Are policy tools and governance modes coupled? Analysing welfare-to-work reform at the frontline

ABSTRACT

This paper considers the link between policy tools and governance modes – the characteristic ways frontline staff are meta-governed. It asks: Are substantive policy tools coupled to procedural tools (governance modes) that can guide local service delivery agencies and the work of individuals delivering welfare services? The substantive policy tools in this case are those typically utilised to reform welfare-to-work services: contracting-out of services and competitive tendering, and the regulation of quasi-markets. These are hypothesised to flow through to procedural policy tools in the form of corporate and market incentives and regulatory (bureaucratic) methods that shape how work is done (governance modes), privileging certain practice orientations at the frontline. Policy makers seek to shape these meta-level governance modes because they should result in systemic change, based on a reconfiguration of policy actors and their interrelationships, for both service delivery agencies and the individuals working in them. We identified four ideal-type governance modes (bureaucratic, corporate, market and network) and tracked which of these were dominant in-practice at the frontline in Australia and the UK at two levels: office and personal, at four points in time (1998, 2008, 2012 and 2016). We found that the dominant mode of organisation at the office level was corporate, followed by bureaucratic in both nations. But the bureaucratic mode had grown in strength over time, particularly in Australia, and as a personal priority for staff, as re-regulation occurred. The results indicate a coupling between substantive policy tools and governance modes at the frontline of welfare-to-work.

KEYWORDS:

• Policy tools
• governance modes
• welfare-to-work
• frontline staff

Introduction

The way welfare services are both imagined and enacted has changed dramatically in most Organisation for Economic Co-operation and Development countries over the last thirty years (Fimreite & Lægreid, 2009). Driving such changes has been the shift in our understanding of welfare payments and the role of the service delivery office. First and foremost, welfare benefits are no longer seen as an entitlement. Welfare payments now come with obligations. To be eligible for payments, job seekers should be actively looking for work and be able to demonstrate this activity; hence, the emergence of distinctive welfare-to-work regimes made up of certain common tools.

Second, while the mechanism for welfare support would traditionally involve a public bureaucracy administering income support payments or access to other public resources such as housing, using rules and standard operating procedures to handle eligibility and redress, by the end of the 1980s, there developed a distrust of the permanent bureaucracy by the political class. This development rendered ‘bureaucracy’ a universally pejorative term, leading many reformers to re-think the ‘how’ of welfare service provision, which allows a variety of new delivery instruments to be tried. The achievement of this vision would potentially be enabled by a desire by system designers to separate strategy from delivery or in the language of the times, to focus upon ‘steering not rowing’.

Foundational to these deliberate changes to the production, consumption and distribution of welfare-to-work services are New Public Management (NPM) inspired ‘substantive’ policy tools (Howlett, 2000) that notably involve contracting out of services and the introduction of contestability of purchasing. Procedural policy tools then follow (Howlett, 2000). For instance, output targets, performance-based contracts, and outcome budgeting (Christensen & Lægreid, 2001; Considine, 2001; Hood, 2005) are deployed to steer the networks of actors involved in policymaking (Leik, 1997) and in managing service delivery (Kickert, Klijm, & Koppenjan, 1997). This new menu of incentives and regulatory devices would include attempts to alter the behaviour of jobseekers and employers and to empower the work of frontline or ‘street-level’ staff who: ‘constitute the services delivered by government’ (Lipsky, 1980, p. 3). This meta-level system steering, we call governance modes – conceiving of them as procedural policy tools.

While governance modes have long been a focus of research on welfare-to-work (see, for Considine and Lewis (1999, 2003); Considine (2001); Considine, Lewis, O’Sullivan, and Sol (2015)), this research has mainly examined governance modes as ideal-types and compared how these have changed over time in Australia, the UK, and other nations. The link between policy tools, substantive and/or procedural, and governance modes, has not been a central concern in that previous research. In this paper, we aim to understand the link between government attempts to reform a policy system at a high level (which can be considered substantive policy tools based on the spectrum of Howlett and Ramesh (1995) cited in Howlett (2000)), and the consequent effects on governance modes at the frontline of service delivery (procedural policy tools). We ask: Are substantive policy tools coupled to procedural tools (governance modes) that can guide local service delivery agencies and the work of individuals delivering welfare services?

Welfare-to-work reform

In reforming welfare-to-work since the mid 1990s, employers were seen to need greater engagement by agencies seeking to place jobseekers who had various barriers to employment. Jobseekers themselves were seen to require complex support, coaching, and even discipline to traverse the journey from welfare to paid work. And the private and public agencies delivering these services and their frontline staff would want the maximum profit from their government contracts. Employment services exemplified most of the core features of the NPM changes being rolled out around the world since the 1980s (Aucoin, 1995; Aberbach & Christensen, 2001). They would no longer simply distribute or allocate programs and entitlements in line with rules and budgets. They would rather enact a program logic somewhat of their own making under contracts that had a ‘black box’ character, creating a menu that was locally determined. Consequently, frontline work, it was assumed, would be transformed.

The reform, by design, drives a normative and pragmatic shift in the delivery of welfare-to-work services (O’Toole (2000) as cited in Howlett, 2000). At the normative level, NPM-driven incentives are the new language and currency for steering policy processes in the desired direction. Given an assumption that staff and clients are highly responsive to the new policy delivery ambitions, the frontline work is expected to be influenced in ways that are characteristic of core NPM reform themes – we call them ideal-type governance modes in this study. The ways in which frontline staff manage their daily work given the incentives they face in practice are emergent and constitute the governance orientations we have tried to empirically measure in this research.

This paper is structured into four additional sections. Next, we outline the successive rounds of welfare-to-work reform in Australia and the UK and discuss the corresponding ideal-type governance modes that are signalled by these reforms. We then briefly describe the development of governance mode measurements in the methods section, and report research findings on how well any of these accounts of ideal-type governance modes fit the reported frontline practices in the two countries. Following this we make some concluding remarks.

Welfare-to-work reforms in Australia and the UK and corresponding ideal-type governance modes

The choice of Australia and the UK is driven by the common institutional arrangements and policy goals and the shifts in policy in each case over time (although not at the same time in both nations), which helps isolate the impact of specific governance instruments. The UK employment services framework is comparable to Australia's, but while the UK was escalating its use of black box methods, Australia was moving in the other direction. Including both countries in this study thus enables analytical comparison of the methods used in the two countries while allowing for a more nuanced understanding of the coupling of reform policies and governance modes, both ideal and emergent. Table 1 provides a summary of the major reforms occurring in the lead up to each of the four rounds of surveys we conducted in the two countries, and the main governance modes signalled by those reforms are shown in bold. A more detailed description of the reforms in both nations follows, and the governance modes are further explained in Table 2.

Table 1. Major reforms to welfare-to-work systems in Australia and the UK (adapted and expanded from Considine et al. (2015)) and governance modes

Survey year	Australia	UK
1998	1994 – Labor Government introduces Working Nation, semi-privatised system, one-third of services contracted out, 300 private providers begin operating 1998 (after survey) – Job Network launched by Conservative Government, most services outsourced, ‘mutual obligation’ introduced, underpinned by ‘work first’ principle	1997 – New Deal launched by Labour Government, contracting out to private services for some functions/clients 1998 (after survey) – green paper introduces idea of ‘making work pay’
	Corporate, Market	Corporate, Market
2008	1999 – second round of tenders for Job Network 2001 – Star ratings (performance measures) introduced, tied to contracts 2003 – all employment services privatized	2000 – Employment Zones established in 15 high unemployment areas 2001 – ‘work first’ welfare state declared 2002 – Jobcentre Plus launched, maintains public sector focus on monitoring and enforcement, about one-third contracted to private sector 2007 – Freud report on ‘reducing dependency’, introduced prime contracting model
	Corporate, Bureaucratic	Corporate
2012	2009 – Job Services Australia launched by Labor Government 2011 – services centralized and consolidated into a single contract, financial incentives to focus on harder-to-help strengthened	2009 – Flexible New Deal launched with prime providers 2010 – Conservative Coalition Government cancels Flexible New Deal 2011 – Work Programme launched, move to a prime-contracting model with 18 prime providers 2012 – introduced performance measures based on job outcomes, tied to (re)allocation of funds 2013 – Universal credit introduced, replacing 6 existing benefits for people of working-age. Benefit payments reduced, in line with working hours reported by the providers, tougher conditionality and the ratcheting up of sanctions
	Corporate, Bureaucratic	Corporate, Bureaucratic
2016	2015 – Jobactive introduced by Conservative Government with emphasis on payment by results, work for the Dole, collaboration bonus and minimum job search requirements (44 providers)	Next major policy changes in 2017 (post our survey)
	Corporate, Bureaucratic	Corporate, Bureaucratic

Table 2. Governance modes

	Source of Rationality	Form of Control	Primary Virtue	Service Delivery Focus
Bureaucratic	Law	Rules	Reliability	Universal Treatments
Corporate	Management	Plans	Goal-driven	Target Groups
Market	Competition	Contracts	Cost-driven	Price
Network	Culture	Co-Production	Flexibility	Clients

Adapted from: Considine and Lewis (1999)

Australia

Embracing NPM ideals, in 1994, the Australian government introduced Working Nation – the first among multiple initiatives to search for a more efficient and effective system for public employment services delivery. Key features that distinguished Working Nation from previous programs were: (1) service delivery was partially contracted to private providers; (2) a case management approach was used for service delivery, according to which jobseekers would be treated as ‘whole clients’, barriers to employment would be addressed and service provision would be tailored to the clients’ needs (Considine, 2005); (3) the adoption and implementation of the principle of reciprocal obligation as a condition of welfare receipt (Thomas, 2007); and (4) the development of an outcome-driven mechanism for rewarding the service providers.

In 1998, the newly elected Coalition Government argued that the Working Nation programs were ‘expensive, poorly targeted, and ineffective’ (Fowkes, 2011, p. 5). The Government thus created the Job Network (JN), to address these problems. Retaining key features of the Working Nation, the JN comprised three contracts, with the first one running from 1998 to 2000, the second from 2000 to 2003; and the third from 2003 to 2006 with an extension of 3 years from 2006 to 2009. The reforms under the JN saw the closing of the training programs (Considine, 2005) and of the more extensive government interventions of the Labor years, particularly ‘job creation efforts as vehicle for very long-term unemployed people to gain work skills and engage in employment’ (Fowkes, 2011, p. 6). JN was designed around ‘work first’ principles, with the fundamental objective to ‘get people into full time work and get them there fast’ (Fowkes, 2011, p. 6).

The next rollout – Job Services Australia (JSA) in 2009 aimed to address the shortcomings of JN. The reformers attacked the fragmentation and the complexity of services, the decreased flexibility in service delivery and a jobseeker compliance regime that was said to be overly punitive (Perkins, 2008, p. 4). JSA changes included adjustments to outcome payments incentives; the introduction of incentives to encourage providers to work closely with local employers; and a less aggressive compliance system (Considine, O’ Sullivan, & Nguyen, 2014). However, underlying such changes remained the fundamental architecture and associated assumptions laid out in the JN. The changes introduced under JSA represented modifications, instead of fundamental shifts as had occurred with the transition from Working Nation to JN.

The same could be said about the current program – Jobactive, which replaced JSA in July 2015. With no significant variations to JSA’s key features, Jobactive instead promised a shift in the intensity of the service experience, primarily related to greater flexibility in service delivery, even more focus on outcomes, reduced competition via a smaller number of providers, and the introduction of a provider collaboration bonus to try to meet the criticism that contracted agencies would tend to advance their own interests rather than cooperate in the interest of their clients (Department of Employment, 2014).

Notably, starting with case management in the 1990s, under which private providers enjoyed a considerable degree of autonomy, the Australian system had become more and more regulated over time (Considine, Lewis, & O’Sullivan, 2011), with providers subject to various quantitative Key Performance Indicators, including the ‘famous’ Star Rating system introduced in 2001 (Jobs Australia, 2015).

UK

Since 1998, the UK’s employment services sector has also undergone several reforms: New Deals (NDs), introduced in 1997 and supplemented by Employment Zones (EZs) in the early 2000s; EZs from 2000 until 2009; Flexible New Deals (FND) between 2009 and 2011; and the Work Programme (WP) which commenced in 2011 and was replaced by the much smaller Work and Health Program (WHP) in 2017. Before 2000, private providers only delivered certain agreed parts of the service menu, such as job search or short-term vocational training, while the pivotal task of job brokerage rested with the public agency. At first, these private providers acted as sub-contractors to public agencies.

However, when the EZs were launched in 2001, the involvement of private actors expanded and intensified (Lane, Foster, Gardiner, Lanceley, & Purvis, 2013). For the first time, they were commissioned to deliver the whole programme and were awarded ‘all of the relevant provision in the area for which they won the bid’ (Finn, 2005, p. 8). This change resulted in direct contestability of the services delivered by the public service (JCP) and the private agencies in some designated cities. The market share enjoyed by private for-profit and not-for-profit agencies significantly increased under WP, including more than 20 programmes, 40 contracts and 18 regional areas under WP (Department for Work and Pensions, 2012). The WHP, however, saw a decrease in the scale of contracting out when delivery changed to just five primes across six regions for jobseekers with disabilities only. The provision of employment support to jobseekers on unemployment benefits had been largely returned to the public agency, Job Centre Plus (JCP).

Over time, the reform in the UK also saw the introduction and deepening of outcome-based payment, the black box approach to service management and emphasis on integrating services for jobseekers with the needs of local labour markets and employers. Specifically, the concept of outcome-based payment, first introduced in the 1980s (Lane et al., 2013), intensified following the Freud Report in 2007. In 2012, WP was described as ‘a major new payment-for-results welfare-to-work programme’ (Department for Work and Pensions, 2012, p. 2). Payment-by-results continues to underscore the WHP’s contracting model, with outcome payments accounting for approximately 70% of the total payments available to providers (Department for Work and Pensions, 2018).

Further, the earlier NDs were more process driven and their norms for agency conduct required certain standard actions to be taken with all clients. In 2000, when EZs were outsourced to the private sector, the providers did not have to deliver a prescribed program (Johnson, 2012). This might be considered the UK’s first serious attempt at a black box model. However, according to Johnson (2012), it was not really a black box at all, with many standard requirements remaining. With the introduction of WP, the black box approach became more pronounced. Unlike FND, no mandatory service components were prescribed in WP contracts and agencies were free to decide which interventions to offer to help participants into sustainable employment (Lane et al., 2013, p. 9).

Table 1 lists the major reforms occurring in the lead up to each of the four rounds of surveys we conducted in the two countries. Our surveys in Australia correspond to Working Nation (as the first survey was conducted before the introduction of Job Network in 1998), Job Network (the third contract), Job Services Austrlia and Jobactive. In the UK, they correspond to New Deals, Employment Zones, and Work Programme when it was first rolled out and finally wound up.

This narrative of reform indicates that the transformation of both the Australian and the UK systems has in fact unfolded through a gradual process of institutional layering, displacement, and conversion. Instead of simply replacing one reform and a single governance mode with another, each iteration has mainly involved layered changes to the contract management and funding model, as well as the jobseeker compliance framework, with the quasi-market design remaining an enduring feature of both countries’ approaches to welfare-to-work program delivery (Considine et al., 2014). Considine and Lewis (1999) found governance modes, at both ideal-type and in-practice levels, that corresponded to all three core NPM reform themes: corporate management, pro-market, and network (or ‘joined up’), in addition to a traditional bureaucratic mode. This was in relation to one round of reform in Australia. Hence, it seems that each round of reforms, as well as the layering of multiple reforms over time, is likely to produce mixed modes of governance with no single mode dominating frontline work.

Table 2 describes the four ideal-types of governance that we first postulated in the 1990s. Traditional bureaucracy is characterized as using laws and rules to deliver services (bureaucratic). Newer NPM-driven variants rely on strengthening the central direction through plans and targets (corporate), others seek to use competition and performance contracts to motivate actors (market), while others emphasize the mechanisms of joint working (network) to get the desired results.

An extended discussion of this typology can be found in Considine (2001) and Considine and Lewis (1999, 2003). Without recounting all this earlier discussion, an important point to note is that these – like all ideal-types – are core properties of constructs, which sometimes overlap at the edges of practice and in the rhetoric of protagonists. The corporate and market modes exemplify this and can often be found combined with conceptual descriptions of NPM, as well as in practice. Likewise, the flexibility that is claimed as a primary virtue of networks in this four-way classification can equally be claimed to be a virtue of quasi-markets. And as noted above, layering and mixing occurs over time.

Methods

To examine the governance modes that are practiced at the frontline in response to a change in substantive policy tools, we developed measures of the ideal-type governance modes and surveyed frontline staff in Australia and the UK at four different points in time. The survey was first developed and refined between 1996 and 1998 following policy document analysis and on-site interviews with staff in each jurisdiction. It was subsequently updated and repeated in both countries in each of 2008, 2012 and 2016. It comprises nearly 100 questions in total, which are focused on how service providers operationalise welfare-to-work policy and how frontline staff orient themselves to their work and conduct their tasks on a daily basis while helping jobseekers find employment. The number of respondents in each country and at each survey period is shown in Table 3.

Table 3. Number of respondents included in the analyses of governance constructs¹

Year of survey	Australia	UK
1998	537	133
2008	883	808
2012	783	143
2016	680	211
Total	2883	1295

¹ In each survey round from 2008 onwards, prizes have been offered as an incentive to staff to complete the survey. For the original surveys it was possible to gain reasonably accurate figures for the total population of interest and to calculate response rates. This has become more difficult over time as the number of agencies has contracted into smaller numbers with multiple offices and often loosely estimated (or no) information provided about the numbers of people employed in these offices. Our best estimates (reducing in certainty over time) are that response rates were: 56% (all countries 1998); 44% (Australia 2008); 45% (UK 2008); 45% (Australia 2012). We have no information on the total sample size for the UK in 2012 or for either country in 2016.

All our respondents were frontline staff working for one of the contracted employment service providers in Australia and the UK. The employment services staff surveyed at each point in time all worked directly with jobseekers to help make them job-ready, place them into work or help them maintain their job post-placement. Given the length of time between surveys, and the turnover of staff in these roles, there is little chance that the same staff were surveyed in different time periods. Staff working at the frontline are in a strong position to understand and assess the dynamics of the service delivery system, making clear the impact of otherwise opaque system reforms on their work. They are also able to disclose important attitudes and orientations to their work via our survey. All of the responses are, of course, their own reported perspectives, which provide indicators of their behaviour, but do not disclose their actual behaviour. While these ‘street-level bureaucrats’ (Lipsky, 1980) might be expected to express support and some slight bias in favour of a system they help to implement, they arguably have less incentive to distort the system’s virtues and vices than either clients or managers, owners of the agencies or the government purchaser of employment services.

To construct items that constitute the governance scale, in the first of our surveys, frontline employment services staff were interviewed, and surveyed. In the interviews, they provided the key descriptions of the imperatives driving their work, and we converted these into a battery of 40 statements for use in the survey questionnaire. They include items regarding the role played by rules, the use of discretion, the nature of supervision, which kinds of incentives matter most, how technology impacts, and the use made of outside agencies when assisting clients. The original 40 items that constitute the governance scale were divided into four groups of 10 statements that were considered to belong to each of the four ideal-type modes (bureaucratic, corporate, market and network).

The governance scales were then used in a questionnaire with a sample of 345 Australian staff working in public, for-profit and not-for-profit agencies in 1996. The exploratory factor analysis from this pilot survey indicated that there were three rather than four modes, with bureaucratic and network modes being evident, plus a third hybrid of the corporate and market modes. This initial analysis also indicated that several scale items were insignificantly contributing to the three factors that arose from the factor analysis (bureaucratic, corporate-market, and network), or were reducing the three factors’ reliability (based on Cronbach's alpha coefficients). A reduced set of 28 of the original 40 items was therefore used for the remainder of the data collection in the first set of surveys. With the results from the pilot survey being further confirmed by the subsequent analysis of the full 1998 survey dataset (Considine & Lewis, 2003), this same set of 28 scale items was then used in the 2008, 2012 and 2016 surveys.

The original list of items derived from the interviews with frontline staff in 1996 also included four core statements that were intended to reflect (in summary) the four hypothesized ideal modes. An analysis of the responses to these ‘core’ items in the pilot survey indicated that many staff agreed with all four modes. That is, frontline staff (probably quite rightly) perceived that all four of the modes were familiar to them to some extent. Many potentially conflicting signals about what is valued by the government and their own agencies are received by these staff. We therefore decided to explore this further by introducing a forced-choice question so that we could get an estimate of which was the most imperative to them. Specifically, in the later rounds of the 1998 study, and in the 2008, 2012 and 2016 surveys, staff were asked to indicate which of these alternatives they regarded as the most important priority in their office:

• knowing the rules and official procedures (bureaucratic),

• meeting the targets set by management (corporate),

• competing successfully with other service providers (market) or

• having the best possible set of contacts outside the organization (network).

Given the potential overlap among core properties of four governance modes at the edges of practice and in the rhetoric of protagonists as previously discussed, analysing the 28 scale items that constitute the governance scale proved to be insightful for interrogating the extent of their distinctiveness or, put differently, whether identifiable modes of governance could be observed in multiple reform-minded systems. Indeed, as mentioned above, in practice the four modes resolve into three with a blend of corporate management and market reforms (NPM). Amongst these, the network mode of governance was the most embryonic of the four modes when this study first began in the 1990s. Its form of organization compared with traditional procedural bureaucracy and the newer corporate and market versions of this (Considine & Lewis, 2003) was novel. Since then, much more has been written about network governance, and a growing interest in it has generated an entire branch of research (see, for instance, Lewis (2011)).

These governance modes were also found to rest on a narrative concerning the nature of complex welfare-to-work services. Reformers initially embraced the idea that the discretionary shaping of services by frontline staff, including especially those in private agencies, would support flexible and tailor-made techniques to assist clients. After the first round of major reforms, and partly because of the diverse (and sometimes perverse) organizational and individual behaviours this generated, new forms of regulation by governments were introduced, to increase the pressure on agents to adhere to the central government directives specified in contractual agreements. At the same time, the performance pressures on frontline staff to meet new goals and targets in assisting clients increased, as noted in Table 1.

The focus of this paper is on analysing the changes of dominant modes of governance across time and assessing how linked the observed effects of (procedural) policy tools are to the substantive policy tools expressed in government reforms. As discussed earlier, we recognise that there will not be a single governance mode guiding this work for agencies and the individuals who work within them. The identified changes were tracked against successive rounds of policy reforms, thereby affording insights into the links between policy tools and governance modes. Fulfilling such a research objective is therefore primarily reliant on the survey responses to the forced choice of the most important governance mode, to draw out what is having the largest orienting effect on the work of frontline staff in practice.

Results: changing governance modes in Australia and the UK

As we have already explained, the work of frontline staff is likely to be shaped by multiple governance logics. Indeed, in the initial survey, we found that frontline staff regarded each of the different modes as guiding their work to some extent, even though in some cases this would mean that they were following conflicting signals (Considine & Lewis, 1999). So, which of these takes precedence and becomes the standard set of principles that guide frontline work at any point of time? We focus on identifying dominant modes, so we can observe changes in the concentration of governance modes over time.

The responses to the forced-choice question have been compared across the two countries and the four time periods, and these data are presented in Tables 4 and 5. There are significant differences between the countries within each year (Table 4), and there are also significant differences for each country between the time periods (Table 5).

Table 4. Dominant office governance modes – Australia vs. the UK comparisons (percentages)

1998	Australia (n = 243)	UK (n = 144)
Knowing the rules and official procedures (Bureaucratic)	9.1	9.7
Meeting the targets set by management (Corporate)	58.4	88.2
Competing successfully with other service providers (Market)	21.4	1.4
Having the best possible set of contacts outside the organization (Network)	11.1	0.7
Chi-squared (3) = 51.070; P < 0.0001
2008	Australia (n = 1,098)	UK (n = 915)
Knowing the rules and official procedures (Bureaucratic)	37.6	11.4
Meeting the targets set by management (Corporate)	47.7	77.3
Competing successfully with other service providers (Market)	11.3	5.2
Having the best possible set of contacts outside the organization (Network)	3.4	6.1
Chi-squared (3) = 234.654; P < 0.0001
2012	Australia (n = 876)	UK (n = 303)
Knowing the rules and official procedures (Bureaucratic)	36.9	21.1
Meeting the targets set by management (Corporate)	54.7	72.6
Competing successfully with other service providers (Market)	5.3	4.3
Having the best possible set of contacts outside the organization (Network)	3.2	2.0
Chi-squared (3) = 30.787; P < 0.0001
2016	Australia (n = 792)	UK (n = 344)
Knowing the rules and official procedures (Bureaucratic)	30.9	23.0
Meeting the targets set by management (Corporate)	57.4	70.1
Competing successfully with other service providers (Market)	7.4	3.8
Having the best possible set of contacts outside the organization (Network)	4.2	3.2
Chi-squared (3) = 17.243; P = 0.001

Table 5. Dominant office governance modes – Country-specific inter-temporal comparisons (percentages)

	Percentage
Australia	1998	2008	2012	2016
	(n = 243)	(n = 1,098)	(n = 876)	(n = 792)
Bureaucratic	9.1	37.6	36.9	30.9
Corporate	58.4	47.7	54.7	57.4
Market	21.4	11.3	5.3	7.4
Network	11.1	3.4	3.2	4.2
Chi-squared (9) = 157.079; P < 0.0001
UK	1998	2008	2012	2016
	(n = 144)	(n = 915)	(n = 303)	(n = 344)
Bureaucratic	9.7	11.4	21.1	23.0
Corporate	88.2	77.3	72.6	70.1
Market	1.4	5.2	4.3	3.8
Network	0.7	6.1	2.0	3.2
Chi-squared (9) = 57.897; P < 0.0001

What these show us is that frontline staff absorb the dictates of new policy and governance reforms. They prove to be highly responsive to the shift in substantive modes of organizing the service, including the need to compete or cooperate in new ways. But the fit is not exact, perhaps because the new system logics are themselves prone to ambiguity. For example, the drive to get people off benefits by sanctioning bad behaviour often threatens the positive working relationship needed to get clients to engage well with the service. What shows up in the four modes discussed in this study is that despite such ambiguity and the many tensions in the policies driving the changes, coherent differences in service delivery orientation do emerge and do help explain a great deal of the difference between systems and within systems over time. These changes also provide a unique vantage point from which to assess the top-down policy changes themselves and to distinguish parliamentary rhetoric from local reality. If changing the policy requires changing the way that frontline staff are organised and directed to do their work, then it could be expected that this will also change which particular strategies they value in practice.

Table 4 shows that in 1998 the corporate mode dominated in both countries but was more dominant in the UK (88.2%), followed by Australia (58.4%). The greater market emphasis in the system in Australia in 1998 shows through clearly, with 21.4% of people saying that competing successfully is the most important priority in their office, with the UK (which at this time still had a single public provider), at 1.4%. Bureaucratic was the most consistent across the countries, being quite low and varying between 9.1% and 9.7%. The network mode was also around this level in Australia, but almost non-existent in the UK.

In 2008, the corporate mode still dominated in both countries, but had fallen for both Australia and the UK (see Table 4). The bureaucratic logic had become much more important in Australia, with 37.6% nominating this as the most important local office priority. The market mode, although low, remained more popular in Australia than in the UK. Both the market and the network modes rose in the UK to 2008. The network logic had become more common for the UK than for Australia.

The corporate mode remained dominant into 2012 with (respectively) 54.7% and 72.6% of the Australian and UK staff choosing this orientation. The bureaucratic mode remained the second most frequently chosen response in both cases, while competition had fallen as an option for the Australian staff and was now very similar to the percentage of UK staff choosing this. The network mode was chosen by a similar percentage of the Australian and UK staff in 2012. Indeed, by 2012, the only clear difference between these two countries was a stronger preference for corporate norms in the UK and for the bureaucratic in Australia. The 2016 figures are very similar, with corporate governance topping the list in both nations, but with this still being stronger in the UK, and bureaucratic coming second and remaining stronger in Australia than the UK. Chi-squared analyses of these national comparisons show significant differences between the countries in each of the time periods.

Table 5 presents these same data across time rather than across countries. For each country, the differences over time are statistically significant. They clearly show the remarkable increase in the bureaucratic mode over the decade for Australia, the relative stability of the corporate mode and the decline of both the market and network modes. Initially, this appeared puzzling, given that the system had become more marketized over this period and that there had been a shift to a fully private sector (for-profit and not-for-profit) provision of services.

We suggest, however, that this is consistent with the increasing level of regulation and oversight that was introduced as the JN evolved in Australia, driving staff to pay more attention to rules enforced by the purchaser and by their local head office. The most dramatic shifts were between 1998 and 2008, in line with the more significant changes made to contracting in that period. By 2016, there had been some reversal of the dominance of a rule-driven approach, and a rise in both the corporate and market modes, likely related to the more recent round of contracting and the new system that was in place by then.

Bureaucratic governance has re-grown steadily in the UK, particularly between 2008 and 2012, while corporate governance has fallen further each time the survey has been conducted. The increased marketization of services and the introduction of many new organizations delivering these services, compared with the single public service that was in operation in 1998, is clear in the 2008 figures, which have since been more stable. The network mode rose from 1998 to 2008 but then dropped to 2012 and rose only slightly to 2016.

The low subscription to the network mode appears as an anomaly, given extant discussions of network governance and the policy commitments of both governments in relation to ‘joined-up’ arrangements, following the fragmentation of previously unified services. Yet when frontline staff are asked to make a choice between this and other priorities in explaining what guides their work at the office level, the network mode is rarely given a high priority. The highest percentage indicating a preference for a network mode was 11.1%, recorded for Australia in 1998. This has since declined to between just 3.2% and 4.2%. It rose for the UK to 2008 but has since fallen to low levels. Clearly, very few people are viewing this as the most compelling way to explain the priorities in their office – although this is stronger at the personal level, as can be seen in Table 6.

Table 6. Dominant personal governance modes – Australia vs. the UK comparisons (percentages)

2008	Australia	UK
	(n = 1,087)	(n = 891)
Bureaucratic	45.2	28.4
Corporate	39.5	49.4
Market	7.6	7.9
Network	7.7	14.4
Chi-squared (3) = 67.754; P < 0.0001
2012	Australia	UK
	(n = 875)	(n = 298)
Bureaucratic	49.5	38.6
Corporate	37.8	42.3
Market	4.3	9.1
Network	8.3	10.1
Chi-squared (3) = 16.460; P = 0.001
2016	Australia (n = 789)	UK (n = 343)
Bureaucratic	42.8	33.2
Corporate	43.5	46.1
Market	6.0	5.5
Network	7.7	15.2
Chi-squared (3) = 19.173; P < 0.0001

We suggest that this network orientation might be more indicative of the activity associated with periods of substantial system change – hence the relatively high level in Australia in 1998 and the UK in 2008, when the systems in these countries had recently undergone major reforms. We speculate that local offices experience these changes as periods of uncertainty and seek to manage this by greater external relationship building, heightened interest in the strategies of others and greater investment in information sharing, in order to stabilise their new work environment. They then return to a different repertoire. But overall, the picture is an office-level focus upon the corporate, followed by the bureaucratic mode of governance. Markets and competition as a dominating explanation have only had any real purchase in Australia in 1998 in the first round of devolution and contracting-out.

In 2008, we added a second version of this question, and asked about staff’s individual priorities, separate from the priorities of their local office (the focus above). The results from this can be seen in Table 6. We expected that the network mode might be higher for these staff personally, even if it was not their office’s dominant governance mode. This was the case in 2008, with the percentage nominating the network orientation more than double, compared with the results in Table 7 (7.7% in 2008 compared with 3.4% in 2012 for Australia and 14.4% compared with 6.1% in the UK).

However, it is the corporate logic that still dominates in the UK in 2008. The bureaucratic mode outstrips the corporate in Australia in individual terms in contrast to office priorities. This is true in both 2008 and 2012. The bureaucratic mode was higher and the corporate mode was lower in the UK (in terms of personal rather than office priorities), but the corporate was still the dominant orientation in both periods – although not by very much in 2012. In 2016, the corporate mode had become stronger in both nations, while the bureaucratic had fallen. These two are at level pegging in Australia, while the corporate mode dominates in the UK. There is also a substantial rise in the network logic for the UK to 15.2%, taking this above its previous high-water mark in 2008 for the UK. Chi-squared analyses of these national comparisons show significant differences across nations for each of the time periods (see Table 6).

These results show a high level of individual resonance with the importance of sticking to the rules and hitting the targets in both cases, but with Australian staff adhering more strongly to the rules than their UK counterparts, who tend to be guided more by targets. There is some evidence in both countries of a stronger subscription to the network mode at the individual staff level than the office level. This fits with the notion that these actors try to make sense of the work environment they inhabit, especially in the UK, by staying in touch with what others in the sector are doing, even if this is not such a high priority for their office. But the general dominance of corporate and bureaucratic ideas is clear from this analysis.

Conclusions

This paper considered how changes at the public service frontline answer larger questions about systems of governance, including the characteristic tools they employ. We tested four different ideal-type governance modes, viewed as procedural policy tools, against the experiences of frontline staff in two different countries at four different points in time. We sought to compare the salience of different governance modes at different times, for those on the frontline of welfare-to-work systems, and to examine the links between these and the predominant substantive policy tools being used at those times.

From an agency point of view, the dominant focus clearly rests strongly on targets (corporate) and rule (bureaucratic) driven modes of delivering services. Despite the high levels of privatisation and competition, the market mode is relatively weak. Nevertheless, the changing policy settings and the instruments of contracting out, financial incentives and the increasingly tight specificity of contracts and performance measures are all impacting on the orientations of frontline staff. While the strongest fidelity is to the tools and practices of the corporate mode in terms of office priorities and personal priorities, and in both countries, bureaucratic rules remain important, particularly in the Australian case.

Indeed, sticking to the rules strengthened to become the most pervasive orientation for Australian staff in 2008 and 2012 after the system had become more heavily regulated, and local flexibility was discouraged. This indicates that the organisational environment in which they were working was strongly averse to moving outside the growing list of standards and regulations, put in place to bring the behaviour of provider organizations back into line with government expectations and desired policy outcomes. While this trend in rule following was also apparent for UK staff, it was targets that really dominated their orientation to service delivery. These two together account for around 80% of the modes orienting the work of these staff in the UK case and slightly more in the Australian case.

We uncovered a subtle difference in how frontline staff experience and use these tools, with more emphasis placed on their networks as a personal resource which might be useful to them in placing clients into work, albeit well behind the importance of meeting the targets and knowing the rules. The relatively low level of use of a network approach to the work speaks strongly about the management signals being received at the frontline and the signals being internally integrated as personal priorities, by staff. While much is said about the importance of joining up services for the benefit of clients, the signals received by staff are about meeting targets and avoiding the penalties associated with breaking key bureaucratic rules. Finally, the lower subscription to competition as a priority orientation and tool for managing work indicates that the marketization of employment services might be better described as ‘thin contracting’ where the purchaser has limited tools to direct local services and mixed motives for doing so.

Frontline staff are working in an interconnected world of disparate, profit-seeking agencies, and there is, of course, discussion about the competition between these agencies. But after years of reform and significant shifts in governance modes, it is rules and targets that appear to be most important in steering the work of those on the frontline. This fits with the return to more bureaucratic policy tools, demonstrating the tight coupling between these and governance modes in practice. The street-level bureaucracy literature (Lipsky, 1980) emerged when these individuals had significant ability to use discretion in applying policy. It seems clear that the menu from which they are now choosing is highly constrained by substantive policy and rules that limit their discretion for flexibility and experimentation.

This examination of the relationship between policy reforms in welfare-to-work over two decades describes the introduction of substantive new policy tools, which help explain the resulting governance modes that can be considered as procedural tools in Howlett et al.’s (this volume) terms. Our research suggests a coupling between these in this policy sector. It is difficult to imagine that the application of substantive tools, such as contracting-out services and applying financial incentives and performance measures, alongside increasing the level of regulation over time to steer the contracted services, will not be accompanied by substantial changes to the approaches to this work developed by local actors’ (individual and collective), expressed as different ways to work with clients and key actors. Our surveys indicate clear and observable changes in the governance modes orienting the work of these frontline service delivery staff in practice, as well as an apparent coupling of substantive and procedural policy tools.

This study is, however, not exempt from limitations. The inclusion of two countries with a similar policy toolkit in this study, and the multiple points of observation afforded by four periods of different reform strategies and four staff surveys, provides some assurance that the tight coupling is not a single nation effect, at one point in time. This enhances the strength of our research findings. But there are likely to be other factors, such as organisational culture and socialisation that contribute to the emergence of in-practice governance modes at the frontline. What is clear from this research, which aligns changes in substantive policy tools with changes in governance modes, is that there is a strong link between central reform ambitions and the meta-level of governance that is experienced in practice by frontline staff. Whether or not the reforms have the desired effects on welfare-to-work staff and the jobseekers they aim to help, it is certainly the case that such policy tools are capable of changing frontline work in practice.

Acknowledgments

This research was supported by an Australian Research Council Linkage Project Grant (LP150100277).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Jenny M Lewis

Jenny M Lewisis Professor of Public Policy at the University of Melbourne and is the President of the International Research Society for Public Management. She is a Fellow of the Academy of Social Sciences Australia, and an expert on policy making, policy design and public sector innovation. She is the author of six books, and more than 80 journal articles and book chapters.

Phuc Nguyen

Phuc Nguyen is a lecturer at La Trobe University, Australia. Phuc’s research interests include welfare state, especially the delivery of employment services; and supply chain sustainability. She has published three book chapters and several journal articles in Public Management Review, Journal of Social Policy and Administration, Australian Journal of Political Science, Third Sector Review, Australian Political Asia Pacific Management Review and Asia Pacific Journal of Marketing and Logistics.

Mark Considine

Mark Considine is Professor of Public Policy at the University of Melbourne and a Fellow of the Institute of Public Administration Australia and the Australian Academy of Social Sciences. His research areas include governance, comparative social policy, employment services, public sector reform, local development, and organizational sociology. He has produced six books and numerous journal articles and book chapters.