2003) – have been keen to find ways to achieve organizational turnaround. Some scholars (e.g., Walshe, Harvey, Hyde, & Pandit, Reference Walshe, Harvey, Hyde and Pandit2004; Boyne, Reference Boyne2006) have hailed turnaround management strategies (TMS), which have been successfully used in the private sector, as a means to help struggling public-sector bodies achieve recovery (Pandit, Reference Pandit2000). The abbreviation TMS refers to a unique set of actions taken by the organizational leadership ‘to bring about a recovery in performance in a failing organization’ (Pandit, Reference Pandit2000), including actions to reduce or cut unproductive activities, to rechannel the resources freed up by such reductions and to restructure the organization as needed to improve its efficiency (Boyne, Reference Boyne2004; the three sets of strategies will be discussed further below). However, most studies of turnaround have failed either to draw sufficiently from extant theory or to develop useful new theory on the subject (Pandit, Reference Pandit2000). Further, as noted by numerous scholars (e.g., Barker & Duhaime, Reference Barker and Duhaime1997; Paton, Reference Paton2003; Cornforth & Paton, Reference Cornforth and Paton2004; Jas & Skelcher, Reference Jas and Skelcher2005), our understanding of cycles of failure and recovery is still marred by many gaps – particularly in relation to public bodies. To be more precise, it remains unclear whether TMS are either needed or useful in the public context.
The current research takes a step towards enhancing our understanding of these cycles while, first, clarifying the institutional and structural barriers to public recovery, and second, empirically exploring whether TMS actually bring about recovery in poor-performing local authorities. Using insights from institutional theory and the resource-based perspective, along with empirical research on changes in the government sector, we argue that TMS have poor chances to succeed in the public context. Furthermore, the failure of TMS as a strategic tool to heal failed bodies in the public sector tells us something deeper about the inherent limitations of NPM. In a nutshell, we argue that while TMS are suitable for organizations, government bodies are more often than not institutions. Being institutions, they are charged with political considerations and power relations, and less readily amenable to actions such as cutbacks, downgrades and restructuring. In short, the difficulty of implementing TMS in the public context serves as an example of how government reforms, including those in the spirit of NPM, are likely to fail if we do not conceive of improving public management first of all as a political goal, and only secondarily in financial terms.
TMS and Recovery
A turnaround situation has been described in the literature as a low-probability, high-impact situation that lasts for long periods and that can be seen as a decline in performance (Schendel, Patton, & Riggs, Reference Schendel, Patton and Riggs1976; Hofer & Schendel, Reference Hofer and Schendel1978; Meyer & Zucker, Reference Meyer and Zucker1990; Pearson & Clair, Reference Pearson and Clair1998). Scholars agree that turnaround situations are a consequence of multiple and interrelated factors (Balgobin & Pandit, Reference Balgobin and Pandit2001), some of them potentially external (i.e., political and economic) and others internal (these tend to be bureaucratic in nature) (Slatter, Reference Slatter1984; Grinyer, Mayers, & McKieran, Reference Grinyer, Mayers and McKieran1990; Shamsud, Reference Shamsud2002; Boyne, Reference Boyne2006). Importantly, failure in the public context does not necessarily involve a decline in financial resources, but is characterized first and foremost by a deterioration in the quality of public services (Pearson & Clair, Reference Pearson and Clair1998).
TMS are generally classified into 3Rs: Retrenchment, Repositioning and Reorganization (Boyne, Reference Boyne2004). Retrenchment deals with efficiency and stability. It may involve reducing the scope of the organization's activities – for instance, by selling assets, contracting out services to external providers or cutting the costs of services (Boyne, Reference Boyne2006). Retrenchment is chiefly reactive in nature and must be supported by proactive activities (Robbins & Pearce, Reference Robbins and Pearce1992). Repositioning is considered the complementary strategy to retrenchment, as it involves rechannelling the resources freed up when unproductive activities are cut. Repositioning emphasizes growth, efficiency and innovation (Schendel & Patton, Reference Schendel and Patton1976; Borins, Reference Borins1998). It requires organizations to redefine their mission and core activities in order to actively improve existing services, or to introduce new services to either current or new consumers (Boyne, Reference Boyne2004).
The third R, reorganization, typically involves intensive internal organizational change designed to reshape the structure of the organization. This can include replacing key leadership figures and shifting or replacing personnel (Thain & Goldthorpe, Reference Thain and Goldthorpe1989; Mueller & Barker, Reference Mueller and Barker1997), as well as making structural or cultural changes within the organization. Reorganization may stand alone, or may be intended to support other TMS. In private-sector literature, the most cited reorganizational act is the replacement of senior managers (Walshe et al., Reference Walshe, Harvey, Hyde and Pandit2004).
Conceptually, TMS are implemented in a manner that is far more radical, urgent, centralized and hierarchical than ordinary improvement/change strategies (Beeri, Reference Beeri2012a, Reference Beeri2013a). When TMS are implemented, five main conditions tend to hold. First, the strategies are introduced in order to bring about radical and extensive change. Second, they tend to be introduced in an urgent manner. Third, they tend to come under intensive monitoring and scrutiny. Fourth, they tend to be executed by either poor or new management. Fifth, they are implemented in an unproductive organizational climate (Pearce & Robbins, Reference Pearce and Robbins1993; Balgobin & Pandit, Reference Balgobin and Pandit2001; Mellahi, Jackson, & Sparks, Reference Mellahi, Jackson and Sparks2002; Boyne, Reference Boyne2004; Paton & Mordaunt, Reference Paton and Mordaunt2004; Turner, Skelcher, Whiteman, Hughes, & Jas, Reference Turner, Skelcher, Whiteman, Hughes and Jas2004; Walshe et al., Reference Walshe, Harvey, Hyde and Pandit2004; Boyne, Reference Boyne2006; Beeri, Reference Beeri2009).
A review of empirical studies on turnaround-related organizational dynamics in the private sector shows that on average, TMS have had a positive impact (Boyne, Reference Boyne2006). Indeed, the literature suggests that TMS are more likely than other sets of strategies to successfully achieve a radical improvement in organizational performance and outcomes when implemented by management after a crisis (Schendel, Patton, & Riggs, Reference Schendel, Patton and Riggs1976; Boyne, Reference Boyne2004; Walshe et al., Reference Walshe, Harvey, Hyde and Pandit2004; Boyne, Reference Boyne2006).
In the public sector, turnaround may be a political act as much as a managerial one, and so may raise extensive resistance (Walshe et al., Reference Walshe, Harvey, Hyde and Pandit2004). Opposition may arise in relation to any stage in the process of identifying and reacting to a crisis: its start and end points, its multiple interrelated causes, its severity and consequences, and whether a crisis even exist or, if it exists, whether it is necessary to do anything about it (Mordaunt & Cornforth, Reference Mordaunt and Cornforth2004). Even if most stakeholders agree on the broad outlines of a crisis, any attempt to implement the three sets of TMS (the 3Rs) may trigger resistance at the political level. Political obligations and considerations may hinder retrenchment and repositioning, making it difficult to stop funding non-profitable activities (Boyne, Reference Boyne2004). Internal stakeholders may also resist repositioning because it requires flexibility in moving between areas and markets and innovative thinking, while managers in failing public organizations are usually mired in traditional work patterns (Boyne & Meier, Reference Boyne and Meier2009). The political obstacles that make retrenchment and replacement difficult in the public sector (cronyism and tit-for-tat arrangements) are even more apparent with regard to reorganization (Meyer & Zucker, Reference Meyer and Zucker1990). In addition, of course, the replacement of elected members in local authorities is subject to political and democratic processes (Boyne & Meier, Reference Boyne and Meier2009).
Early researchers noted that the roots of TMS lie in the private sector (e.g., Schendel, Patton, & Riggs, Reference Schendel, Patton and Riggs1976; Hofer & Schendel, Reference Hofer and Schendel1978), and these strategies have primarily been studied in the context of large American firms (e.g., Ketchen, Reference Ketchen1998; Balgobin & Pandit, Reference Balgobin and Pandit2001; Mellahi, Jackson, & Sparks, Reference Mellahi, Jackson and Sparks2002). Despite this, a number of scholars have suggested that many of the concepts and findings regarding decline and recovery that have been developed in the for-profit sector have at least some application in public organizations (Boyne, Reference Boyne2004; Cornforth & Paton, Reference Cornforth and Paton2004; Joyce, Reference Joyce2004; Mordaunt & Cornforth, Reference Mordaunt and Cornforth2004; Paton & Mordaunt, Reference Paton and Mordaunt2004; Walshe et al., Reference Walshe, Harvey, Hyde and Pandit2004; Jas & Skelcher, Reference Jas and Skelcher2005; Turner & Whiteman, Reference Turner and Whiteman2005). Indeed, there is both theoretical and empirical work to backup such claims. First, the few empirical studies of public-sector turnaround offer evidence that TMS are widely employed by poor-performing local authorities and can be valuable (Boyne & Meier, Reference Boyne and Meier2009; Beeri, Reference Beeri2013a). Second, in terms of theory, an initial look at institutional theory suggests that strategies that have succeeded in the private sector can be adapted to the public sector. After all, we are dealing in both cases with bodies that may have similar structures and similar goals – to serve people and meet their needs.
However, empirical evidence on failure and turnaround within the public context is rare (Walshe et al., Reference Walshe, Harvey, Hyde and Pandit2004; Beeri, Reference Beeri2013a, Reference Beeri2013b), making the effectiveness of these strategies in the public sector still largely an open question. In addition, a closer look at institutional theory and, more specifically, the question of legitimacy suggests that the effectiveness of TMS may in fact be quite limited in the government sphere. We now turn to institutional theory before presenting our research design and empirical findings.
Barriers and Obstacles to Public Recovery: An Institutional Perspective
From an organization theory standpoint, it seems clear that not every outcome is the result of a conscious decision process (March & Olson, Reference March and Olson1983). Thus, the selection and implementation of TMS in the public context may end not only in turnaround, but in either continued failure or, in the worst cases, terminal decline (Miah & Mia, Reference Miah and Mia1996; Cabrero-Mendoza, Reference Cabrero-Mendoza2000; Boyne, Reference Boyne2006). Theoreticians in the area of turnaround suggest various models for cycles of decline and recovery, most of them derived from the business arena (e.g., Pearce & Robbins, Reference Pearce and Robbins1993; Arogyaswamy, Barker, & Yasai-Ardekani, Reference Arogyaswamy, Barker and Yasai-Ardekani1995; Harker & Sharma, Reference Harker and Sharma2000; Balgobin & Pandit, Reference Balgobin and Pandit2001; Chowdhurry, Reference Chowdhurry2002; Shamsud, Reference Shamsud2002). McKiernan (Reference McKiernan2003) suggested a model whereby both private and public organizations undergo cycles of failure and recovery. The framework includes six stages that reflect an ideal throughway to organizational recovery: causes and symptoms of failure, triggers for change, diagnosis and plan formulation, organizational stabilization, recovery and renewal. Yet, this process is fraught with risk, as at each stage things can go wrong and lead to a deeper failure than in the first place. As will be detailed next, a variety of factors that have a particularly strong presence in the government sector and political context may interfere and thwart attempts at government sector recovery (Turner & Whiteman, Reference Turner and Whiteman2005; Boyne & Meier, Reference Boyne and Meier2009).
Paradoxes of power relations and politicization
Central government, as a superordinate authority, has the legitimate power to regulate, constrain and oversee the authorized powers decentralized to local authorities. Accordingly, national governments are likely to create bureaucratic arrangements that centralize discretion at the top and allow relatively little autonomy to local managers (Simon, Reference Simon1983; Scott, Reference Scott1987; Eisenhardt & Martin, Reference Eisenhardt and Martin2000). Such arrangements constrain municipal efforts at recovery and lead to a pattern of coercive isomorphism (Pilcher, Reference Pilcher2011; on coerciveness as a mechanism of change in general see DiMaggio & Powell, Reference DiMaggio and Powell1983). Coercive isomorphism in local authorities results from both formal and informal pressures exerted on the local leadership by the central government. In other words, organizational change and recovery processes are direct responses to government dictates and not necessarily a reaction to local incentives (Lodge & Wegrich, Reference Lodge and Wegrich2005).
The environment in which local authorities operate is complex, creating gaps between people's needs and local resources, on the one hand, and democratic and managerial expectations, on the other hand. Moreover, local authorities must frequently contend with conflicting policies from above. As noted by McNulty and Ferlie (Reference McNulty and Ferlie2004: 1409):
There is a rich seam of irony here: previous public sector reforms launched in the name of better management have strengthened vertical lines of authority and reporting… They are still embedded and may be proving antithetical to the growth of post-NPM process-based organisations supported by a new policy rhetoric. The resilience of these ‘deep structures’ predicts continuing limitations to reform strategies designed to achieve ‘big bang’ change in public service organisations.
There is a triple contradiction here that tends to strongly hinder the effective implementation of TMS. First, local authorities may be caught in a double bind by central imperatives to ‘show strong leadership’ but, at the same time, to ‘do as we say’ (Turner & Whiteman, Reference Turner and Whiteman2005). Second, local authorities must manoeuvre between two ideals: the ideal of a local authoritative leadership with the formal and informal power to declare a ‘state of emergency’ and to institute urgent and radical reforms, and the ideal of governance and pervasive local participation, which requires respecting the people's voice. Third, local authorities tolerate the expectation that with the transfer of responsibilities through decentralization, local managers will remain both professional decision makers and, simultaneously, accountable and responsive – that is, politicians.
Massive government intervention may reflect an aspiration to pursue values and standards that symbolize the pure spirit of NPM. But the more the central political system perceives local government as a contractor rather than a partner, the more resistance there is likely to be from key stakeholders (Saunders, Reference Saunders1984; Kingdom, Reference Kingdom2003). These stakeholders – including not only members of the local authority, but also external interest groups (e.g., workers’ unions facing massive layoffs and contractors that have to compete by tender) – may use their position and power to blunt the influence of government agencies even if they are not themselves involved in planning and implementing reforms (McNulty & Ferlie, Reference McNulty and Ferlie2004). In short, both local authority members and interest groups may, for their own reasons, fail to co-operate with national plans and policies (Meyer & Zucker, Reference Meyer and Zucker1990). They may reject effective TMS and retard full organizational recovery even when these serve national goals of improvement and modernization.
In addition to this, national policy initiatives offer double message that ultimately result in incoherent reform within public-service organizations. Without abandoning the theme of strong leadership, policy makers speak grandly about bringing power back to local people by renewing local democracy and revitalizing civic government. This approach promotes a rhetoric of ‘managed networks’, ‘partnership’, ‘joined-up government’ and ‘governance’, which reflect a lateral rather than a vertical perspective towards organizational change.
Ideally, both policies may work simultaneously. Nevertheless, public recovery strategies, particularly retrenchment, require strong leadership able to make radical decisions (Paton & Mordaunt, Reference Paton and Mordaunt2004) that are likely to damage the interests of some key partners. Moreover, a poor-performing local authority, which is likely to suffer from deterioration in the delivery of services, reduced financial support, mismanagement and internal conflicts, would probably find it difficult to balance two policy types at the same time (Shamsud, Reference Shamsud2002; Boyne, Reference Boyne2006). The concern here is that poor and politicized leadership, which is subject to great pressure, will prefer unrestrained partnerships in order to survive, while avoiding taking tough decisions that are necessary for recovery (Turner & Whiteman, Reference Turner and Whiteman2005).
Avoiding the use of unpopular TMS may result in the recovery process becoming partial, biased and consequently ineffective. Skipping the retrenchment phase may mean reducing the resources available for reinvestment in productive areas – that is, repositioning – along with unnecessary reductions in productive activities. Hence, while TMS can be adopted and implemented as a radical top-down managerial intervention designed to revolutionize the local authority, it may be incompatible with the governance tendency to share responsibilities and powers and thus, be antagonistic to effective public recovery.
The cost of the marketization approach
The practices that accounted for the success of science in modernity, in particular the act of ‘attentive witnessing’, also accounted for the rise of liberal democracy (Ezrahi, Reference Ezrahi1990). In a similar vein, to increase the legitimacy of institutional change, liberal democracies since the 1980s and 1990s have been led to draw on managerial knowledge accumulated in the private sector under principles of strict marketization, transferring private-sector practices to public institutions as if in a laboratory setting. Yet, to be effective, a recovery plan must be organic, not artificial (McNulty & Ferlie, Reference McNulty and Ferlie2004). That is, the diagnosis must be translated into operative actions while taking into account local needs and circumstances (Pearce, Robbins, & Robinson, Reference Pearce, Robbins and Robinson1987). Thus, the adoption of strategies based on private-sector managerial practices according to strict ‘scientific’ principles may carry a high price (Mordaunt & Cornforth, Reference Mordaunt and Cornforth2004). In the contentious atmosphere that characterizes the governmental sphere, especially where central–local relations are conflict ridden, such requirements can have a negative impact on the selection and quality of TMS.
In addition, central inspection takes into account a narrow definition of accountability based on fixed performance indicators dictated by national agencies, rather than on more legitimate factors such as local targets and priorities (Brooks, Reference Brooks2000; Wilson, Game, Leach, & Stoker, Reference Wilson, Game, Leach and Stoker2006). Furthermore, some local leaders are pushed to ‘tick boxes’ when choosing TMS. In other words, local leaders may be pushed to select strategies that generally result in measurable outcomes and satisfy governmental agencies’ representatives, a situation that reflects a ‘compliance culture’ (Turner & Whiteman, Reference Turner and Whiteman2005). Such imposed changes are likely to be more superficial, and organizational managers are likely to be less committed to them, in comparison with strategies selected based on internal considerations (Scott, Reference Scott1987).
In sum, a harsh political atmosphere may motivate local authorities to adopt TMS. Nevertheless, the same agenda may affect the suitability of these strategies and result in resistance to principles of improvement and innovation – thus reducing the effectiveness of TMS for public recovery.
Managerial and local barriers: The resource-based perspective
Resources are defined as tangible and intangible assets that affect the organization's ability to implement strategies to improve its efficiency and effectiveness (Pandit, Reference Pandit2000). Tangible assets are property based and include plant and equipment, geographic location, physical technology and access to raw materials. Intangible assets are knowledge based and include organisational routines and the abilities of individual employees (Miller & Shamsie, Reference Miller and Shamsie1996). Organizations in recovery face the problem of having to balance the exploitation and exploration of resources in order to ensure reasonable performance in the long run and maintain the possibility of radical recovery (Pandit, Reference Pandit2000). Thus, taking into account that by definition key resources are rare, difficult to imitate and valuable, the effectiveness of TMS in failing local authorities may be determined also by the interplay of contextual forces, including resources and intra-organizational dynamics. There are internal resource-based systems, both behavioural and political, which differentiate poor from excellent performers and persistently poor from recovered local authorities (Carmeli & Cohen, Reference Carmeli and Cohen2001).
In addition to resources, a recovering organization may have to overcome further internal barriers that decelerate change, including denial (Boyne, Reference Boyne2004), success traps (Turner et al., Reference Turner, Skelcher, Whiteman, Hughes and Jas2004) and persistent failure (Meyer & Zucker, Reference Meyer and Zucker1990). The recovering organization must perceive at least one event as a trigger for change, which usually takes place at the point of the widest gap between aspirations and actual performance (McKiernan, Reference McKiernan2003). Then, it may stimulate (i) preliminary diagnosis of the causes of failure and determining present needs, (ii) search for new TMS to prevent terminal decline, (iii) selection of TMS and lobbying for legitimacy and (iv) implementation of TMS (Boyne, Reference Boyne2006). Yet, what can guarantee that an event is perceived as a trigger for change that stimulates these strategies?
Possibly, radical organizational change requires a commitment to a culture of either competition or reformation, so that less-effective departments or positions are eliminated or restructured. In short, commitment to change is required for recovery (Greenwood & Hinings, Reference Greenwood and Hinings1996). However, turnaround involves real and personal costs that may create organizational barriers to this commitment (Turner & Whiteman, Reference Turner and Whiteman2005) – particularly in the government sector, which functions within a political context. Since radical organizational change is likely to result in public cutbacks and downsizing, rational stakeholders may lobby vigorously for keeping the status quo to protect their jobs and power base (Meyer & Zucker, Reference Meyer and Zucker1990).
Since public and private organizations are fundamentally different, practices and tools from the private sector cannot be directly transferred to public management tasks (Boyne, Reference Boyne2006). As McNulty and Ferlie (Reference McNulty and Ferlie2004) point out, cloning managerial practices from the private sector is not always helpful, and may produce limited achievements in organizational transformation. In accordance with the need for commitment to reformation, radical change cannot occur unless local managers, politicians and decision makers have the knowledge and experience to take action, especially in the case of strategies characterized by being radical, urgent, intensively monitored and implemented in what is most likely an unproductive organizational climate – that is, TMS (Boyne, Reference Boyne2004; Paton & Mordaunt, Reference Paton and Mordaunt2004; Walshe et al., Reference Walshe, Harvey, Hyde and Pandit2004). However, knowledge and experience regarding public failure, recovery and turnaround often fall sadly short (Greenwood & Hinings, Reference Greenwood and Hinings1996). As Cornforth and Paton (Reference Cornforth and Paton2004: 198) point out, ‘little is known about whether, how often and why “normal” cycles of [public] decline and recovery may be arrested’. Thus, public managers lack the appropriate tools to predict failure in their organization, prevent it and generally deal with a public crisis after it has occurred (Paton, Reference Paton2003). Indeed, the gap in our knowledge regarding the recovery of failing public organizations is not accidental. Jas and Skelcher (Reference Jas and Skelcher2005) and Cornforth and Paton (Reference Cornforth and Paton2004) explain that this gap is rooted in government policy and the predominant ethos, which is focused on excellence: excellent performance and excellent performers. The focus on excellence has centred research on good performers, creating serious gaps in our understanding of the processes at work in failing public organizations (Meier & Bohte, Reference Meier and Bohte2003).
Furthermore, beyond the required but absent commitment to a culture of reformation described earlier, radical organizational change requires a culture of competition. Yet, competition may become destructive rather than constructive. Local authorities may compete with one another in a way that hinders appropriate co-operation. They may be unwilling to share principles of best practice for fear of losing their competitive advantage (Bevir, Reference Bevir2001). But without knowledge sharing, local authorities cannot benefit from the experience and hard-won insights of others.
In sum, there seem to be three significant barriers to public recovery in the public sphere, some intrinsic to local authorities and some extrinsic: power relations and the politicization of unintended results; paradoxes that result from the tensions between science and politics; and a lack of organizational commitment, managerial tools, experience and knowledge, which may bring about the adoption of inapplicable TMS. Added to these intrinsic managerial barriers, elements of national policy and the political atmosphere can further impede public recovery. Accordingly, we hypothesize that:
Hypothesis 1 : In the public-sector context, greater employment of TMS will not result in recovery.
Method
Background and research setting
The current research took place between 2002 and 2006 in the United Kingdom, where public policy under the Conservative and New Labour parties has been influenced by NPM practices. Consistently, local authorities were expected to modernize services while public spending decreased. The Local Government Act of 1999 – a response to the functional and economic failure of some local authorities – created a framework whereby the government could audit, inspect and intervene in local affairs (Wilson et al., Reference Wilson, Game, Leach and Stoker2006). (The Act applied only in England and Wales, where local authorities act under the principle of Ultra Vires, meaning that authorities may operate only within the scope of powers explicitly granted by the national legislature). Audit and control functions were to be conducted chiefly through the Audit Commission (AC) (settled by the Audit Commission, 1998), which worked alongside other regulators.
Throughout the 2000s, the AC was heavily involved in what became a complex and erratic relationship between the central and local governments. The problematic nature of this relationship can be seen in the government's frequent shifts to new mechanisms of audit and control, from the best value regime to the comprehensive performance assessment to the comprehensive area assessment. BV – the central performance management framework for local government during the early 2000s (Bovaird & Downe, Reference Bovaird and Downe2006) – required local authorities to agree on performance plans and report performance against a set of national indicators, with provision for central intervention. Then Prime Minister Tony Blair declared that his government would be quick to intervene if local authorities failed to carry out their mission effectively (Brooks, Reference Brooks2000).
By the end of 2001, early results indicated that best value was potentially a disaster. Inspection reports had categorized 60% of local services as poor or fair and 41% were described as unlikely to improve (Bovaird & Downe, Reference Bovaird and Downe2006). Consequently, the comprehensive performance assessment was introduced in 2002, with the aim of more effectively evaluating councils’ overall performance capabilities. At the same time, significant extra intervention was imposed on local authorities judged poor or weak in order to enforce desired behaviour. However, according to the comprehensive performance assessment reports for 2002 through 2005, among all the poor and weak performers in 2002, 12% had made no progress at all after 3 years of effort, and 38% had achieved only minor gains. These ratings raised questions about the effectiveness of the AC (Game, Reference Game2006). Accordingly, in 2009, the comprehensive performance assessment was superseded by the comprehensive area assessment, which shifted its focus to the combined efforts of local councils and partners in particular geographic areas, and from processes to the achievement of outcomes (Bell, Reference Bell2008). Yet, the comprehensive area assessment has also suffered from strong criticism, on the grounds that it has failed to reduce the burden on councils and is poorly co-ordinated (Williams, Reference Williams2009). In late 2010, the Secretary of State for Communities and Local Government disbanded the AC and ordered new arrangements for auditing England's local public bodies.
Research design, population and sample
This study takes a quantitative survey-based approach to test the hypothesis developed earlier in this paper. The population of this study was defined as English local authorities that have suffered from organizational failureFootnote 1. The sample consisted of 34 poor- or weak-performing local authorities based on their AC performance rating from 2002. These 34 local authorities were split into two groups according to their 2005 ratings: 17 persistently failing (stagnated or moved up only one category) and 17 recovered local authorities (improved their status by two or three categories) (see Table 1).
Sample groups of local authorities
The quantitative survey was based on a questionnaire distributed to 553 senior leaders of 34 underperforming local authorities. Eighty-six senior councillors and executive directors responded, a response rate of 15.4%. Of those who responded, 56% were from persistently failing and 44% from recovered local authorities. In terms of demographics, 82% were male, their average age was 54.2 (SD = 9.99), they averaged 16.4 years of education (SD = 3.45) and average tenure in the position was 6.7 years (SD = 7.85). Eighty-three per cent were councillors and 17% were non-elected officials. A comparison of respondents and non-respondents was carried out to test for non-response bias. T-tests and χ2 tests revealed that in terms of the authority's performance ranking and the local leader's position, the respondents were distributed in accordance with the potential sample as a whole, indicating that the response rate did not skew the results.
Data collection, measurement tools and data analysis
To test the hypothesis, following Beeri (Reference Beeri2009), the items of the TMS scale were drafted into a questionnaire. The survey measured the extent of implementation of TMS, based on the participants’ own experience with their current local authority over 2002–2006, on a 5-point Likert scale (1 = ‘hardly at all’, 5 = ‘to a very large extent’). Tests confirmed the scale's validity and reliability (Cronbach's α = 0.89).
An exploratory factor analysis – a reduction of a large number of items to the smallest number of latent meaningful dimensions – was employed. First, to confirm the appropriateness of exploratory factor analysis, the factorability was tested and was found to be good. The Kaiser–Meyer–Oklin value (KMO = 0.810) exceeded the recommended value of 0.6, and Bartlett's test of sphericity reached statistical significance (p < .001), both indicating relatively compact patterns of correlations between the items and data that should yield distinct and reliable clusters of items. The items had roughly normal distributions, which is another good sign for factorability. An inspection of the correlation matrix of all items revealed the presence of significant inter-correlation coefficients of 0.3 and above (p < .001 to p < .05) between the items. These findings indicate that the items measured the same dimensions of content, supporting the use of exploratory factor analysis.
Second, exploratory factor analysis analysis showed that the proposed scale was consistent with the 3Rs theoretical structure of TMS. The scale yielded eight factors, reflecting three internal dimensions (the two figures in brackets show eigenvalues and % variance, respectively): (i) retrenchment of services (3.963, 11.92); (ii) retrenchment of expenditures (1.384, 4.19); (iii) repositioning as reaching out (2.276, 6.89); (iv) repositioning as innovative services (1.537, 4.65); (v) repositioning as renewed relationship (1.205, 3.65); (vi) reorganization at the institutional level (9.706, 29.41); (vii) reorganization at the personnel level (1.159, 3.51); and (viii) reorganization as extent of centralization (1.852, 5.61). These factors explained 69.8% of the total variance, which is quite large.
Third, following Milfont and Duckitt (Reference Milfont and Duckitt2004), and in order to reduce the number of variables in the full regression model, we used the eight first-order factors to extract second-order factors. The analysis revealed clear content similarities with the original 3Rs within three factors (again, the figures in brackets show eigenvalues and % variance, respectively): (i) repositioning and reorganization at the institutional level (2.748, 34.35); (ii) retrenchment (1.460, 18.25); and (iii) reorganization at the personnel level and centralization (1.389, 17.36). These factors had a total loading of 70%.
The variable recovery was defined as a dummy variable for moving up two categories or more and was based on data published by the AC. An additional operational definition was employed: extent of recovery, as improvement in terms of ratings from 2002 to 2005. In addition, we controlled for three characteristics of the local authorities: their size (measured by the log of population; M = 5.44, SD = 0.21); their type (1 = ‘unitary authorities’ – towns or urban centres with one level of local government; 0 = ‘county councils and district councils with a two-tier structure of local government’); and their relative deprivation. For this last variable, we used as a proxy the Index of Multiple Deprivation for 2004 (Office of the Deputy Prime Minister, 2004). This index is based on seven indicators of deprivation: income, employment, health, education, barriers to housing and services, crime, and living environment; M = 28.83, SD = 10.33. Each of the controlled variables potentially affects the likelihood of recovery for any given local authority.
Statistical analyses, including independent samples t-tests, Spearman's correlation coefficient and hierarchical multiple regression were employed using SPSS 18.0 software. In addition, we conducted a power analysis and hierarchical linear modelling using HLM 6.08.
Findings
Given the theory of cycles of turnaround (McKiernan, Reference McKiernan2003; Boyne, Reference Boyne2006) and common sense, one might think that greater implementation of TMS would contribute to organizational recovery. However, this section shows that, as we hypothesized, greater employment of TMS in the public-sector context does not necessarily result in recovery.
The first set of analyses employed independent sample t-tests that compared mean scores of TMS within persistently failing and recovered local authorities (see Table 2). The t-tests reveal that for overall TMS and most of the factors there was no significant difference between the groups. Moreover, reorganization at the personnel level was significantly implemented to a higher extent (t = 2.328, p < .05) within persistently failing local authorities (M = 4.21, SD = 0.66) than within recovered local authorities (M = 3.74, SD = 1.09).
T-tests for TMS scores for persistently failing and recovered local authorities
Note. TMS = turnaround management strategies.
*p < .05.
To examine the relationship between TMS and extent of recovery in our sample, three tests were employed. First, Spearman's correlation coefficient analyses revealed that greater implementation of TMS did not necessarily produce greater recovery. Furthermore, there was a significant negative correlation between reorganization at the personnel level and extent of recovery (rs = −0.286, p < .01). In other words, the greater the extent to which a local authority implemented reorganization at the personnel level, the less it recovered.
Second, we employed power analysis to confirm that our study provides adequate statistical power (Cohen, Reference Cohen1988) for testing the effect of TMS on recovery. Power analysis is required to avoid a Type-2 error (i.e., a false finding that the groups do not differ, or in our case, that there is no relationship between greater employment of TMS and the extent of recovery) – a critical step in our study, given that we framed Hypothesis 1 as a null hypothesis, that is, expecting a non-significant relationship. The power of the test is influenced by three factors: sample size, effect size and α level (Stevens, Reference Stevens1996). The power analysis revealed that our N is larger than the required sample size, and that our study provides adequate statistical power (0.85 with an 0.05 α) to test the effect of TMS on recovery.
Finally, we employed a hierarchical multiple regression analysis. In the first step of the equation, the local authority-level variables were included in the model to control for the size, type and deprivation of the local authority (see Table 3). Extent of recovery was negatively related to population size, implying that smaller local authorities are more likely to recover. Alternatively, since recovery is incremental and gradual (Price & Chahal, Reference Price and Chahal2006), recovery may slow as local authorities grow in size. This finding, by itself, is interesting and in line with Boyne and Meier (Reference Boyne and Meier2009), who found that large school districts were less likely to achieve recovery.
Findings of hierarchical multiple regression analysis (standardized coefficients) for the relationship between TMS and extent of recovery
Note. TMS = turnaround management strategies.
N = 86.
**p < .01, ***p < .001.
In the second step of the equation we added the three higher-order TMS and tested for a relationship between these strategies and extent of recovery; no positive relationship was found. Moreover, a negative relationship was found between reorganization at the personnel level and centralization and extent of recovery (β = −0.35; p < .01). This interruption to recovery will be discussed in the next section. Overall, the findings support our hypothesis, which expected that greater employment of TMS would not lead to organizational recoveryFootnote 2.
Discussion and Conclusions
The first contributions of this research are empirical. In contrast to the predominant rhetoric of excellence, this study offers observations on the ineffectiveness of TMS in the public sphere. It is as yet unclear what relative weight should be assigned to extrinsic and intrinsic factors in diminishing the effectiveness of TMS, much as scholars still struggle to weigh the external and internal roots of public failure. Probably, a mixture of interrelated factors is at fault.
Our results suggest that, despite arguments that successful recoveries tend to follow a path of top-management replacement (e.g., Bibeault, Reference Bibeault1982; Slatter, Reference Slatter1984; Stopford & Baden-Fuller, Reference Stopford and Baden-Fuller1990), and in line with other studies (see Daily & Dalton, Reference Daily and Dalton1995; Boyne & Dahya, Reference Boyne and Dahya2002), persistently poor performers execute reorganization at the personnel level and centralization (i.e., ‘replaced senior and middle managers’, ‘changed internal organizational structure’ and ‘took centralisation steps’) to a greater extent than recovered public organizations. These results point to the manifest differences between public and private organizations in this regard. In the private context, as suggested by Upper Echelon theory (Hambrick & Mason, Reference Hambrick and Mason1984; Carpenter, Geletkanycz, & Sanders, Reference Carpenter, Geletkanycz and Sanders2004) and the top management team approach, long-tenured management is less effective, especially in uncertain and confusing situations like failure and turnaround. Indeed, in the private-sector literature, the most cited reorganizational act is replacement of senior managers (e.g., Thain & Goldthorpe, Reference Thain and Goldthorpe1989; Walshe et al., Reference Walshe, Harvey, Hyde and Pandit2004), and this leads to successful recovery over 80% of the time (Schendel, Patton, & Riggs, Reference Schendel, Patton and Riggs1976). However, in public administration, political restrictions and democratic processes can complicate replacement of the top leadership (Boyne & Meier, Reference Boyne and Meier2009). In addition, there is evidence that in the public context internal leaders are both more experienced and more innovative than leaders brought in from outside to replace them (Borins, Reference Borins1998). Accordingly, Borins (Reference Borins1998) argues that in the public context, internal leaders should lead the recovery process.
Taken as a whole, it appears from our findings that the combination of over-replacement together with over-centralization is destructive to public recovery. While new managers are freer to introduce organizational change, too extensive, quick or frequent replacement of top personnel may mean that TMS are executed prematurely by a centralized, unqualified and inexperienced leadership. Thus, replacement should be balanced with other tactics (such as stabilizing a productive organizational climate, gaining subordinates’ trust, employing restrained decentralization or using organizational learning techniques) before introducing radical policy changes such as those inherent in TMS.
The research exposes a further interesting relationship between TMS and recovery in poor-performing local authorities. Some poor performers improve their rankings while others persistently stagnate. Yet, both groups appear to implement TMS to a large and, more often than not, similar extent. The explanation for our findings may lie in the internal resources of the organizations under investigation (Pandit, Reference Pandit2000). It may be the case that some managers simply make poor judgements as to the limits of the resources available to them. They therefore may overextend their resources, delay pursuing a balance between exploitation and exploration of resources, or mistakenly choose to reposition resources to new areas rather than refocusing resources on current core activities (Wild, Reference Wild2010).
At the theoretical level, the current research clarifies barriers to radical improvement in performance inherent in the political environment. We have tried to show theoretically and empirically that implementing TMS in government organizations is first and foremost a political task. That is to say, successful reform requires successful politics, including the appropriate balance between responsibility and authority. The recovery of failed local authorities really depends not only on power, but also on legitimacy. The alternatives, then, are radical change that accepts the dependence on consent, or radical change that is based on pure power. Both alternatives may succeed or fail, but only the alternative which accepts that radical recovery is a political task can also be legitimate. Furthermore, this paper suggests that normative considerations may have an effect on the probability of success. In other words, if we do not consider turnaround and TMS in political terms, we may not be able to understand public failures and promote desired changes.
Looking deeper into this study's results, it is clear that government agencies and national inspection programmes are key players in the processes studied. Such agencies and programmes attempt to measure performance through inputs, outputs and outcomes, and they have a hand in all three sets of explanatory factors discussed in this research. They execute national policies, they shape the political atmosphere and they accompany managerial decision making.
Taking into account that, as a whole, the legitimation of power, constraints and oversight imposed by central government on local authorities did not generally result in effective recovery, our findings provide evidence for institutional theory in its worst form. Central government seems to hold the role of a powerful super-authority that mandates structural reforms and practices on local authorities. Yet, the heavier the pressure on the municipal leadership to adopt national practices – that is, TMS – the less TMS are adapted to local needs and the more they are blurred, distorted, biased and distanced from local incentives. This conclusion goes hand in hand with that of Pilcher (Reference Pilcher2011), who researched national financial reforms imposed on Australian local authorities. The author concluded that institutional coercive isomorphism is apparent in Australian local government, and that NPM may have been pushed to the extreme in that case. Under the pressure imposed by external stakeholders, local councils complied with national policies. However, as with TMS, the usefulness of the mechanisms employed were found questionable. Clarifying these points is particularly important given the worldwide economic crisis in the second decade of the 21st century.
Another point worth mentioning is that this research highlights the dual role of any government agency in charge of a national inspection programme. On one hand, as a quasi-governmental body and a non-interested party, such an agency is charged with independently inspecting, measuring and reporting on public performance. On the other hand, it advises and supervises poor performers, putting it in charge of promoting recovery. This dual role creates a potential for bias. Any government agency in charge of a national inspection programme has a vested interest in showing improvements in poor-performing organizations in order to justify its existence. The result is further top-down pressure on poor performers.
The implication of this finding is that implementation of TMS without serious political work is likely to be useless. In addition, municipal managers should be aware of the mixed messages arising from trends in national policies – for example, strong leadership and the contrasting ideas of governance, local democracy, bringing power back to local people and civic government. Since the governance model and decentralization risk giving disproportionate power to groups inside and outside the local council, this model may encourage weak control mechanisms (Ionita, Reference Ionita2005). Moreover, the skills needed for co-production are related to persuasion and consultation rather than dictatorial decision making. Thus, considering that TMS must be implemented in a manner that is radical, urgent, centralized and hierarchical (Boyne, Reference Boyne2004; Cornforth & Paton, Reference Cornforth and Paton2004; Beeri, Reference Beeri2012a), and in line with Harker and Sharma (Reference Harker and Sharma2000), municipal managers should adopt throughout public recovery processes a national policy of strong leadership and give it preference over governance (yet remember in this context that according to our data, over-centralization employed together with over-replacement is destructive to public recovery). After organizational stabilization and the first signs of recovery and renewal are achieved, this preference may be reconsidered.
Our findings suggest that government administrators and municipal managers should pay more attention to the very first stages of organizational recovery – that is, diagnosis and plan formulation. During this sensitive phase, those charged with coming up with the diagnosis and plan must withstand pressure from higher tiers of government to ‘tick boxes’ and choose TMS based on fixed national priorities. Rather, striving for municipal recovery means putting local priorities ahead of satisfying government agencies or surrendering to a ‘compliance culture’. Only this can ensure that in importing TMS from the private sector, planners choose an appropriate set of well-ordered and co-ordinated strategies.
Another trap to avoid in planning turnaround is automatically applying massive layoffs. Some public administrators employ reorganization at the personnel level because it looks like the only possible thing to do. Later, the very new management brought in to replace them poorly implement other TMS and cause further deterioration. More often than not, when large-scale layoffs are implemented, local authorities have already cut redundant staff members, and further cuts may impede rather than support attempts at recovery. Reorganization at the personnel level was found to have a negative impact on the organizational climate and to weaken bonds among staff members (Beeri, Reference Beeri2012b). Thus, municipal managers should employ reorganization at the personnel level with caution, and should balance it with decentralization. In particular, it should be employed as a last resort if the true target behind public replacement is cutting local expenses through retrenchment – a strategy whose value to recovery itself is doubtful (see, e.g., Hambrick & Schecter, Reference Hambrick and Schecter1983; Barker & Mone, Reference Barker and Mone1994). Instead, municipal managers should seek alternative means of reducing manpower-related spending, such as encouraging voluntary early retirement, freezing executive salaries, initiating shutdowns of the municipality during the summer and consolidating agreements with employee unions on contribution of days off.
Another point should be borne in mind. According to AC data along with the analyses employed here, public recovery appears to be gradual and not revolutionary. Indeed, revolutionary change requires incremental progress (Price & Chahal, Reference Price and Chahal2006). Moreover, the findings strengthen March & Olson's (Reference March and Olson1983) argument that comprehensive governmental reorganization does not have a major impact on administrative costs, efficiency or control and rarely accomplishes its stated objectives. In other words, while TMS are probably necessary for recovery, they are almost certainly not an exclusive factor. Thus, as noted by Stoker (Reference Stoker2004), government administrators should back up local recovery with appropriate support, employing strategies aimed at informing and guiding local leaders and not merely detecting, pursuing and punishing problematic behaviour.
Finally, the limitations of this study should also be mentioned. The Local Government Act of 1999 and the AC method may extend some of the barriers to recovery to their fullest. Put differently, national government policy and the national political atmosphere might be less salient for public organizations in other settings than suggested here. Hence, though the effect of TMS on English local authorities’ recovery is questionable, TMS might be feasible and more effective in other settings. This possibility should be supported by empirical evidence in future studies. Second, the division between failing/recovering authorities is somewhat arbitrary, given the complexity of public performance. As Jas and Skelcher (Reference Jas and Skelcher2005) observed, the performance of public organizations over time is typically cyclical. Third, our sample is of modest size, quite homogeneous and share features common to the English political environment. This fact limits our ability to generalize the findings to other public organizations and cultures. Hence, future studies should try to control internal organizational factors (e.g., causes of failure, leadership and role) and external organizational factors (e.g., eligibility for national grants, residents’ socio-economic status, relationship with government agencies) that may further explain what differentiates recovered from persistently poor performers.
