Institutional Logics and Low Skills: The Case of the Private Security Sector in Singapore

Introduction

Policy interventions in many economies in the area of skill formation have traditionally been informed by human capital theory (Becker, 1975; Little, 2003). As such, not only are skills policies ‘front end loaded’, focusing on education (i.e. supply-driven), but employers are deemed to be the major, if not the only, actor to shape and develop the skills agenda. In the event of market failure, active labour market policies may kick in to help individuals get back into jobs. In the late 20th and early 21st centuries, new policy measures were introduced with their academic origins in institutional economics and more specifically in the ‘Varieties of Capitalism’ literature (Hall and Soskice, 2001). Initially inspired by the work of Finegold and Soskice (1988) to tackle the problem of low skills, this ‘institutional’ approach has been further developed by others such as Buchanan et al. (2001) and Anderson and Warhurst (2012) to deal with skill formation issues across the labour market.

In spite of the advances made by these institutional approaches, which has helped shift policy attention to the demand for skills (i.e. at the workplace), the policy interventions inspired by these subsequent developments have not been very successful. This article argues that previous institutional approaches suffer from two major weaknesses. The first is that they treated the sector institutions as separate and independent from other elements within the sector. The mere presence of a certain set of institutions would generally be linked to a particular outcome (e.g. higher skills requirements in jobs). Not only can we not assume that a particular outcome is inevitable because of the presence of a certain institution, we tend to witness an absence of analysis of the relationships among these institutions. Secondly, while the presence of a set of institutions is deemed to be conducive to producing some favourable outcomes, the lens for these institutions almost always focuses on their complementarities. We argue in this article that a far more realistic examination of these institutions is less likely to reveal a system of well-intended cooperation among the institutions and more likely to reveal institutions striving to achieve their own objectives, only some of which may overlap with others, which may therefore lead to competitive behaviour, especially in the area of ‘power’. In this process, the more powerful institutions tend to shape the behaviour of other actors and the eventual outcomes.

By focusing on the relative power, legitimacy and interactions, we are able to identify why institutions in some countries work and others fail. Through the use of a case study of low-skilled and low-paid security workers in Singapore via the concept of institutional logics, we show how important it is to examine the relative power of the various actors within an industrial sector. The concept will also show how policy levers can be derived as a result of being able to identify what/which institutions shape behaviour within a sector. To summarise, the main research question of this article is to examine the usefulness as well as the weaknesses of the existing institutional approaches to analysing and formulating skills policy. This critique then leads directly to our proposed alternative ‘institutional logics’ (IL) model that places less emphasis on institutional complementarity, and instead identifies a more powerful source of policy formulation that draws on the analysis of how relative power and legitimacy affect the behaviours of institutions and skills policy outcomes.

Institutionalist approaches and their weaknesses

A number of academics have developed approaches from an institutionalist perspective that have provided insights into the conditions that favour a particular outcome between pay and skills. Hall and Soskice (2001) focused on the relationships among employers, state agencies, industrial relations and vocational training institutions, distinguishing between the liberal market economies (LMEs) where the state merely establishes the legal framework and relies on the market to coordinate activities, and coordinated market economies (CMEs) where the state plays a more active role, together with employers and unions, in coordinating the production process.

Gallie (2007) introduced a slightly different approach, focusing on the power of organised capital and labour and the way this is mediated by the state through employment policies, industrial relations and skill formation policies, creating inclusive, dualist or market regimes. Which of these regimes materialised depended largely on the involvement of organised labour in the decision-making process. More recently, Thelen (2014) has sought to extend this form of analysis by tracing how countries with different types of institutional characteristics have responded to the pressures of globalisation and the growth of the service sector by adapting their institutional characteristics (e.g. how the CMEs have maintained their inclusiveness with regard to the manufacturing sectors but at the price of greater exclusiveness in the service sector). Together, these early institutionalist approaches revealed important differences, especially between the production regimes (i.e. LMEs and CMEs) and the employment regimes (i.e. inclusive, dualist and market), in terms of the basic characteristics of the societal arrangement and the embedded economy, and how they affect pay and job skills.

While these approaches have made significant advances to our understanding of the impact generated by institutions, they nevertheless encounter problems. The first concerns the problematic nature of the concept of institutions. Salverda (2008) argued there was always a certain amount of fuzziness to the concept of institution. For example, it is used to refer to large organisations, small informal organisations, specific regulations, policies, informal rules, and public programmes. Used in this manner the concept starts to lose analytical precision. In addition, there is the question of whether the concept is used as an analytical tool to refer to isolated single institutions or groups of similar organisations. Salverda also cautioned against the risk of failing to see that such institutions exist as part of a larger structure.

In effect, early institutionalist approaches tend to fall back onto the use of typologies. Indeed, other scholars have also realised that these typologies are not adequate in encapsulating the varying practices across countries (e.g. Hall and Soskice, 2001). As Appelbaum et al. (2010: 26) point out:

… a clear understanding of specific national outcomes requires moving beyond typologies, which are often too functionalist and too static and tend to overemphasise complementarities. National institutional arrangements, in practice, often display many conflicts, contradictions, tensions and inconsistencies that can be brushed aside to fit into a particular typology.

In addition, most of the studies adopt a macro approach to the analysis of economic data and then draw conclusions about firms’ behaviour at the job level. This means that their analysis remains rooted to the national level, with the result that there is no attempt to study the ‘real interaction’ between firms and institutions at the local or sectoral level. Worse, there is no attempt to understand how the economic and sectoral environment may affect business strategy and how business strategy impacts on technology adoption, use of labour and work processes in the workplace. We argue that these omissions at the business level negate our effort to understand why jobs are low skilled and/or low paid. In so doing, previous approaches also fail to consider or explain the rationale and the actions of the respective stakeholders within the same business sector. Put simply, the failure to recognise the competing legitimacy of the stakeholders in their socio-economic position (e.g. within a business sector), and the rationale of their actions, has resulted in ineffective policy intervention.

Apart from Thelen (2014), one group who did attempt to study employers and institutions in detail at the sectoral level was the Russell Sage group through their analysis of low pay in six countries (Gautie and Schmitt, 2010; Lloyd et al., 2008; Westergaard-Nielsen, 2008). As this research was conducted by different researchers in each country, the conceptual framework they shared tended to be loose. Broadly informed by Finegold’s concept of the high and low-skills equilibrium (1988 and 1999), the studies focused on the high road and low road modes of organisation (Solow, 2008), on the understanding that employers have a choice of mode of production which are all economically viable. While accepting that competitive intensity and technology are important determinants of labour market outcomes, researchers in the Russell Sage group acknowledge that other institutional factors, such as labour market policies, minimum wage legislation, employment protection, welfare arrangements, collective bargaining, industrial relations and education and training systems, were also important. The impact of these factors on the types of jobs created was then examined in five sectors of low-wage and low-skilled employment.

While these studies made a significant contribution to our knowledge of how different institutional arrangements facilitated or inhibited the low-wage route, they too had limitations in linking the macro data analysis with sectoral and job level analysis. While the detailed empirical case studies identified the various institutional factors that create pressures to either drive down wages and conditions, as in the UK, or to sustain wage levels and improve job quality, as in Denmark, the connection between the two levels was never clearly spelt out. For example, in their analysis of the food and meat processing industry in the UK, James and Lloyd (2008) identify employers’ business strategies and the consequences of these for work organisation and pay levels. Trends in the product market were identified and institutional factors, such as the national minimum wage, health and safety legislation, the presence and use of agency workers and migrants, were all identified as limiting the actions of employers and explaining the downward pressure on wages and conditions. What is missing is an analysis on the interaction of the range of actors (e.g. purchasers, employers, regulatory agencies, labour market supply, trades unions) that influence outcomes, both intended and unintended, for employers and workers and the other actors involved. Rather than employers and other actors being passive reactors to external pressures, it is important to recognise that some may seek out new opportunities, and that all actors have their own interests and seek purposive outcomes, but the end result of the interaction is not necessarily the intended outcome of any one actor.

Another related problem, shared with the other institutional approaches, is that the intellectual framework does not provide any insights into which of the levers of change held by the various institutions and other stakeholders are the most significant. The study merely delivered a picture of the institutional arrangements reinforcing each other in such a way as to create an institutional arrangement that merely reproduces existing outcomes. All of this means that in these conventional approaches there is a level of interaction, crucial to an understanding of the behaviour of the respective parties, that is missing from the analysis.

The concept of institutional logics

The research objective of this article is to develop a new analytical approach which addresses the weaknesses of the existing institutionalist approaches. This new concept is called ‘institutional logics’. The approach is based on the relational sociology approach. This approach argues for the need for the analysis to focus on stakeholders and system interaction, social and economic structures (institutions), and the affordance or constraints experienced by the institutions, including societal norm and values (Bourdieu, 2005; Nee, 2005; Spence and Carter, 2014). By doing so, the analysis seeks to unveil the otherwise hidden competition for the actors’ legitimacy within the socio-economic field. This is extremely important because microeconomic analysis adopts the simplistic view of exchange among equals, which tends to be transactional in nature with assumed neutrality (Smelser and Swedberg, 2005). Weber has insisted that ‘[it] is essential to include the criterion of power control and disposal in the sociological concept of economic action’ (1922: 67). The study of high-skill and low-skill roads by institutionalists is essentially located within the economic field, which includes the decisions of firms as they interact with the exogenous forces that result in the varied outcomes of skills use and pay for the workers. The vision of multiple actors interacting and struggling within the economic field is aptly described by Bourdieu in the following passage:

The economic field of forces is also a field of struggles, a socially constructed field of action in which agents equipped with different resources confront each other in order to gain access to exchange and to preserve or transform the currently prevailing relation of force. Far from being faced with a weightless, constraint-free world in which they develop their strategies at leisure, they are oriented by the constraints and possibilities built into their position. (Bourdieu, 2005: 199)

In order to capture the interaction of multiple actors and present their interactions within the low-skilled, low-wage sector, we have adapted the ‘productive systems’ map originated by Robinson (1966) and developed by Wilkinson (1983). The ‘productive systems’ map attempts to capture the totality of the social relationships and interaction through the structure of production, and the process of production (Wilkinson, 1983). The map illustrates actors’ positions, actions and rationale. While this tool remains a relatively underdeveloped idea, it has informed some sector-based studies (Birecree et al., 1997; Konzelmann et al., 2006). The advantage of using this map is that it allows us to establish the locus of power within the interaction of the actors. Figure 1 shows the productive systems map of the private security services industry with which we will illustrate the interactions and power relations among the stakeholders leading to low skills and low wages.

OPEN IN VIEWER

Figure 1.

The productive systems map of the private security services industry.

On the y-axis, we identify the ‘hierarchy of power’ to illuminate the different power positions of actors in a system – the higher the position, the more powerful the actor. In Figure 1, these actors include the sectoral regulator, security service buyers, private security agencies and the security guards. The x-axis represents the process of production and consumption of security services. This entails the securing of site contracts by security agencies, recruitment and assignment of security guards, and the daily dealings of security guards and the facility management of the site. Hence, actors on the left tend to be involved early on in the production process. The security officers, who have no say in how their services are delivered, are on the extreme right of Figure 1. Some ‘stakeholders’ – notably the Union of Security Employees (USE) and the security agencies associations – are omitted in Figure 1 because they do not exert any power or systemic influence over other actors. This omission reflects a major analytical distinction between the proposed institutional logics approach and the previous institutional analyses, where the latter focuses mostly on the typology of (cooperative) institutions involved.

The arrows symbolise the actions of one actor imposed onto the others. The dotted line dialogue box explains the action. The actors located toward the top-left of Figure 1 are therefore more ‘powerful’ in shaping the action of those towards the bottom-right. Using Figure 1, we define institutional logics as a set of relationships between the outcomes of the production of the services or products and consequences of positional power within a definable sector or industry.

The (un-armed) private security services industry in Singapore provides a relevant case study for the concept of institutional logics. After decades of adopting one of the most pro-business environments, income inequality has been a major policy concern; as such, public skills strategy is geared towards supporting the twin objectives of increasing economic competitiveness and improving social inclusiveness. For many years, the private security industry has been on the radar of the government for its inability to get out of the low-skilled/low-wage equilibrium. Many attempts have been made to improve job skills and wages in this sector, but without success. For example, the Inclusive Growth Programme was introduced to encourage firms to adopt security technologies to increase productivity and pay higher wages; in the more recent Progressive Wage Model, the regulator mandates compulsory deployment of trained security officers for specific job roles. As will be explained, none of the initiatives have changed the continuous entrenchment of low-skilled and low-wage practice in the sector. The reason is that despite the repeated efforts, the underlying institutional logics in the sector have not changed. Security jobs remain low skilled and poorly paid.

Research design

Within our institutional logics framework, we began with a stakeholder mapping to facilitate the examination of the power relationships which are examined at two levels. The first are those among the macro institutions (i.e. those that affect employment practices, wage setting and the sectoral regulation). The second are at the sectoral level, where we seek to examine the institutional logics that condition practices, actors’ decisions and their actions to protect their respective interests.

At the macro-level, we conducted six interviews with policymakers between 2011 and 2013 when the Singapore government was calling for Skills-Productivity-Innovation driven growth. The interviewees were from those ministries, agencies and other organisations that were directly concerned with the private security sector. These include: the Ministry of Manpower (MOM), responsible for labour market regulations (e.g. working hours); the Police Licencing & Regulatory Department (PLRD), responsible for the licensing of security officers (SOs) and their employers, the security agencies (SAs); the Workforce Development Agency (WDA), responsible for training SOs; SPRING, the agency responsible for innovations in small and medium-sized enterprises (SMEs); the National Trades Unions Congress (NTUC) that was implementing the Progressive Wage Model; and the Union of Security Employees. These interviews were supplemented by the analysis of policy documents and secondary data on wages, productivity and skills. Many of these interviews were guided by prior examination of the available policy documents.

At the sector level, the buyers, who purchase the services of SOs from the SA, are the owners and managers of physical properties (e.g. commercial centres, retail malls, residential sites and sensitive locations such as government buildings). We interviewed eight of these buyers who owned/managed different sizes and types of property in different locations. Of the 273 SAs in 2013, 70 per cent were small and medium firms employing fewer than 100 SOs (Singapore Police Force, PLRD, 2013). The small SAs usually consist of the owner, a few administrative staff and a larger group of SOs on short-term contracts. We interviewed 10 of these SAs, which ranged from the largest security firms undertaking national contracts for the government to very small SAs employing less than 20 SOs. We targeted our questions at factors that drive the actors’ behaviour (e.g. how competition and regulation affect their business approaches and SO job content) and their own analysis of what could have been done differently.

Seventeen interviews were conducted with SOs, in which we explored the SOs’ day-to-day experience of work, skills issues, their motivations for entering the job and continuing within it, their perceptions of SAs, and the opportunities or lack of them for entering other jobs. Besides the interview, we also conducted three focus groups with SOs. These were chosen randomly from the training centres at which mandatory certification is required. The focus groups enabled the researchers to explore issues of motivation, the status of the job, career and why they were attracted to security employment. We also interviewed the Union of Security Employees and the two SA associations regarding their role in relation to the construction of SO jobs and other matters such as contracts and wage issues. These are the formal representation bodies of SOs and SAs.

All individual interviews took the form of guided conversations, which enabled the interviewers to probe different areas relevant to identification of institutional logics in the private security industry. The transcripts were manually coded based on the broad criteria of: social position, disposition, rationale, actions, relational interactions with other actors, affordances and constraints encountered. The transcripts were individually analysed, and then cross-analysed by group and across groups, to identify consistencies and inconsistencies, as well as their relationships with one another. Together with the documentary information, the interviews provided rich data to reconstruct a full picture of how security jobs were created, designed, remunerated and how the specific institutional logics for the security industry were constructed.

The institutional logics of the private security industry

This analysis starts by tracing the relationships depicted in Figure 1. As illustrated, two of the most powerful institutions that shape the business strategies and behaviour of employers (SAs) are the PLRD and the service buyers. The security industry in Singapore is viewed as a strategic sector by the ministry and agencies looking after the national security of Singapore. The police authorities treat the security agencies and the security guards as their junior partner in protecting Singapore from the threat of terrorism. The PLRD shared its view on the primary function of the private security services industry:

They form the network of eyes and ears on the ground that not only perform ‘detect and report’ functions, but also act as critical deterrents in ensuring the national security of Singapore.

This perceived importance of national security means that the PLRD uses its regulatory levers to manage the SAs and to license individual SOs. The SAs are generally perceived to have poor management practices because of the general low-cost approach, poor working conditions and minimal training investment in SOs. In an attempt to change this and inculcate good practices, the PLRD provide an annual grading exercise of SAs to ensure ‘the industry maintains a high standard of security services’. It involves allocating SAs to one of four grades, A to D, based on their operation, training and backroom support and management (including human resources) practices, which together are known as the ‘three pillars’. The annual grading is the PLRD’s tool for professionalising the sector:

The grading aims to provide an assessment of the performance of the SAs, help service buyers to discern service quality, motivate Security Agencies to improve operation capability, and to strategically elevate standards of private security industry in Singapore. (Singapore Police Force, PLRD, 2011: 1)

However, interviews with SAs show that grading is of little relevance to the contractual outcomes in terms of skill requirements. As one SA explained:

I think it is because of [the buyers’] limited budgets. Most of them don’t specify grade [of SAs]. Mostly bigger companies or more special locations would do so. [The] majority would pick the best price that comes with the best grade. The grading didn’t help us in securing [a] contract.

The security (only) focus means that the PLRD dictates the mandatory skills for SO licensing. The competencies are subsequently endorsed at the Sectoral Tripartite Committee at which SAs are not represented.

For the entry level SOs, there are only two mandatory competences – ‘handle security incidents and services’ and ‘provide guard and patrol services’ – skills that can be acquired in a short period of time and which do not provide room for variety and discretion. Indeed, SOs revealed little variation in the task structure of these jobs because the job content was driven by the mandatory (security-only) competences and the specifications of the contract between the SA and the buyer of services.

Beyond dealing with the PLRD, SAs felt greatly constrained by the service buyers as they had to accept non-negotiable contract terms, the risk of premature termination of the contract, and the monthly deduction of fees under the ‘liquidated damage’ (LD) terms in the contract. One manager explained:

Every new client is a risk. I had a really bad experience with this client where he imposed unreasonable liquidated damages in the first month of operation. Ended up I have to pay them an amount more than the contract sum. We can’t work like this, so I said, ‘sorry, let’s part ways’.

The tender specifications dictate the number of security officers and supervisors needed per shift. We just quote based on the specification. Sometimes [the buyers] may not require a supervisor on site. Sometimes, they may ask us to quote for walkie-talkies and clocking equipment.

If only the customer would listen to us, we are not here to sell them many guards. We are here to tell them the bare minimum they can get away with, but they said ‘just follow the spec’; so we just quote accordingly.

From the SA interviews, we learnt that it was a standard practice for buyers to regularly impose LDs. Most of the deductions were caused by the poor performance of the SOs. One security agency explained how this creates a high level of mistrust in the relationship between the buyers and SAs:

In principle, we always tell our customers: we have no problems with LDs. Because I do understand that you need to use that as an instrument to manage your vendors, but it is how you use it.

LDs feature extensively in all contracts, in which most of the penalties are due to poor performance of the SOs deployed to the site. This may be attributed to a number of root causes. Firstly, the SOs are normally recruited and deployed for the period of a site-specific contract. Thus, the employment relationship is transient in nature. Secondly, the SAs have little contact with the SOs once deployed to site. Thirdly, the manpower shortage perpetuates poaching of SOs. This leads SAs to use double shifts to make up for the missing headcount on site. To overcome LDs, most SAs have adopted technology to enforce surveillance on SOs (e.g. using bio-metric attendance systems, real-time clocking, remote monitoring cameras on sites). The following remarks present the various views from the SAs:

[Security guards] are mercenaries. They would leave for the extra dollars. It is very frustrating! It does not help when PLRD allows the licensed guards to work for two employers; it should be one! Some work for one agency in the day shift, change uniform, and work for another agency in the night. Next day, too tired; did not report for duty.

At the moment, the guards are getting away with murder. PLRD should hold them [SOs] individually accountable. Agencies like us are toothless. I told the guard: You sleep on duty, I sack you! Sack lor [i.e. reply: go ahead and sack me; not bothered]! He goes to the next employer. Then, we are penalised by our customers!

Essentially, this Virtual Supervision System (VSS) was to help us. Firstly, [it is] to encourage punctuality, good attendance and discourage people from leaving the site without proper authority. And at the same time to make sure that whatever we promised the clients in terms of security patrols are carried out diligently. Lastly, it enables us to observe what’s going on at the site without us having to be there. Hence, my operations managers can control multiple sites [in] real-time, via iPhones. So that’s very productive.

The other actor responsible for shaping the behaviour of the SAs and SOs is the Ministry of Manpower which, as Figure 1 illustrates, occupies a very powerful position in shaping relationships in the sector. As the private security sector is regarded as part of the wider national security system, foreign labour is not allowed to work as SOs. This means that SAs have been confronted with a manpower shortage due to a tight labour market while the low-skilled, low-waged work makes it difficult to attract labour. As a result of the shortage of labour, the Ministry of Manpower has granted exemption to the working-time regulations for the industry. The exemption has resulted in the prevalent practice of 12-hour work shifts, which means the security guards are required to work an average of 72 hours work per week. The unintended consequence is a suppressed wage. A representative from the USE explained:

The overtime exemption is historical. The pay is low, so overtime gives workers the extra bucks. With the labour shortage, we cannot abruptly stop the exemption. Otherwise, the employers would suffer. The drastic move may force many SMEs [small SAs] to go out of business. What happens next, employees would be out of jobs. I think it is also a lack of political will from the government. … Actually, with one stroke, the government can revert many things, the overtime and the wages. It is a lack of political will!

The objective of helping the industry to overcome manpower shortages has inadvertently resulted in excessive overtime and helped sustain poor working conditions and poor work performance in the form of high levels of absenteeism. High levels of absenteeism lead to LDs being imposed by the buyers. This creates pressure on the SAs to minimise operating costs, and as they compete against each other for contracts, this results in even greater downward pressure on wage costs. Security jobs are not just poorly paid but are also insecure. Once a contract is secured, the SA is free to fill the jobs with any SOs, so long as the required number of SOs is met. At the same time, SOs are free to contract their service to other SAs, so long as they can deliver those hours. As a result, in a tight labour market, a casual and non-committed employment relationship between SAs and SOs has developed.

Our interviews also revealed that security jobs are often described as dull, low status and boring by the SOs. From their perspective, there is little to tie them to their immediate employer. After leaving one job, an SO can easily pick up another. Those working overtime or with a second security job do not have any incentive or time to think about up-skilling. For some, once the contract ends, the employment relationship may dissolve for various reasons, such as the absence of a new contract or the new contract is located too far from home. Under these circumstances, it also means that employers (SAs) cannot take a long-term view regarding capability building and investing in the skills of employees.

The use of institutional logics would suggest that employers – being so far down in the institutional logics chain – have little power and room to shape their business strategy, the service design, alternative ways of creating a competitive advantage, skills options and the pay and conditions of their employees. They operate in a regime where the other more powerful actors, in the exercising of their legitimacy, have created unintended consequences that perpetuate the conditions conducive for reproducing the low-skill road.

Yet there are pressures for change. The first concerns the power of the PLRD in shaping business strategies and job design. The current PLRD regulations focus on the primacy of the ‘security only’ aspect of the job of SOs. However, our research reveals that in many of the ‘non-sensitive’ properties, such as shopping malls and residential sites, SOs are often approached by customers asking for general assistance, advice or reporting incidents. Such activities would normally be regarded as part of estate management. SAs report that they would like to change their business strategies and branch out into the various disciplines of estate management so that they could add value to SOs’ jobs. If the PLRD regulations were to modify the ‘security only’ focus, taking into account the levels and types of security required by different properties, the SAs would modify their business strategies to take advantage of the new opportunities. As such, there would be a demand for new skills for the more broadly defined jobs. It would also change the dynamics of competition among SAs in this part of the market, creating more emphasis on the quality of the service they provide, moving away from the ‘body-count’ default approach so prevalent in the industry. The second pressure for change concerns the role of the MOM. As long as the MOM provides the SOs with exemption from the working-hour exemptions, they will continue to use excessive overtime and second jobs to maintain their income. The enforcement of regular working hours and minimum wage could change the dynamics of the labour market, placing pressure on employers (and buyers) to make more effective use of labour when labour is in short supply. ¹

Under institutional logics, the above changes enable us to identify the pressure points for policy change and to establish a different kind of business environment that provides more room for job redesign, career building, changing the technology–person mix in the job content and the scope for buyers to value-add. As one of the major buyers of security services, the government is in a position to introduce such changes through their procurement activities. Thus, while the institutional logics continue to entrench the private security sector in the low-skills, low-pay road, the analysis also reveals areas where change can take place – the logics are not deterministic. Yet the points at which change can be effectively introduced are within the context of exiting power relations within the sector. They are not in the areas where voluntaristic government programmes, similar to those used in the past by governments in the UK, Australia and Scotland, can be used to persuade employers to adopt ‘better’ HR practices (Anderson and Warhurst, 2012; Buchanan et al., 2014). Change in the institutional logics in Singapore requires action directed at changing power relationships between the actors identified above, rather than seeking to persuade employers to collaborate in changing their ‘HR’ practices.

Conclusions

In criticising the institutionalist approaches, we highlighted their failure to make explicit links between macro factors such as legislation, government ministries and the sectoral level participants such as the employers and the buyers of security services, and how the link may consequentially affect jobs at the worker’s level. We have sought to rectify this by using the concept of institutional logics to focus on the interrelationships between the actors involved. This has revealed the importance of the power relations in shaping the dynamics of the relations between the various interdependent actors (though with different powers), which in turn shape the business strategies of firms and relationships within the labour market. This has revealed a situation in which firms that are usually the focus of policy interventions are relatively powerless to influence the reproduction of the low-skills road in Singapore, at least in the security sector. This is in contrast to the tendency in the literature to view firms as the culprit in the low-skilled sector for their failure to embrace the high-skills model and their inability to develop high-performance working (e.g. Bernhardt, 1999; Ehrenreich, 2001). Instead, our approach highlights relationships at the meso- and sector levels as crucial to understanding the persistence of the low-skills route.

The concept of institutional logics has also revealed the centrality of power relations in understanding labour market outcomes. Previous theoretical approaches treated the sector institutions as separate and independent and tended to focus on their complementarity in producing outcomes, and in policy terms emphasised the importance of collaboration between them. In contrast, our approach reveals the importance of power relations in shaping the dynamics of the relations between the various interdependent actors, simultaneously co-existing in forms of competition and collaboration. By focusing on the institutional logics, we have been able to identify the relative power balance within the sector and thereby identify the institutions or actors that have the greatest impact on shaping the dispositions and business strategies of the employers and other actors. All this has important policy implications.

One added strength of the institutional logics approach is that it focuses more on the business model in a normal business environment, rather than an artificial business environment created by government programmes seeking to generate collaborative behaviour induced by incentives. Thus, previous policy interventions that were based on the skills-ecosystem approach have proved largely unsustainable because they are not changing the institutional logics that underpin the business models for high skills. These were interventions that were targeted at employers and their partner institutions with the intention of demonstrating the advantages of employing more highly skilled workers. However, once the incentives such programmes provided were taken away with the end of funding, the partnerships withered (Buchanan et al., 2014). As the institutional logics approach focuses on restructuring the business environment via changing the power relationships and the behaviour of employers and other actors, changes are more likely to be long-lasting. It provides the basis for interventions that take cognisance of the realities of the business environment. What this also means is that because the institutional logics tend to vary across sector, then forms of policy interventions, the policy levers, are likely to be sector-specific.

Finally, it is very important to recognise that we are not necessarily arguing for the use of a new policy tool or form of regulation. The use of the concept of institutional logics illustrates that the market mechanism is a versatile tool that can work within different institutional environments yet lead to very different outcomes. It provides a means of changing the cost/benefit calculations of employers in relation to skills, thereby creating a business case to make skills a driver for change within the business environment.