A promise made is a promise kept: Union voice, HRM practices, implicit contracts and workplace performance in times of crisis

Funding information Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung, Grant/ Award Number: PBZHP1_143612 Abstract Sociological debates on financialisation emphasise that workplace-based productivity bargains between employers and employees are often precarious. This literature suggests that where employers are unable to keep their side of the bargain, participative arrangements with employees are likely to become unsustainable in generating high workplace performance. Recessions substantially increase the pressures on employers to breach (unwritten and invisible) implicit bargains with employees. We propose that a robust (rather than shallow) implementation of union voice and Human Resource Management (HRM) practices will make implicit bargains more resilient and benefit workplace performance if recession impact is high. Using two waves of the British Workplace Employment Relations Survey (2004 and 2011), we examine our hypotheses in the context of the deep 2008–2009 recession and find supportive evidence – especially in (a) firms without a dominant individual or family owner and (b) growing markets.


Practitioner Notes
What is currently known about the subject matter • Workplace-based productivity bargains between employers and employees are often precarious.
• Workplace performance crucially depends on (unwritten and invisible) implicit contracts between employers and employees.
• External economic pressures, such as a recession, can induce managers to breach implicit contracts with employees.
What our paper adds to this • Union voice and HRM practices are mechanisms that can support these (unwritten and invisible) implicit contracts.
• Therefore, especially when recession impact is high, strong union voice and a full implementation of HRM practices can benefit workplace performance.
• This key finding depends on the company's context.
• Specifically, the finding is corroborated for firms without a dominant individual or family owner, and in growing markets.

Implications of the study findings for practitioners
• A shallow implementation of union voice and HRM practices may not sufficiently support implicit contracts when there is an 'external shock' such as a recession.
• Strong union voice and a full implementation of HRM practices are likely to benefit workplace performance especially after such a 'shock'.
• Focussing on the short-term costs of these practices can therefore be myopic.
• Managers need to consider their company's context (i.e., ownership structure and market conditions) when evaluating these practices.

| INTRODUCTION
Research frequently suggests that participative arrangements with employees can be a source of mutual gains for employers and employees and can lead to high workplace performance (e.g., Kochan, 1995). Such arrangements often rely on (unwritten and invisible) implicit bargains that, in return for employees' increased involvement and participation, management will aim to safeguard employees' interests (Appelbaum, Batt, & Clark, 2013;Thompson, 2013). As implicit bargains cannot be enforced in courts, employees need to be able to rely on management's commitment to protect them (Mayer, 2013). Crucially, it is often difficult for management to safeguard employees' interests when their firms are under short-term financial pressures. To the extent that implicit bargains are protected, such bargains facilitate efficient collaboration and high workplace performance (Blair, 1995).
The sociological literature on financialisation, though, is sceptical of the sustainability of these bargains (e.g., Clark, 2009;Cushen & Thompson, 2016;Dore, 2008;Perraudin, Petit, & Reberioux, 2013). Thompson (2003) suggests that management's commitment to implicit bargains is often shallow: there is often a discrepancy between the stability needed to support these bargains and the insecurity in today's economies. The notion of 'financialized economies' captures this insecurity and includes various trends, such as the deregulation and globalisation of markets and the important role of finance in management's decision making (Thompson, 2013). If employers are unable to keep their side of the bargain, participative arrangements are likely to become unsustainable in generating high workplace performance (Clark, 2009).
Thus, an important question arises: How sustainable are these bargains in the light of external shocks, such as a recession? Existing research outlines two opposing views (Butler, Glover, & Tregaskis, 2011) based on national business systems: (a) Streeck (1997) suggests that workplace partnerships are unlikely to be robust in a voluntarist environment such as the UK (with a lack of 'beneficial constraints'that is, little hard regulation protecting employees' participation in decision making). Managers will be tempted to defect from workplace partnerships in the short term, even though cooperation would be beneficial in the long term; and (b) Roche and Geary (2002) who maintain that even in a voluntarist environment, robust partnerships may arise as long as there are 'functional equivalents' to the beneficial constraints that are legally imposed in other countries such as Germany. The question whetherand under which conditionssuch resilient partnerships are possible is a key motivation for our study.
In this article, we propose that a robust (rather than shallow) implementation of union voice and HRM practices will make implicit bargains more resilient and benefit workplace performance if recession impact is high. The distinction between robust and shallow implementations of labour-management practices/processes has been used elsewhere (e.g., Dobbins & Dundon, 2017). As indicators for a robust implementation, we will consider the strength of union voice in decision making (i.e., the extent to which unions can negotiate, rather than being consulted or informed, on a range of relevant issues; e.g., Oxenbridge & Brown, 2004) and the accumulation of HRM practices (i.e., the extent to which workplaces have adopted a full array of practices; e.g., White & Bryson, 2013). Research on employee involvement and participation (EIP) distinguishes between formal representative, direct formal and informal forms of EIP (Marchington, 2015;Wilkinson, Gollan, Marchington, & Lewin, 2010). The first key variable in our analysisunion voiceis a formal representative EIP form, whereas the second key variable -HRM practicesincludes direct formal EIP as a key component. The literature on HRM practices typically focuses on bundles of practices that comprise not only direct EIP but also complementary practices including work organization, incentives, training and recruitment (Mowbray, Wilkinson, & Tse, 2015;White & Bryson, 2013). We will use 'participative arrangements' as an umbrella term covering union voice and HRM practices.
The theoretical argument developed here is that a severe recession represents an 'external shock' that substantially increases the pressures on employers to breach implicit bargains with employees. If the participative arrangements (i.e., union voice and HRM practices) are sufficiently robust, however, we may expect implicit bargains to survive and the workplace to achieve good performance after the deep recession. Thus, we propose that union voice and HRM practices can serve as functional equivalents to legally mandated 'beneficial constraints' if they are deeply embedded in the workplace. This is key to understanding the long-term future of employee participation, because recessions may disrupt productive employment relationships.
To address this question, we examine the interaction effects between recession impact and union voice and HRM practices, respectively, on workplace performance. This empirical approach will help to make inferences about the resilience of the (unwritten and invisible) implicit contracts. We use two waves of the British Workplace Employment Relations Survey (2004 and 2011). The first wave was during one of the longest periods of economic growth in Britain, whereas the second wave took place after the global financial crisis and the deep 2008-2009 recession. Crucially, the recession affected different workplaces to different degrees (an item captured by the survey in our empirical analysis). We find evidence supporting our hypothesis that a robust implementation of union voice and HRM practices is beneficial in terms of workplace performance under high recession impact. Subsample analyses show that this association holds particularly for (a) firms without a dominant individual or family owner and (b) in growing markets. To the best of our knowledge, this is the first study that examines these associations using large-scale regression analysis. We conclude with suggestions for future research.

| EMPLOYEE PARTICIPATION AND IMPLICIT EMPLOYMENT CONTRACTS
It is widely acknowledged that employment contracts are to a large degree incomplete (Goldthorpe, 2007). Although employment contracts typically outline employees' roles and responsibilities, many aspects of work (including, e.g., cooperation, adaptation and taking initiative) are difficult to specify in formal contractual terms. Employers and employees are often unable to write down the terms of their cooperation in great detail due to their inability to foresee all future contingencies (Williamson, 1985). Therefore, it can be too costly or impossible to make a contract that is enforceable in court with sufficient reliability (Posner, 2000). As a consequence, cooperative employment relations depend to a large extent on implicit contracts (e.g., Granovetter, 1985).

| Employee participation and the need to protect implicit contracts
Implicit contractsalso known as 'relational contracts' 1are unwritten agreements and understandings that cannot be enforced in courts (Macneil, 2000). The concept of implicit contracts has roots in economics and sociology, though with slightly different nuances. The economic view emphasises the 'shadow of the future', suggesting that current violations of implicit contracts may (under certain conditions) be punished in future interactions (e.g., Baker, Gibbons, & Murphy, 2002). The sociological view, though, emphasises the 'shadow of the past', implying that the past development of social norms may discourage the current violation of implicit contracts (e.g., Granovetter, 1985).
In both views, there is a concern that managers might act myopically (if they are under short-term pressures) and breach implicit contracts.
Employee participation in organizations often relies on building trust (Timming, 2012), but breaches of implicit contracts can destroy trust. Such breaches can take different shapes and forms (e.g., reducing job security, intensifying work or cutting pay and benefits), weaken norms of cooperation and reduce the employees' willingness to engage in efficiency-enhancing improvements. A breach of implicit contracts can generate immediate gains today (e.g., by cutting employees' wages) at the expense of distant gains in the future (e.g., by reducing employees' future willingness to cooperate).

| Recessions increase pressures to breach implicit contracts
Economic recessions represent an 'external shock' that increases the short-term pressures on managers, potentially inducing them to breach implicit contracts. Teague and Roche (2014) provide evidence that managers sometimes react to recessions by making myopic decisions, such as overly intensifying work and creating employment insecurity see also Gallie, Felstead, Green, and Inanc (2014) and Felstead, Gallie, Green, and Inanc (2013). Roche and Teague (2014) examine patterns of adaptation in response to the short-term recessionary pressures and find evidence for myopic adaptation in some workplaces, in reaction to 'the tyranny of the short-term', but also for more long-term oriented adaptations in other workplaces, trying to maintain staff morale during the crisis. Relatedly, Roche, Teague, and Coughlan (2015) and  identify different types of concession bargaining with different degrees of cooperation (or confrontation) with unions.
Overall, these studies support the assertation in the financialisation literature that implicit contracts are precarious and vulnerable to disruption (Thompson, 2003). They also show that employees and their representatives tend to be more willing to make sacrifices during a deep recession. Thus, not every concession or sacrifice needs to be considered as a breach of implicit contract. However, there are clear differences between employers in the way that they seek such concessions. In contrast to a growing economy, a recession requires employers and employees to share the pain rather than the gain. Thus, economic recessions put employers' commitment to implicit contracts to the test.

| The robustness of union voice and HRM practices
Union voice and HRM practices can support implicit contracts in workplaces. As discussed above, managers can implement these arrangements in a shallow or robust manner. How does robustness fit in the above discussion of precarious implicit contracts in a recession?
We propose that if recession impact is high, a more robust implementation of these participative arrangements will provide a better protection of implicit contractsthat is, stronger union voice and a fuller implementation of HRM practices will be beneficial to workplace performance after recession impact. Thus, our main interest is the interaction effect between these participative arrangements and recession impact on workplace performance.
Regarding union voice, a deep recession poses particular challenges for the unions' 'collective voice' face (Doucouliagos, Freeman, Laroche, & Stanley, 2018;Freeman & Medoff, 1984). Employees' suggestions for productivity improvements are particularly valuable when the workplace's profitability is under pressure. This economic pressure also exacerbates the risk that management might use productivity improvements to breach implicit contracts.
Thus, weak union voice offers little insurance for employees that managers will not use their suggestions against them. Strong union voice can serve as an institutional constraint against excessive short-termism that might damage implicit contracts (Butler et al., 2011;Deakin, Hobbs, Konzelmann, & Wilkinson, 2006). Furthermore, a union's function of improving information flows between management and employees (Kaufman, 2004) becomes more important in times of turbulence and crisis.
Regarding HRM practices, a deep recession increases the value of HRM practices in making the workplace more flexible. External economic pressure is likely to require the workplace to reconfigure its resources and adjust to the changing environment (Wright & Snell, 1998). The use of direct voice practices (which are a key part of HRM practices) can increase the speed of decision making and the extent to which managers involve employees under time pressure (Bryson, Charlwood, & Forth, 2006). During times of crisis, HRM practices may make managers more responsive to employees' needs (Boxall & Winterton, 2018) and may engender a paternalistic concern for employees' welfare (Cullinane & Dundon, 2014). We expect that managers' temptation to breach implicit contracts (e.g., by reducing job security, intensifying work, cutting pay and benefits, etc.) is lower if the workplace has a fuller implementation of HRM practices, because the adoption of HRM practices is costly: if a workplace has made high investments in HRM practices and employees do not cooperate, the workplace will forego the returns on its investment (Ichniowski & Shaw, 2003). Moreover, a fuller implementation of HRM practices is likely to ingrain stronger norms of cooperation in the workplace's culture, thus acting as a constraint on managers (Miller, 1992).
Thus, we propose: Hypothesis 1 There will be a positive interaction effect between the strength of union voice and recession impact on workplace performance.
Hypothesis 2 There will be a positive interaction effect between the extent of HRM implementation and recession impact on workplace performance.
The above hypotheses are part of a contingency approach, suggesting that the effectiveness of (strong) union voice and (a full implementation of) HRM practices may depend on the workplace's context. This contingency approach is inspired by past research. For example, Freeman and Medoff (1984) assert that a key issue for understanding union voice concerns the relative importance of two facesthat is, the 'monopoly face' (which may reduce workplace performance by monopolising the supply of labour) and the 'collective voice' face mentioned above (which may increase workplace performance by improving collaboration, employee morale, etc.). Our first two hypotheses suggest that recession impact is an important contingency. In the subsequent section, we aim to refine this approach by considering additional contingencies, which shall be tested in subsample analyses.

| Contextual conditions: Ownership structure and market growth
The above hypotheses may not apply to all workplaces equally. Specifically, we consider two contextual conditions: (a) firms with/without a dominant individual or family owner and (b) growing/non-growing markets.
Regarding ownership structure, a dominant individual or family owner is likely to provide more stability in times of crisis. Such owners often have a long-term horizon for their investment, driven by their intention to protect their There is empirical evidence that family firms' long-term orientation contributes to stability: for example, family firms are overrepresented in cyclical industries with wide fluctuations in prices, such as construction and personal and household goods (Zellweger, 2007).
A second stabilising factor is that dominant individual or family owners constitute a countervailing power to employee voice. Coupled with long-term involvement, countervailing powers can lead to more balanced negotiations (Galbraith, 1952) and can therefore facilitate more productive implicit contracts between employers and employees (van Essen, van Oosterhout, & Heugens, 2013). Furthermore, there is evidence that 'relational' owners (i.e., owners with long-term involvement and interests beyond financial returns) tend to safeguard implicit contracts with employees during macro-economic crises (David, O'Brien, Yoshikawa, & Delios, 2010;Krausert, 2018;Yoshikawa, Phan, & David, 2005). Due to their long-term involvement, such owners are likely to be more concerned about the negative long-term consequences of breaching implicit contracts.
In sum, we expect that dominant individual or family owners will act as a stabilising force and that the effectiveness of participative arrangements will be less sensitive to recession impact as a result. Thus: Hypothesis 3 If the firm has a dominant individual or family owner, the effects of union voice and HRM on workplace performance will be less sensitive to recession impact. That is, we are less likely to find a significant interaction effect between union voice (or HRM practices) and recession impact in the subsample of workplaces with a dominant individual or family owner.
Regarding market conditions, growing markets are typically characterised by 'unprogrammed decision making, competitive variation, and poorly understood means-ends linkages' (Hambrick & Abrahamson, 1995, p. 1429. That is, managers have a wider set of investment opportunities (than in non-growing markets) and need to be creative and willing to exploit these (potentially risky) opportunities. Managers in such markets tend to receive higher salaries and more variable incentive compensation to reflect the greater responsibility and risks they take (Collins, Blackwell, & Sinkey, 1995) as well as their greater influence on the company's success (Finkelstein & Boyd, 1998).
In growing markets, breaches of implicit contracts are likely to be especially damaging. If managers have a wide range of strategic options and can use creative courses of action, they rely on the employees' flexibility and adaptability to implement their strategies successfully. Yet employees may resentand refuse to supportsuch flexible courses of action after breaches of implicit contracts. Weak support for implicit contracts is likely to harm workplace performance especially where managers require employees to engage in new and risky activities to exploit market opportunities.
We propose, therefore, that a robust implementation of union voice and HRM practices is particularly important in growing markets in order to protect implicit contracts if recession impact is high. Workplaces that manage to uphold implicit contracts during the crisis will likely have a strong advantage of adaptability. In mature or declining markets, this advantage is likely to be less pronounced, given the lower requirement of quick adaptation and the fact that conflicts over the distribution of profits are paramount even in 'normal' times (Barry & Wilkinson, 2011). Thus: Hypothesis 4 If the workplace operates in a growing market, the effects of union voice and HRM on workplace performance will be more sensitive to recession impact. That is, we are more likely to find a significant interaction effect between union voice (or HRM practices) and recession impact in the subsample of workplaces that operate in growing markets.  Table S1 presents a detailed definition of the dependent, independent and control variables. Workplace performance is an unweighted additive index of the manager's ratings of three performance dimensions (i.e., financial performance, labour productivity and quality of service or product), in comparison with competitors in the same industry (see, e.g., Bryson, Forth, & Stokes, 2017 for a recent application). 3 Although WERS also collected objective financial data, this data suffers from substantial sample attrition biases, due to the sensitivity and confidentiality of the establishment-level data. The two types of performance measures often produce similar results when relating them to a set of exogenous variables (Forth & McNabb, 2008). Thus, both types of measures have advantages and disadvantages (see Bryson et al., 2017;Forth & McNabb, 2008 for a comprehensive discussion). For the purposes of the present study, the subjective measure is beneficial both for comparability with prior studies on related topics and for our theoretical objective: that is, if managers consider workplace performance to be high in the presence of strong employee participation, we may infer that they are likely to be supportive of, or not opposed to, the underlying implicit bargains.

| Independent variables, control variables and subsample criteria
Consistent with our theoretical argument, our independent variables represent intensity measures for union voice, HRM and recession impact. Union voice is an additive index based on whether management normally 'negotiates with', 'consults with', 'informs' or 'does not inform' unions about six workplace-related issues (i.e., rates of pay, hours of work, holiday entitlements, pension entitlements, training of employees and health and safety). The Cronbach's alpha of the union voice index for the whole sample is 0.966, suggesting very high internal consistency. Our HRM index follows White and Bryson's (2013) definition, which used WERS 2004 and constructed two alternative HRM indices (with 43 and 20 items, respectively). We used their extended index covering five domains: participation (e.g., meeting time and consultation), team-working (e.g., team interaction and responsibility), employee development (e.g., training and induction), selection (e.g., skills and personality tests) and incentives (e.g., merit pay and profit sharing). We checked the comparability of the items in WERS 2011 and were able to retain 38 items that were coherent across the two waves. 4 Recession impact is a five-value measure of the degree to which the workplace has been adversely affected by the 2008-2009 recession. The year 2004 was part of one of the longest periods of economic expansion in Britain (i.e., we coded the variable as 'no recent recession'). An advantage of using WERS 2004 and 2011 is that our sample comprises a range of workplaces under favourable as well as unfavourable macro-economic conditions.
To facilitate comparability, our choice of control variables follows that of related studies (Bryson et al., 2017;Bryson, Willman, Gomez, & Kretschmer, 2013). We control for non-union voice, workplace size, establishment age, largest occupational group, region, industry, multi-site organization, foreign ownership and year 2011. 5 Finally, we use two criteria for the subsample analysis. Dominant individual or family owner indicates whether such an owner owns at least 50% of the organization to which the workplace belongs. Market growing indicates whether the market for the workplace's main product or service is 'growing' (as opposed to 'mature', 'declining' or 'turbulent').  1992up to 2007. In 2008-2009, the global financial crisis plunged Britain's economy into its deepest post-war recession (Seager, Kollewe, & Hopkins, 2009). Then the economy recovered but with lower growth rates than in the pre-crisis period (World Bank, 2018). Fewer workplaces were in growing markets in 2011.

| Analysis
We apply survey-weighted, linear, pooled panel regression, using a Huber-White clustered sandwich estimator to allow for intragroup correlation (clustered by establishments) with robust standard errors. Due to WERS's sample design, we apply Stata's 'svy' suite to ensure a correct treatment of the survey weights. Following Bryson et al. (2017), we treat workplace performance as a continuous variable, because the additive performance index approximates a continuous scale.
Consistent with prior literature (Bryson et al., 2006;Willman, Bryson, & Gomez, 2007), Table S2 shows that our main predictor variables (i.e., union voice and HRM practices) are relatively stable over time. Therefore, we applied our panel regressions to the absolute levels rather than changes in the independent and dependent variables.
To calculate interaction effects, we mean-centred the respective variables: union voice index, HRM index and recession impact (Aiken & West, 1991). Table 1 shows a summary of the estimated effects for union voice and HRM practices, respectively (see Table S3, for the coefficients and t-statistics). For each of the subsamples, we can see whether union voice and HRM practices (a) significantly interact with recession impact; (b) have a significant direct effect or (c) have no statistically significant effect on workplace performance. If the interaction effect is significant, a short description explains the predicted effects.

| RESULTS
1 and 2 suggest that union voice and HRM practices will positively interact with recession impact on workplace performance. Table 1 confirms these hypotheses for the whole sample. Figure 1 illustrates the same: as recession impact increases (from left to right), the benefits of union voice and HRM practices increase too: that is, the lines for the union index and the HRM index turn anti-clockwise.
The main effects in Figures 1-3 are consistent with the prior literature. HRM practices are sometimes described as 'best practice' (but not unanimously; see Freeman & Kleiner, 2000). The effects of union voice on performance tend to vary across countries, time periods and samples of workplaces (Doucouliagos & Laroche, 2003). Using samples of British workplaces, Bryson et al. (2013) find overall no statistically significant effects on workplace performance. In our analysis of the whole sample and the subsamples, the direct effect of union voice on performance is never negativeit is either not statistically significant (at the 5% level) or significantly positive (in the subsample where a dominant owner is present). This finding is also consistent with Freeman and Medoff's (1984) assertion that HRM x recession: +*** HRM has no effect if recession impact is low; and a positive effect if recession impact is high.
3 Market growing Union x recession: +** Union voice has a small negative effect if recession impact is low; and a positive effect if recession impact is high. HRM x recession: +*** HRM has no effect if recession impact is low; and a positive effect if recession impact is high.
4 Market not growing No effect HRM: +*** Note: This table summarises the results of the regression analyses that are significant at least at the five-percent level (**p < .05, ***p < .01). If the interaction effect between union voice (or HRM) and recession impact is significant, a short description illustrates the predicted effects while holding all other (explanatory and control) variables constant at their mean value (see Figures 1-3 for visual illustrations). If the interaction effect is not significant, the cell indicates whether the direct effect of union voice (or HRM) is significant. If neither the interaction nor the direct effect is significant at the five-percent level, the cell indicates 'no effect'. All regression models incorporate control variables, as indicated in the text. unions do not harm efficiency overall. Furthermore, Figures 1-3 illustrate the interaction effects that were confirmed in Table 1 (i.e., the different lines with increasing levels of union voice and HRM practices rotate, rather than being parallel to one another).
In sum, Figure 1 suggests that the robustness of the two participative arrangements matters in overcoming a deep recession, and that this robustness is particularly beneficial if recession impact is high.
The subsamples provide additional nuance. 3 and 4 suggest that the subsamples for workplaces in (a) firms with a dominant individual or family owner and (b) non-growing markets will be less sensitive to recession impact. Indeed, in both subsamples, we find no significant interaction effect. Subsample (1) is consistent with the idea of highly institutionalised bargaining that may lead to productive agreements: the effect of both union voice and HRM on workplace performance is positive (irrespective of recession impact). In subsample (4), union voice has no effect and HRM a positive effect, on workplace performance (again, irrespective of recession impact).
Consistent with our hypotheses, the impact of the recession is more visible in subsamples (2)

| DISCUSSION AND CONCLUSIONS
The key interest of this paper has been on the interaction between participative arrangements (i.e., union voice and HRM practices) and recession impact on workplace performance. Our results suggest that the robustness of these arrangements is key to explaining performance under high recession impact. We also found systematic differences in subsamples, implying that robust implementation of union voice and HRM practices is particularly important in firms without a dominant individual or family owner and in growing markets.
Our study's primary contribution is to the literature on financialisation and implicit contracts. Consistent with financialisation (e.g., Thompson, 2003), our findings support the idea that workplace bargains are precarious and may break down under short-term external pressures, such as a recession. A shallow implementation of union voice and HRM practices is unlikely to provide the stability needed for implicit productive bargains to survive under turbulent conditions. However, sufficient stability may be possible if implicit bargains are bolstered by robust participative arrangements. Interestingly, we find qualitatively similar findings for union voice (i.e., representative EIP) and HRM practices (of which direct EIP is a key component) in the whole sample and subsamples. This suggests that the benefits of robustness can apply to various forms of EIP.
Our findings are also consistent with the idea that 'beneficial constraints' (Streeck, 1997) may arise endogenously even in a country with a voluntarist regulatory system (Roche & Geary, 2002) here, in the form of the workplace's institutionalisation of participative arrangements. As noted above, such arrangements tend to be relatively stable over time; thus, past managerial decisions and path dependency are likely to play a key role in enabling beneficial constraints at the workplace level to support implicit contracts.
Although implicit contracts are unwritten and difficult to observe, it is hard to overstate their influence on productive activities in workplaces (e.g., Gibbons & Henderson, 2012). The disruptive impact of macro-economic crises on implicit contracts has been documented around the world. Even in Japan where implicit contracts have strong cultural support, the crisis in the 1990s made it difficult for employers to uphold some commitments, such as life-long employment and refraining from downsizing (Ahmadjian & Robbins, 2005;Ahmadjian & Robinson, 2001;Yoshikawa et al., 2005). In the United Kingdom, management sometimes marginalised union voice, and employee representatives often lacked power to challenge management in crisis (Marchington & Kynighou, 2012).
It is notable, though, that in some circumstances even shallow participative arrangements (i.e., a limited implementation of union voice and/or HRM practices) may sustain functioning implicit contracts. Shallow arrangements might even appear preferable to some employers because they initially incur lower costs. For example, the costs of strong union voice might include overcoming heterogeneous interests, slower decision-making and suboptimal compromises (Hansmann, 1996). The costs of a fuller HRM implementation might include investment in reorganising work, setting up incentive and appraisal systems and sharing the surplus with employees (Ichniowski & Shaw, 2003).
Due to these costs, the net benefit of robust participative arrangements might appear small (Godard, 2004), especially if there is no recession impact.
However, this view can be myopic. Most workplaces have a lifespan that comprises multiple periods of economic expansion and recession. Therefore, shallow arrangements may miss the opportunity to uphold and benefit from implicit contracts over longer periods. Many companies retain a competitive advantage over decades, in large part because their implicit contracts with employees are hard to imitate by competitors (Gibbons & Henderson, 2012). Therefore, our findings may also inform the resource-based view of the firm (Barney, 1991), in particular by highlighting the robustness of participative arrangements as a possible precondition for companies to achieve a sustainable competitive advantage.
Our study gives rise to many interesting follow-up questions. Having highlighted the importance of ownership structure and market growth, future research can use qualitative case analyses to investigate specific case companies in each of these circumstances. Furthermore, there is an opportunity to theorise and empirically examine additional contextual conditions regarding the exposure to financialisation depending on different types of foreign ownership (Morgan, 2014) and other company characteristics (Allen, 2014;Whitley, 1999).
Moreover, qualitative studies could give additional insights into the contents of implicit contracts in workplaces.
Specifically, an important question is the extent to which union voice and HRM practices help reduce the likelihood or the impact of a breach of implicit contracts. We also need to note that our empirical models did not explicitly measure the companies' strategies and the managers' perceptions of how important the employees are to fulfil these strategies. Future research can address these issues directlyfor example, do managers who believe in the importance of employees pursue a more robust implementation of HRM practices? It would be useful to examine the recession impact on participative arrangements in other countries with different institutions governing employment relations. For example, how did companies in countries with legally mandated 'beneficial constraints' protect implicit contracts, and were there variations across companies and workplaces?
Finally, although we focused on private-sector workplaces, examining public-sector workplaces could lead to fresh insights. In a deep recession, the public sector is subject to different pressures and influences (e.g., the role of politicians and a different type of public scrutiny).
For organisational policy, this article implies that we need to take into account the contextual conditions of participative arrangements. Just like the business cycle fluctuates, so does the workplace's product cycle. A workplace in a growing market today might find itself years later in a declining market, with different implications for its participative arrangements. Although infrequent, changes in ownership structure (e.g., from family ownership to stock market listing) can also influence the context of employee participation. Employers might need to adapt their participative arrangements to these changing circumstances. However, in the employees' perceptions, changing implicit contracts might appear like reneging on them. Thus, adapting employee participation requires careful planning and open and honest communication to arrive at new and productive agreements.
Social Research as the originators of the 2004 and 2011 Workplace Employment Relations Study data, and the UK Data Archive at the University of Essex as the distributor of the data. The National Centre for Social Research was commissioned to conduct the survey fieldwork on behalf of the sponsors. The usual disclaimer applies.