Introduction: theoretical perspectives on innovative capacity and job quality

1 While there was widespread consensus that Taylorism had been the master trend in the development of work from the 1930s to the 1950s, the first misgivings about extrapolating the trend into the distant future came from researchers working on the implications of advanced automation for work organisation (Blauner, 1964; Woodward, 1970; Gallie, 1978). The argument was that automation necessarily implied a reversal of the historic trend towards an ever-increasing division of labour and a reintegration of work tasks that restored significant control to employees in the running of complex and highly integrated production systems. The combination of the relative volatility of such systems and the high costs of downtime meant that the active intervention of employees to anticipate and manage problems became crucial to economic performance.

2 The notable development in the 1980s was the emergence of more general theories of management that predicted an increased focus in workforce policies on improvements in job quality as a condition for productivity and competitiveness in an era of intensified international trade, increased diversification, faster rates of change, and rising demands for product and service quality. Generically, these have come to be referred to as “high performance” workforce policies. Richard E. Walton (1985) argued that a revolution was under way in the management of work, involving a shift from seeking to increase productivity through detailed control of employees’ task performance to winning their commitment by improving the quality of their jobs. In the new era, employees would be given greater discretion over their work tasks, would have good opportunities for training, would be given a voice in organizational matters and have higher levels of job security. In a similar vein, Edward E. Lawler (1986) argued for the benefits to employers as well as to employees of systems of “High-Involvement Management”, a general line of thinking which, under different names, became increasingly influential in the 2000s (Appelbaum et al., 2000; Lawler et al., 1995; Wood and Wall, 2007).

3 Another influential strand of theory underlined the importance of continuous learning and the competitive advantages of “learning organizations” in contexts of uncertainty and rapid change (Lundvall and Nielsen, 2007; Valeyre et al., 2009; Lorenz and Lundvall, 2011). The information technology revolution, it was argued, transformed the dynamics of modern economies by enhancing the rate of change. It is not just that the rate of innovation in ICT itself is rapid and provides pressures for more frequent organizational change, but ICT speeds up innovation in other sectors through the improvement it brings to communications. In situations of rapid change, knowledge becomes more rapidly obsolete and learning new skills and competencies becomes more crucial to improvements in performance. Learning new skills is increasingly “learning through experience” and “learning through doing”, rather than through formal channels of knowledge acquisition. But such learning depends upon the way an organization is structured—its design, practices and culture (see also Evans et al., 2006; Hoyrup, 2010). In particular, it requires a culture that promotes the importance of learning, relative flat organizational structures and forms of job design that give employees the discretion that allows them to learn through experience.

4 The factors pointed to by these broad theories of high performance and discretionary learning workforce policies have received support from more recent research on employees’ “innovative work behaviour”, defined in terms of “finding, suggesting and implementing new and beneficial work-related ideas” (De Spiegelaere et al., 2014). In particular, in addition to the provision of opportunities for formal and informal learning, a wide range of studies has confirmed the importance of control over the immediate job (autonomy, task discretion) for innovative work behaviour (Hammond et al., 2011; De Spiegelaere et al., 2016). Further, there is evidence that job security may be important for employees’ capacity to be innovative. Surveying the literature, Tahira M. Probst (2009) finds that job insecurity may lead to risk-averse thinking, behavioural rigidity and a lower willingness to engage in organizational citizenship behaviours whereby people are prepared to step outside their formally defined roles to benefit the organization. Individuals threatened with lay-offs are less able to solve a creative task (Probst, Stewart, Gruys and Tierney, 2007); and downsizing significantly reduces aspects of the work environment that favour creativity, such as individual freedom on the job, access to resources, supervisory encouragement and work group support (Amabile and Conti, 1999).

5 This theoretical and research literature points to the growth of managerial practices that favour a number of job characteristics that have also been regarded as central to good job quality. These can be termed “innovation-conducive job quality” (ICJQ). They include in particular employee participation in decision-making, the importance of good learning opportunities at work and job security. While there has been research on the overall prevalence of such practices, there has been little attempt to examine their implications for inequalities between different categories of employee [1]. The primary focus has been on class differentials and, even in this respect, the arguments developed in the literature are ambivalent. Some point to a positive effect of new workforce practices on the work conditions of less skilled occupations, while others imply that the benefits are likely to be primarily for the most highly skilled occupations.

6 The advocacy of high involvement management was inspired partly by the relative success of the Japanese car industry in drawing on employee initiatives to enhance production quality and by the Volvo experiment in Sweden, which involved a more radical delegation of decision-making to work teams. Both examples suggested that improvements in the quality of work could be beneficial for innovation for those in relatively routinized work, allowing semi-skilled workers to make a significantly greater contribution. In this scenario, the extension of high performance policies could be expected to lead to a degree of convergence in job quality across the workforce.

7 The “learning organization” literature, however, which places a stronger emphasis on the role of ICT and rapid technical change in reshaping the challenges faced by organizational structures, implies that such policies will bring greater benefits for those in the most advanced technical sectors and hence more skilled employees. This is also an implication of a literature on the effects of technical change on the occupational structure (Autor et al., 2003; 2010; Goos et al., 2009) which suggests that advanced technology is complementary to the skills of those in higher occupations but tends to displace the skills of employees in intermediate-level jobs. These differential effects would imply that the beneficiaries of policies to improve job quality in the interests of innovation would tend to be the higher skilled and that such policies would increase workforce polarization with respect to the quality of jobs.

8 While the existing literature offers contrasting scenarios of possible trends with respect to class differentials there has been little consideration of whether the spread of high performance workforce policies increase or decrease the relative job quality of other disadvantaged labour market groups, such as those in less developed regions of Europe, women and employees on temporary contracts. There is a need, then, to provide a broader assessment of changes in differentials in ICJQ by region, occupational class, gender, and contract status.

9 Further, both perspectives assume that changes in production processes lead to a relatively general process of change in advanced societies. There is evidence, however, that some aspects of the quality of work are affected by the nature of institutional systems in societies (Gallie, 2007a; 2013). In particular, those countries in which organized labour has a stronger institutional position in the formulation of national policy generally have better working conditions and lower levels of class inequality. It seems plausible that such countries will have lower class differentials with respect to those aspects of job quality that are conducive to innovation.

10 Employer policies with respect to job quality also may vary with labour market conditions, a factor that could be particularly important given the scale of the economic crisis that followed the banking collapse. Arguably, in periods of economic crisis, employers have an incentive to cut costs by reducing labour rather than improving conditions, while in periods of expansion, the emphasis shifts to labour retention and improvement of performance by stimulating motivation through better quality work (see for instance Ramsay, 1977). It may be important then to distinguish developments in the period 2007 to 2010 when, in most Western countries, GDP was still below its pre-crisis peak from developments in the period between 2010 and 2015, when many (although certainly not all) of the European economies were once more expanding.

11 The next sections discuss the data source and the measure of ICJQ, before turning to the analysis of the trends in differentials in ICJQ between occupational classes, men and women and permanent and temporary workers to assess whether the patterns point to convergence, polarization or stability.

1. Measuring Trends in Innovation-conducive job quality

2. Convergence or polarization in the distribution of jobs with innovation-conducive job quality

2.1. Region

17 A first issue is how far developments in management policies favouring the introduction of new ICJQ type jobs have affected the relative position of employees in the more developed and less developed regions of Europe. There could be conflicting expectations. It might be that the trends of change in work organization follow the logic of cumulative advantage, so that workforces in the richer and more technically advanced countries disproportionately reap the benefits of new developments. Alternatively, it might be the case that the rate of change would be faster in the less developed countries because of stronger pressure to modernise, greater incentives for foreign investment and looser regulative controls over employers.

18 The average scores of ICJQ jobs in the different regions of Europe and the trends over time are given in figure 1. The columns show the average ICJQ score of employees in each region. A first point to note is that the Nordic countries had the highest average score in all three time periods. They were followed by the North West countries and the Continental European countries. Average ICJQ scores were lowest in the Mediterranean and East European countries. The pattern is consistent with the view that the prevalence of ICJQ is partly affected by the level of economic development.

19 It is also notable that ICJQ was increasing in the European workforce in the period 2005 to 2015 as predicted by the scenarios discussed in the introduction (Figure 1). ICJQ scores for “all regions” rose during the economic crisis and, even more significantly, in the period of economic recovery. The pattern of change over time varied however considerably between regions: there was an overall growth of such jobs between 2005 and 2015 in the North West, the Continental, the Mediterranean and Central East countries. In contrast, there was a significant decline of such jobs in South East Europe over the period of the economic crisis (although this effect had disappeared by 2015) and in the Nordic countries (once changes in workforce composition had been controlled). There was no consistent relationship between changes in ICJQ and the economic crisis. In the Mediterranean and the Central East countries, the growth was primarily in the period of economic crisis, while in the North West and Continental countries, it occurred over the period 2010 to 2015.

20

Figure 1. Trends in ICJQ scores by European region

Figure 1. Trends in ICJQ scores by European region

21 The implications of such trends for regional convergence or polarization can be assessed by a regression model that includes interaction terms between region and year. Table 2 shows the change in regional scores relative to the Nordic countries in 2010 and 2015 compared with the pre-crisis pattern in 2005. The first column shows that, between 2005 and 2010, there was an improvement in ICJQ scores relative to the Nordic countries in the Mediterranean and Central East European countries. By 2015, however, there was a more widespread reduction of the gap with the Nordic countries, including the North West, Continental, Mediterranean, and Central East countries. The final column, which takes account of differences in age, sex, class composition and industry structure, confirms improvement in ICJQ in these regions relative to the Nordic countries independently of compositional differences. This suggests a trend towards regional convergence in managerial policies with respect to job quality.

Table 2. Interaction coefficients for change in regional differentials compared with the nordic countries

	Year 2010 Interaction Coeffs		Year 2015 Interaction Coeffs		Year 2015 Interaction Coeffs with controls
North West	0.05	n.s.	0.28	***	0.28	***
Continental	-0.02	n.s.	0.21	***	0.33	***
Mediterranean	0.15	**	0.10	(*)	0.14	**
N. East	-0.05	n.s.	0.12	n.s.	0.17	n.s.
C. East	0.17	**	0.19	***	0.22	***
S. East	-0.13	(*)	0.06	n.s.	0.04	n.s.

Table 2. Interaction coefficients for change in regional differentials compared with the nordic countries

Note: Countries weighted by size. Coefficients, derived from an interaction model for employees including all regions (all years), show the change in differentials between each region and the Nordic countries in 2010 and 2015 respectively compared with the differential in 2005. Significant positive coefficients indicate a reduction of differentials, significant negative coefficients indicate an increase. Sig: p<=0.001=***; p<=0.01=**; p<=0.10=(*). Controls for age, sex, class and industry structure. Unweighted Ns with controls: N. West (5447); Nordic (10486); Continental (21043); Mediterranean (10059); N. East (6586); C. East (11962); S. East (4270).

Conclusion

38 A growing literature has argued that, in advanced economies, employers are seeking to improve certain aspects of job quality in the interest of greater productivity and innovation. There has been little examination, however, of whether such policies tend to accentuate or reduce previous divisions between relatively advantaged and disadvantaged categories of worker. This paper addresses this issue focusing on the subset of job quality characteristics frequently cited as favouring greater motivation, capacity to develop new ideas and cooperation in innovative change. These are task discretion (or control over the immediate job task), the ability to use initiative in work, formal and informal learning opportunities and job security. Indicators of these were aggregated to construct a measure of overall “innovation-conducive job quality” that correlated with reports of innovative work behaviour and of innovation in work organizations. The measure was then used to assess change in distributional patterns between 2005 and 2010 with respect to European region, occupational class, gender, and contract status.

39 There was confirmation that such jobs were becoming increasingly prevalent in the overall European workforce in the period 2005 to 2010. They increased both during the economic crisis and, even more significantly, in the period of economic recovery. The patterns at regional level were diverse. There was an overall growth of such jobs between 2005 and 2010 in the North West, the Continental, the Mediterranean and Central East countries. The only evidence for a significant decline of such jobs was in South East Europe over the period of the economic crisis (although this effect had disappeared by 2015) and in the Nordic countries (if one controlled for changes in workforce composition).

40 The regions also differed substantially in the prevalence of jobs with ICJQ. In each time period, they were most common in the Nordic countries and least common in the South East countries. A notable finding of the analysis, however, is that there was a degree of regional convergence between 2005 and 2015, with a reduction of the gap between the Nordic and several other regions. This convergence was evident both for the other EU-15 regions and for the Central East countries.

41 With respect to differences related to personal characteristics, it is notable that there was no evidence that women were generally less likely to be in jobs with ICJQ than men at any period between period 2005 to 2015. This is perhaps surprising given the evidence of the disadvantage women experience with respect to pay and career opportunities. But it is consistent with other evidence on the lack of gender differences in intrinsic job quality (OECD, 2017). It may be related to the fact that women are more likely to be in jobs where the very nature of the work involves direct interaction with other people (pupils, people in need of care, customers). The unpredictability of such interpersonal work is likely to favour significant discretion to the employee.

42 Differentials by class and contract status, however, proved more problematic. The period of the economic crisis led to a significant deterioration of the relative position of the low skilled both overall and in the Continental, Mediterranean and Central East countries. However, with the exception of the Continental countries, the increase in class differentials was generally reversed in the period of economic recovery. These trends largely reflected changes in the job characteristics of the low skilled with ICJQ scores declining between 2005 and 2010 and rising again in the period 2010 to 2015. The quality of jobs of the low skilled appears then to be particularly sensitive to labour market conditions. The changing pattern of class differentials fits well theories of the cyclical nature of employer workforce strategies with respect to class (Ramsay, 1977), with lower level employees bearing the brunt of coercive organizational change in periods of economic downturn but experiencing more accommodative management policies in tighter labour markets.

43 In contrast to the cyclical pattern for the low skilled, there was a deterioration in the relative position of those on temporary contracts both in the period of the economic crisis and in that of the recovery. This largely reflected the changing position of temporary workers in the EU-15 countries—in particular in the North West and Mediterranean countries (and in the Nordic countries in the period 2010 to 2015). An important, but not the only, factor in this growing disadvantage was a decline in security. There was no significant change in the relative position of temporary workers in the East European countries. It is a finding consistent with other research. A study drawing on workplace surveys in France and Britain found a negative association between the use of temporary work and productivity growth (Askenazy et al. 2016), suggesting that a focus on flexibility and short-term cost savings leads to working conditions that are unfavourable for long-term innovation. At least in France, it was associated with changes in the nature of temporary work since the great recession—in particular, the resort by employers to temporary contracts of shorter duration. Very short duration contracts are especially unlikely to allow for the types of task discretion and knowledge acquisition, let alone the job security, that has been shown to be conducive to employee’s capacity or willingness to innovate.

44 Overall, then, the evidence supports the view that there has been an increase in the prevalence of jobs with ICJQ in Europe in recent years. This increase has had quite distinct implications for different types of workforce inequality. It has been associated with a reduction in regional inequalities, relative stability in gender inequalities, a cyclical effect with respect to inequalities of class and a sustained deterioration in the position of temporary workers in the EU-15 countries. Interpretation of these trends must be largely speculative. While the cross-sectional data upon which the analysis is based is invaluable for examining broad trends, it does not permit rigorous exploration of the factors affecting these trends. Individuals are not followed over time and sample numbers are too small to allow for detailed investigation of the potential impact of compositional changes within the different categories of employee that have formed the basis of comparison.

45 On the basis of the evidence for the period 2005 to 2015, the introduction of new management practices to enhance the innovative capacity of employees has proved to be heavily constrained by pre-existing structures of social inequality. Despite a significant overall increase in the prevalence of innovation-conducive jobs, there has been no improvement over time in the social distribution of such jobs. More optimistic predictions of the knowledge-based economy, whereby the growing requirements for high performance and innovation would translate into a greater utilisation of the capacities of the broader workforce and a convergence in the job quality of different categories of employee have not come to pass.

46 It is clear that, if change is to take place, it will not be through some deterministic effect of changing technologies or production processes, but will require sustained intervention through public policy. There are, however, grounds to think that policies to enhance job quality can make a difference. The prevalence of innovation-conducive jobs varies substantially between different European regions. In particular, it is notable that the Nordic countries stand out as having had a significant advantage in this respect throughout the period, even if there has been some narrowing of differentials over time. This cannot be explained in terms of differences in workforce or industry composition. It is most plausibly attributable to quite distinctive policy orientations that have placed improvements in the quality of work high on the policy agenda.