A Slowdown in Human Resource Management Investments?

One area that has not been discussed a great deal in the broader literature on the productivity puzzle is that of human resource management (HRM) investments; that is, the HR practices that managers may implement in pursuit of higher productivity. If the recession had reduced the rate at which HRM investments were made—or lowered the rate of return on such investments—this might have contributed to a slowdown in productivity growth.

The broad literature on HR practices and workplace performance (e.g. Huselid, 1995;Bloom and Van Reenen, 2011) tends to focus on three sets of practices which are expected to have positive implications for productivity: first, work organization practices which give workers a greater level of autonomy, aid collaboration, and raise their skills; second, performance or quality management practices which seek to more closely manage workers' effort and output;and third, incentive pay schemes which seek to motivate workers financially.

It is apparent from existing work (e.g. Wood and Bryson, 2009) that some of the practices cited above, such as team-working and the use of quality targets, became more prevalent in Britain over the period 1998-2004, when the economy was growing strongly. Here we investigate whether the rate of growth of these practices might have slowed since the mid-2000s, or whether the returns to such HR practices might have diminished, in such a way as to have contributed to the general slowdown in productivity growth.

Alongside the three sets of practices considered above, we also look at arrangements for employee voice. Collective employee representation through trade unions was known to be negatively associated with workplace performance in Britain in the 1980s and 1990s, but unionization is known to have weakened in recent decades, whilst arrangements for direct communication between managers and employees have grown in popularity (Blanchflower and Bryson, 2009).

Our analysis again calls on the Workplace Employment Relations Survey, but this time employs data from the cross-section surveys of 1998, 2004, and 2011. We use data on private sector workplaces with ten or more employees and, first, chart the incidence of the HR practices

Table 5.6. Share of employment in private sector workplaces with specific HR practices, 1998-2011

1998

2004

2011

2004 vs 1998

2011 vs 2004

2011 vs 1998

%

%

%

Signif.

Signif.

Signif.

Work organization:

Semi-autonomous team- working+

44

35

48

***

***

Functional flexibility+

79

78

82

**

Training for 80%+ experienced employees+

21

41

49

***

***

***

Quality management:

Problem-solving groups

49

34

30

***

*

***

Quality targets

55

58

63

Appraisals for 80%+ nonmanagerial employees Incentives:

53

69

78

***

***

***

Profit-related pay

53

44

43

***

***

Share-ownership scheme Voice:

32

33

28

**

Representative + Direct

26

31

33

**

***

Representative only

43

28

24

***

*

***

Direct only

11

21

23

***

***

Neither

20

20

19

Base: employment in private sectorworkplaceswith ten+ employees. Notes: + for the largest occupational group.

Key: *** = sig. at 1 per cent; ** sig. at 5 per cent; * sig. at 10 per cent. Source: Workplace Employment Relations Survey

discussed above over the course of the three surveys. We then examine the associations between these HR practices and a subjective measure of workplace productivity in each year, as a rough indication of whether there may have been changes in returns.

Table 5.6 shows the percentage of employees who work in establishments where the specified practices operate.[1] Considering first those practices relating to work organization and skills, we see increases in the use of team-working, in the use of functional flexibility, and in the intensity of training between 2004 and 2011. The rise in team working reversed an earlier decline seen between 1998 and 2004, whilst the increased intensity of training represented the continuation of a prior trend.

Turning to quality and performance-management practices, we see a decline in the use of problem-solving groups and a rise in the use of performance appraisals but, again, neither change was unique to the period 2004-11. On incentive pay, we see a small decline in the prevalence of share ownership schemes, and in respect of voice, we see the continuation of a shift away from sole reliance on representative arrangements and towards the use of direct forms of communication, either alone or in combination with forms of employee representation.

On the whole, these patterns indicate a progressive shift away from formal, collective approaches to the management of employees and employee performance (i.e. problem-solving groups, group-based incentive pay and engagement with unions) towards a more individualistic focus that encompasses upskilling and the direct management of quality and performance. However, there appears to be no obvious change in trajectory between 1998-2004 and 2004-11. These patterns do not therefore suggest that that the recent period of recession in Britain was characterized by any particular slowdown in the diffusion of 'productivity-enhancing' HR practices.

The evidence for any changes in returns is also weak, insofar as we can gauge with our data. WERS only provides accounting data on performance for a small subset of workplaces and so we must rely on the subjective rating given by the workplace manager. As noted earlier, they are asked to rate the level of labour productivity at their workplace relative to the average for their industry and answer on a five-point scale from 'A lot above average' to 'A lot below average'. We can then investigate whether specific practices are associated with levels of productivity in a given year and whether these 'returns' appear to change over time. If the returns diminish, this might suggest that increased diffusion of the practice is making a smaller contribution to productivity growth. One must, however, accept that there are caveats, given the crosssectional nature of the data and the subjective nature of the performance rating.

The results of this analysis are presented in Tables 5.7 and 5.8. In the first of these tables, the individual practices shown in Table 5.6 are included together in an ordered probit regression of the workplace's subjective productivity rating. Once we control for a set of observable workplace characteristics, including the size of the workplace, its industry sector and its location, we see no consistent pattern of changing returns. The most notable patterns are a reduction between 2004 and 2011 in the productivity advantage conferred by functional flexibility, and a reduction between 1998 and 2004 in the productivity

Table 5.7. Ordered probit regression of labour productivity on specific HR practices, private sector, 1998-2011

1998

2004

2011

1998

2004

2011

Controls?

No

No

No

Yes

Yes

Yes

Semi-autonomous

team-workingA

0.162

[1.26]

0.045

[0.44]

0.048

[0.50]

0.097

[0.80]

-0.022

[-0.21]

0.062

[0.63]

Functional

flexibilityA

0.303**

[2.16]

0.278**

[2.51]

0.010

[0.10]

0.393***

[2.94]

0.264**

[2.46]

0.055

[0.53]

Training for 80%+ experienced employeesA

-0.067

[-0.49]

0.027

[0.25]

-0.059

[-0.63]

-0.073

[-0.51]

0.006

[0.05]

-0.112

[-1.15]

Problem-solving

groups

0.071

[0.61]

0.129

[1.06]

-0.049

[-0.39]

0.045

[0.41]

0.119

[0.96]

0.011

[0.08]

Quality targets

0.065

[0.58]

-0.072

[-0.65]

0.196**

[2.04]

0.138

[1.18]

-0.052

[-0.46]

0.157

[1.65]

Appraisals for 80%+ non-managerial employees

0.096

[0.76]

0.218*

[1.93]

0.122

[1.17]

0.024

[0.19]

0.253**

[2.04]

0.157

[1.40]

Profit-related pay

0.011

[0.08]

0.181*

[1.71]

0.098

[0.99]

0.184

[1.36]

0.216**

[2.08]

0.067

[0.66]

Share-ownership

scheme

0.163

[1.18]

-0.213*

[-1.73]

0.050

[0.41]

0.213

[1.59]

-0.211

[-1.62]

0.075

[0.60]

Voice (ref = none): Representative + direct

-0.111

[-0.67]

0.062

[0.39]

0.191

[1.39]

-0.159

[-0.98]

0.237

[1.51]

0.160

[1.05]

Representative only

-0.399***

[-2.75]

-0.053

[-0.34]

0.021

[0.13]

-0.436***

[-2.95]

0.249

[1.54]

-0.001

[-0.01]

Direct only

0.084

[0.43]

0.194

[1.43]

0.153

[1.27]

0.050

[0.26]

0.081

[0.59]

0.133

[1.08]

N

1,259

1,210

1,337

1,258

1,210

1,337

Base: private sector workplaces with ten or more employees.

Control variables: workplace size; industry sector; region; largest occupational group; whether part of multi-site organization; number of competitors in main market; degree of competition in that market; whether market local/regional/national/international; whether market growing/mature/declining/ turbulent.

Key: л questions refer to the largest occupational group at the workplace

*** = sig. at 1 per cent; ** sig. at 5 per cent; * sig. at 10 per cent [t-statistics in parentheses].

Source: Workplace Employment Relations Survey

Table 5.8. Ordered probit regression of labour productivity on count of HR practices, private sector, 1998-2011

1998

2004

2011

1998

2004

2011

Controls?

No

No

No

Yes

Yes

Yes

Count of HR practices

0.113***

[2.77]

0.103***

[2.75]

0.051

[1.53]

0.111***

[2.92]

0.091**

[2.32]

0.057

[1.60]

N

1259

1210

1337

1258

1210

1337

Base: private sector workplaces with ten or more employees.

HRM count is a count of the number of HR practices from (min=0; max=6).

Control variables: as listed under Table5.7, plus whether any profit-related pay, any share ownership scheme, and type of voice arrangement.

Key. *** = sig. at 1 per cent; ** sig. at 5 per cent; * sig. at 10 per cent [t-statistics in parentheses].

Source: Workplace Employment Relations Survey

This corresponded with employees' experience. When asked 'Did any of the following happen to you as a result of the most recent recession whilst working at this workplace?' one-quarter (26 per cent) of private sector employees said 'My wages were frozen or cut', making it the most common response alongside 'My workload increased.' Unsurprisingly the incidence of pay cuts and freezes was strongly associated with the extent to which workplaces were adversely affected by the recession. In four-fifths (82 per cent) of the cases where HR managers reported freezing or cutting wages, it was accompanied by at least one other action, usually to cut costs. For example, over one-third (36 per cent) of those freezing or cutting wages had also instituted a freeze on filling vacant posts, 28 per cent had reduced paid overtime, 28 per cent had 'postponed plans to expand', 27 per cent had made 'changes in the organization of work', and 22 per cent had made compulsory redundancies.

Further insights can be gleaned regarding pay setting during the recession in relation to the last pay settlement for the largest nonmanagerial occupation at the workplace. The percentage of settlements resulting in a pay freeze or cut doubled between 2004 and 2011 from 12 per cent to 26 per cent. Again, the influence of recession was in clear evidence: whereas only 15 per cent of workplaces who reported no adverse effect of the recession had instituted a pay freeze or cut in the last pay settlement for the largest non-managerial occupation, this rose to 36 per cent where the HR manager said the workplace had been affected 'a great deal'.

As noted earlier, the decline in real wages in Britain since the onset of recession is almost unprecedented in a period of low inflation, raising questions as to how management has been able to make such sizeable wage adjustments. One common hypothesis is that the reduced incidence of collective bargaining and a loss of union bargaining power has limited unions' ability to block pressures for wage reductions. The incidence of workplace trade unions and membership density changed little between 2004 and 2011, although there was a reduction in the scope of collective bargaining in the private sector, which may be indicative of unions' reduced ability to maintain influence over a wide bargaining agenda (Van Wanrooy etal., 2013). However, unionization is not correlated with the likelihood of managers saying they froze or cut wages in response to the recession, nor to wage freezes or cuts in the last pay settlement for the largest non-managerial occupational group. Nor has there been a noticeable decline in the size of the union wage premium— instead we see counter-cyclical movement, consistent with previous studies (Figure 5.8). It is therefore difficult to pinpoint a break in union power which may have provided employers with the opportunity to

Union membership wage premium, 1994-2012 Source

Figure 5.8. Union membership wage premium, 1994-2012 Source: Authors' calculations from Labour Force Survey

downwardly adjust real wages. If such a change has occurred, it may date back further than the onset of the recession itself.

There are two other changes which analysts point to as potential reasons for the weakness of real wage growth since the recession: welfare reform and immigration. Welfare reform in the UK has been extensive in recent years and has focused on increasing labour market participation of the inactive and unemployed (OECD, 2013a: 67-77). It can affect employer wage setting and jobseeker behaviour in a variety of ways that can limit real wage growth. For instance, unemployed jobseekers may be prepared to accept job offers at lower rates of pay than might have been the case in the absence of reform. We are able to identify those workplaces most likely to draw applicants from welfare benefit recipients, and thus those most likely to be affected by welfare reform, through two data items in WERS, namely whether the workplace used the public job placement service to fill vacancies for the largest non-managerial occupation at the workplace in the last twelve months, and whether the workplace had special procedures to encourage job applications from those who had been unemployed for at least twelve months. Neither was associated with pay freezes or cuts in the last pay settlement for the workplace's largest non-managerial occupational group, nor was either associated with freezes or cuts in wages, or the reduction of non-wage benefits, in response to the recession. Thus, to the extent that welfare reform might be expected to impact most on employers engaging with the public job placement service and drawing from the unemployed for recruits, there was no discernible direct effect of welfare reform on these aspects of wage setting. Of course, it is quite possible that the reforms have had other direct effects on wage setting, and that they have had broader, less direct effects on the operation of the labour market in general.

Although the UK has experienced a very substantial inflow of migrants in the last few years—a labour supply shock that could, in principle, slow the rate of real wage growth—the empirical evidence on the link between immigration and wages is heavily contested (see Ruhs and Vargas-Silva, 2014). In 2011, for the first time WERS collected information on the number of non-UK nationals employed at the workplace, distinguishing between those from the European Economic Area (the EEA) and those outside.[2] Of those employed in private sector workplaces in 2011 non-UK nationals accounted for a mean of 7.6 per cent, of whom 3.0 per cent were non-EEA nationals. Although the percentage of non-UK nationals employed at the workplace had no bearing on wage freezes or cuts that were directly attributed to the recession, and no effect on cuts to non-wage benefits in response to the recession, the probability of a pay freeze or cut for the largest nonmanagerial occupational group in the last pay settlement rose with the proportion of non-EEA nationals employed by the workplace. One- quarter (26 per cent) of private sector workplaces had instituted a pay freeze or cut for the largest non-managerial group of employees in the last pay settlement. An increase in one percentage point in the number of non-EEA nationals employed at a workplace raised the probability of a wage freeze or cut by roughly 0.4 of a percentage point.[3] The proportion of EEA nationals was not statistically significant. One potential explanation for this finding is that a workplace's ability to employ non-EEA nationals reduces the bargaining power of employees at that workplace, thus limiting employees' ability to resist wage freezes or cuts.

If wages have fallen in response to changes in productivity levels we might expect to see 'Productivity levels within the organization or workplace' featuring prominently as an influence on the last pay settlement for the largest non-managerial occupation in the workplace.

Table 5.9. Influences on the most recent pay settlement for the largest nonmanagerial occupation

2004

2011

All

Freeze/cut

Increase

All

Freeze/cut

Increase

Financial

performance

30

36

29

36

44

34

Productivity levels

21

23

21

19

18

19

Changes in cost of living

24

11

26

21

17

22

Recruitment and retention

21

16

21

13

11

14

Industrial action

<1

<1

<1

<1

<1

<1

None of these

4

14

3

12

11

12

N workplaces

1,750

182

1,587

1,756

379

1,346

Notes: (1) Responses to question: 'Looking at this card, which of the factors listed influenced the size of the pay settlementor reviewfor [largestoccupational group]?' (2) Figuresare column percentages based on N responses, so adding to 100.

Source: Workplace Employment Relations Survey

Table 5.9 compares the influences on pay settlements in 2011 with those in 2004 for all settlements and those that resulted in a freeze or cut versus a pay increase. The most commonly cited influence is 'The financial performance of the workplace or organization': it accounted for over one-third (36 per cent) of responses in 2011, up from 30 per cent in 2004, and was particularly salient in settlements leading to a freeze or cut. 'Rises in the cost of living' was the second most commonly cited factor, and was more salient in cases where the settlement led to a pay rise. 'Industrial Action threatened or taken' rarely featured in employers' considerations at all, perhaps indicating the limitations of unions' influence over pay awards.

Productivity levels accounted for around one-fifth of responses, but they were no more heavily cited in 2011 than they were in 2004, nor did they feature more in cases where there was a pay freeze or cut. There is therefore little to indicate that productivity had become a more common consideration in wage setting as a result of the recession.27

These figures are based on the subset of coded responses available in 2004 and 2011. In 2011 a more extended set of options was provided including reference to the national minimum wage for example. Productivity levels accounted for 16 per cent of this more extended set of influences in 2011, a figure that did not differ according to whether the settlement resulted in a pay increase or not. Financial performance was mentioned almost twice as many times (31 per cent).

  • [1] We prefer this employment share to the share of workplaces with a practice, since largerworkplaces contribute disproportionately to aggregate levels of productivity.
  • [2] The EEA comprises the European Union, Iceland, Liechtenstein, Norway, andSwitzerland.
  • [3] The coefficient on the proportion of non-EEA nationals was -0.52 in the absence ofcontrols (t-stat of 3.95), falling to -0.38 (t=2.23) with controls for number of employees,single establishment organization, industry, region, union, and largest occupational group.Full results are available on request.
 
Source
< Prev   CONTENTS   Source   Next >