Management Formality, Size of Firm and Employee Evaluations of Work Working Paper No. 99 June 2008 David J. Storey, George Saridakis, Sukanya Sen-Gupta, Paul K. Edwards and Robert A. Blackburn Warwick Business School’s Small and Medium Sized Enterprise Centre Working Papers are produced in order to bring the results of research in progress to a wider audience and to facilitate discussion. They will normally be published in a revised form subsequently and the agreement of the authors should be obtained before referring to its contents in other published works. The Director of the CSME, Professor David Storey, is the Editor of the Series. Any enquiries concerning the research undertaken within the Centre should be addressed to: The Director CSME Warwick Business School University of Warwick Coventry CV4 7AL e-mail david.storey@wbs.ac.uk Tel. 024 76 522074 ISSN 0964-9328 – CSME WORKING PAPERS Details of papers in this series may be requested from: The Publications Secretary CSME Warwick Business School, University of Warwick, Coventry CV4 7AL e-mail sharon.west@wbs.ac.uk Tel. 024 76 523692 1 Management Formality, Size of Firm and Employee Evaluations of Work David J. Storey* George Saridakis** Sukanya Sen-Gupta^ Paul K. Edwards^^ Robert A. Blackburn# * Director, Centre for Small and Medium Sized Enterprises, Warwick Business School, University of Warwick, CV4 7AL. Email: David.Storey@wbs.ac.uk ** Lecturer in Business Economics, Loughborough Business School, Loughborough University. ^ Lecturer in Human Resource Management, Cardiff Business School, Cardiff University. ^^ Professor of Industrial Relations, Industrial Relations Research Unit, Warwick Business School, University of Warwick. # Professor of Small Business Studies, Small Business Research Centre, Kingston Business School, Kingston University. We are grateful for comments received on earlier versions of this paper from Jim Curran, Kim Hoque and others following presentations at Cardiff and Glasgow Universities. Paul Edwards and Sukanya Sen-Gupta were funded by the ESRC through the Advanced Institute of Management Research. 2 Management Formality, Size of Firm and Employee Evaluations of Work Abstract A headline result from WERS 2004 is that measures of employee ‘satisfaction’ are higher in small than large firms. This paper pursues this issue further. We conceptualise ‘satisfaction’ more exactly as the employee evaluation of work experience (EEWE) and provide theoretical reasons to expect the observed relationship. We then interrogate it in two main ways. First, we address the size of the workplace as well as the size of the firm, and also make distinctions between workplaces that are part of multi-site companies and stand-alone sites. Second, the relationship between size and the EEWE is often explained in terms of ‘small firm informality’. We develop a measure of formality to test this idea. Formality was indeed related to size, both of workplace and firm. It also had associations with the EEWE. Relationships here were complex, but there was a clear tendency for the effect to be negative in smaller single-site firms. 3 1 Introduction A major finding from the 2004 Workplace Employment Relations Survey (WERS) is that the employees of small firms give higher evaluations of their jobs, on a number of dimensions, than do those in large firms (Forth et al., 2006: 97-8). This result challenges some conventional wisdom, for in terms of measures such as pay levels and standard industrial relations indicators such as trade union presence small firms perform poorly. The result is, however, consistent with commentary on ‘size and morale’ going back over 50 years (Acton Society Trust, 1953). That commentary was, however, often based on poor evidence, a weak conceptualisation of ‘morale’, and a failure to be clear about whether size referred to the work group, the workplace, or the organisation as a whole (Curran and Stanworth, 1981). The WERS result puts the basic fact of high ‘satisfaction’ in small firms on a firmer footing. This paper interrogates the basis of this fact in three ways: by addressing different measures of size; by identifying a key determinant of ‘satisfaction’, the formality with which employment relations are managed and showing that this is associated with size and ‘satisfaction’; and by offering a theory of the sizemorale link. Under the third head, we could use terms such as ‘satisfaction’ and ‘morale’. These terms have, however, considerable conceptual baggage. What we are examining is the employee evaluation of (dimensions of) work experience, which for clarity if not elegance we will call the EEWE. In contrast to Forth et al., our analysis considers both the size of the workplace and the size of the organisation, firm or enterprise owning that workplace/establishment. To address this we make the following distinctions. First, we identify single-workplace firms. Where the workplace is small then, by definition, the organisation is small. Second, we identify workplaces that are part of a multi-site firm. One organisation may own several workplaces with the result that the organisation may be large even when its workplaces are small. Nevertheless, small and medium multi-site workplaces (SMMWs) may also belong to an SME. The WERS sample suggests that about 15% of small multiworkplaces belong to a small firm (5-49 employees) and 16% are part of a medium-sized firm (50-249 employees). We also found that 10% of medium multi-workplaces are part 4 of a medium multi-workplace firm. This leads us to make a further distinction between SMMWs that are part of a SME and those that are part of large firm. We begin by examining formality and how this varies by workplace size within the two categories of workplace ownership (single-site firms and multi-site firms). Formality is measured by the presence of written procedures, and rules and policies to design, measure and regulate the employment relationship. In contrast, ‘informality’ is defined as being dominated by custom and practice, with an absence of written procedures. We find strong evidence that formality increases as the workplace size increases. Then we examine whether the EEWE follows a similar pattern. We find, perhaps unsurprisingly, the reverse results: the EEWE decreases as workplace size increases. An interesting question, however, is whether small and medium single-site firms have different EEWE and management formality, and if so to what extent, from SMMW’s that are either part of a SME or part of a large firm. Our results suggest that single-site SMEs have higher EEWE and lower formality than SMMWs that are part of either a SME or a large firm. Having established the above relationships, the analysis then proceeds by studying empirically the link between EEWE and formality, holding workforce and establishment characteristics fixed. We find that the association between EEWE and management formality is quite complex. Some indicators of formality are negatively associated with evaluations, others are positively associated and others are unrelated. The remainder of the paper is structured as follows. Section 2 reviews existing work and derives hypotheses. It also highlights the distinction between workplace and firm size. Section 3 describes the dataset and our measures of formality, and summarises the empirical method. Section 4 examines the relationships among workplace and organisation size, formality, and the EEWE. Section 5 reports multivariate analyses of the influences on the EEWE. Section 6 draws out some implications. 2 Organisation size, formality and employee attitudes 2.1 Size and formality WERS 2004 and its predecessors clearly show that measures of formality in employment relations increase with the size of the workplace. The presence of an HR manager is one example (Kersley et al., 2006: 40). All indices of formality are positively associated with 5 size. This seems to be a cross-national fact, as studies in the US (Kaman et al., 2001), Australia (Kotey and Slade, 2005), and Canada (Golhar and Deshpande, 1997) suggest. It is important, however, to distinguish between the effects of workplace size and the effects of the size of the organisation owning that workplace [Marginson (1984)]. We return to this distinction after we have presented some expectations about organisation size, initially treating ‘organisation’ as covering both workplace and firm. Many behavioural differences between organisations of different sizes reflect differences in the nature of the uncertainties facing them and hence their employment relations (Storey and Sykes, 1996). The size of an organisation is related to uncertainty in three respects. First, small enterprises are much more likely to cease trading than larger enterprises (Mata and Portugal, 2002). This is a result of a combination of factors but is primarily because they lack control over their external environment as well as other factors such as under-capitalisation and perhaps poor management skills (Watson and Everett, 1998). They lack the necessary market power to influence prices and they are more likely to be dependent on a single or small number of customers. In contrast, large firms lack control over their internal environment. They are more likely than small firms to experience problems in communicating with their workforce. The conventional response to this problem is to develop formal procedures such as newsletters, formal meetings and representative structures. Formality is, therefore, argued to be a response to increased size, on the grounds that it is infeasible for the ultimate decision taker to communicate this directly to all those affected. A second key problem for large firms is to make workers, at all levels, comprehend the significance of their own actions. So, while workers in all sizes of firm understand that absenteeism causes problems for co-workers, workers in small firms may be more likely to understand the impact this may have upon the viability of the whole enterprise. To address this problem larger firms again respond by seeking to raise formality. Team spirit, which may occur more naturally in small firms, is promoted by formal communication channels and seeking to standardise procedures. A third difference is that, in small firms, diversity is more likely to be managed flexibly whereas, in large firms, it is more likely to be managed by formality. So, in large firms, a central theme is 6 to ensure that individuals are treated fairly, by which it is meant that no individual is seen to be advantageously treated, compared with comparable co-workers. We expect these effects to apply to workplaces and firms, though perhaps in different ways. Thus as workplace size increases there is likely to be a more formal division of labour and growing distance between the top management and workers. Edwards et al. (1998) report a study of worker attitudes in six large organizations, and argue that the small size of workplaces in a financial services firm promoted direct faceto-face relationships and a degree of trust, compared with the larger workplaces of other large firms. Growing firm size is likely to increase problems of communication and to promote the need for standard rules across workplaces. These considerations lead to: H1: Small workplaces will exhibit less formality than large workplaces. H2: Small workplaces owned by single-site firms (small firms) will exhibit less formality than small workplaces owned by a large (multi-site) firm. 2.2 Employee ‘satisfaction’ The view that there is a strong link between firm size and employee ‘satisfaction’ was memorably summarised by the Bolton Committee (1971: 21). “Although physical conditions may sometimes be inferior in small firms, most people prefer to work in a small group where communication presents fewer problems: the employee in a small firm can more easily see the relation between what he is doing and the objectives and performance of the firm as a whole. Where management is more direct and flexible, working rules can be varied to suit the individual. Each employee is also likely to have a more varied role, with a chance to participate in several kinds of work and better opportunities to learn and widen his experience.” This statement reflected common ambiguities at the time as to whether it was the size of the ‘firm’ or the work group which was important. The dependent variable was also loosely identified. Ingham (1970) made one development in suggesting that small-firm employees balance intrinsic job qualities such as close working relationships against relatively poor pay and conditions, so that there may be no single dimension of satisfaction. More recent work, however, shows that satisfaction is multi-dimensional 7 (Kalleberg, 1977) while many scholars prefer to address indices of job quality, such as pay and autonomy, as opposed to ‘satisfaction’ (Green, 2006). A key issue with satisfaction is that an employee may be satisfied only because s/he is measuring work against very low expectations. This is certainly true, but it is equally true of the autonomy measures favoured by some researchers, for workers must use some implicit benchmark when they assess how much autonomy they have. The WERS researchers, none the less, produced a set of 22 measures which they labelled ‘employee needs’ (Forth et al., 2006). It comprised evaluations of three broad areas: the nature of the job (8 items); the quality of communication from managers (4 items); and the quality of the manager-employee relationship (10 items). Employees rated these on five-point scales. As noted earlier, we have chosen to call this set the EEWE: it addresses issues broader than satisfaction with the job itself, but it plainly is a measure of employee evaluations of the quality of the work experience, and not direct measures of that quality. We take the view, evidently shared by students of autonomy and other aspects of job quality (e.g. Gallie, 2003; Gallie et al., 2004), that it does make sense to ask employees about these things, since there can be no other source for personal assessments of work experience. We acknowledge that there are potential problems with self-assessments. But there is only a problem if we have reasons to expect that assessments will be systematically skewed according to size of firm on variables of interest: in this case, that small firm workers are, as a group, more or less likely to give a higher evaluation than large-firm workers, for reasons to do with socially structured expectations as opposed to their ‘true’ assessments. It has certainly been shown that some small-firm workers in low-paying jobs are content because of very limited alternatives (Ram et al., 2007). But the whole point about those jobs is that they are on the margins of the economy. There is no reason to expect skewed expectations across workers in small firms as a whole. Empirical evidence comes from Tsai et al. (2007) who examined three contrasting sectors and showed that the employment relationship was negotiated differently in each. But they also found important similarities across the sectors. This led them to identify two kinds of size effect. The first is a ‘pure’ effect of size, which reflected in particular direct personal relationships. The second covers processes that are correlated with size 8 rather than the direct result of it. These are the factors discussed below. In short, we can refine Bolton’s speculations into the expectation that direct personal relationships in small organisations will lead to positive evaluations of work experience. And we can unpack ‘organisations’ to test whether this is a workplace effect, an effect of the size of the firm as a whole, or both. H3: The employee evaluation of work experience (EEWE) is higher in small than in large workplaces. H4: The EEWE in small single-site workplaces (small firms) is higher than in small workplaces owned by a large (multi-site) firm. 2.3 Formality and employee evaluations of work Size may be only a ‘proxy for other more theoretically relevant variables that are usually poorly specified’ (Kalleberg and Van Buren, 1996: 47). Others dispute whether there is any determinate influence of size: once industrial sector and other factors are taken into account, ‘size of firm is not, in itself, an important factor in explaining differences in levels of job satisfaction’ (Curran and Stanworth, 1981: 343). Kalleberg and Van Buren (1996: 63), their reservations notwithstanding, find a high level of job autonomy in small firms, which they attribute to a tendency of large firms to constrain the freedom of choice open to workers. Idson (1990) addressed the size of the workplace rather than the whole organization. He argued that a lack of ‘satisfaction’ in large firms could be explained as a result of the nature of formality, which meant that the rules of any particular workplace would not match the expectations of a significant number of workers. These studies, together with the results of Tsai et al. described above, lead one to expect that there will be a ‘pure’ size effect as stated in Hypotheses 3 and 4 but also that formality may play an independent role. It is plainly also important to control for the sectoral and other effects identified by Curran and Stanworth, to see whether ‘pure’ and ‘formality’ effects disappear when other factors are introduced. H5: Formality has a negative impact on the EEWE, holding workforce and establishment characteristics fixed. 9 3 Data, formality measure and measure of employee evaluations 3.1 Data Our analysis uses data from the WERS 2004 which is a nationally representative crosssection survey, based on a stratified random sample of establishments and a sample of employees at those establishments. Here our focus is on private sector workplaces with a minimum employment size of 5 employees. We have excluded sole UK establishments of foreign organisations. We draw on data from two sources. The Management Questionnaire consists of face-to-face interviews with 2,295 senior managers dealing with industrial, employee or personnel relations at the workplace (response rate 64%); and the Employee Questionnaire, consists of a self-competition questionnaire distributed to a random sample of up to 25 employees in each workplace with 5 or more employees located in Britain (response rate 60%). In the descriptive analysis, the Management and Employee data have been weighted using the ‘empwtnr’ and ‘seqwtnr’ variable, respectively. i The ‘empwtnr’ has been applied to the management questionnaire because the analysis reflects the proportion of employees to whom the formality procedures have been applied. Further information regarding the sample design, selection and weighting can be found in the technical report (Chaplin et al., 2005). The matched employer-employee data comprised 22,451 employees working in 1,733 workplaces. Our number of observations, however, was reduced by 2% because we excluded sole UK establishments of a foreign organisation; this reflected our desire not to conflate genuinely small firms with the single sites of potentially large overseas firms. Given that the small firm literature presented earlier concerns private sectors firms, we have also excluded public sector workplaces. Table 1 provides information on the distribution of employees by workplace size, in workplaces which are part of a multi-site firm and single-workplace firms. This table shows that 47% of employees of the single workplace firms work in workplaces with 549 employees, compared with 26% of those in workplaces that are part of a multi-site firm. It is worth mentioning that only 14% of multi-site employees are in SMEs (5-249 employees). This emphasises that multi-site employees work primarily for large firms (250+ employees). ii 10 Table 1: Distribution of employees by size and type of workplace (%, weighted estimates) Workplace size 5-49 50-99 100-249 250-499 500+ Observations All workplaces A) Workplaces which are part of a multi-site firm B) Singleworkplace firms 31.7 15.9 19.8 16.4 16.2 14,249 26.3 14.9 22.2 18.7 17.9 10,649 47.4 18.7 13.1 9.8 11.1 3,600 Note: The null hypothesis that the proportions are equal between workplaces that are part of a multisite firm and single-workplace firms of the same workplace size was rejected for each case. 11 3.2 Measurement of formality Although informality has long been a defining feature of the small firm debate it is necessarily a matter of degree, and so this presents problems of definition. As Hill (1974) pointed out, small firms have elements of formality, and large firms often practise informality (as a massive literature on the breach or neglect of formal rules testifies). Case study researchers have explored the negotiability of rules and the meaning of formal and informal relationships (e.g. Holliday, 1995). Within WERS it is possible only to derive indicators of the extent to which formal practices and procedures exist. If one were seeking an index of formality, one might identify measures that are not available in WERS or define measures differently, for example with explicit reference to formal systems. Consequently, the WERS measures are imperfect, but they have the benefit of covering a wide range of HR practices including recruitment, appraisal and pay. In particular, they reflect the different dimensions of formality identified above, that is, formality in relation to performance management (e.g. appraisal), training and employee development, and methods of handling communication with employees. We used the management questionnaire to identify formal practices, taking only indicators that reasonably clearly indicated a formal structure or process. The decision of what measures should be included was not straightforward. For example, WERS has a question on benchmarking, defined as ‘examining the way things are done at other workplaces and comparing them with this establishment’. Benchmarking might certainly be part of a structured approach to the workplace, but it could be done through very informal methods. In addition, whether or not comparison with outside organisations takes place may not impinge directly on an establishment’s workers. We therefore excluded benchmarking as a criterion. Similar arguments apply to the use of just-in-time methods and the attainment of ISO quality standards. That is, the focus is on formality in relation to employee relations specifically, though wider studies of workplace production organisation might embrace benchmarking, JIT and ISO. We identified 12 measures of formality, most of which relate directly to the management of employees. Some of these are self-explanatory, for example the existence of formal procedures in such areas as equal opportunities. We also included coverage of 12 what was explicitly called a ‘formal strategic plan’. This question was followed in WERS by a follow-up on the issues covered by the plan. An initial idea was to include only cases where employee issues were mentioned, but there were very few where such issues were mentioned and thus the primary variable on use of a plan was used. Investors in People (IiP) is also a natural indicator because the standard is heavily concerned with documented formal procedures and systems (Hoque, 2003). In relation to communication with employees, we used two measures: whether meetings were held between senior managers and the workforce, and the use of some structured communication device such as newsletters. The latter is clearly formal; the former is less so, but none the less a meeting, as distinct from unplanned face-to-face interaction, is a reasonably clear indicator of a considered approach. Two other measures warrant comment. First, we included the existence of any non-pay benefits for the largest occupational group; WERS listed five such benefits including an employer pension scheme and sick pay in excess of statutory minima. The logic here was that such benefits embrace ‘instrumental-economistic’ rewards other than basic pay and are likely to reflect a structured and long-term approach to rewards, as opposed to payment of a basic wage alone. Second, we took one non-HR measure, the existence of a target for any of a long list of measures including sales, costs, quality, and labour turnover. We took a target to be an indicator of a stated objective which implies a structured approach. iii This measure, in contrast to those that we rejected, may affect employees directly in terms of cost pressures and performance standards. 3.3 Employee evaluation of work experience As noted above, employees were asked to evaluate their work experience on 22 different aspects, using a five point scale. These measures are listed in Table 3. Forth et al. (2006) chose to analyse a summary measure. We did likewise, but also tested whether there were in fact three (or more) underlying factors. We found that three separate factors did indeed emerge, with each factor weighting heavily on one dimension of work experience: evaluations of the job (8 items), evaluations of information (4 items), and evaluations of manager-worker relationships (10 items). iv It is possible to add scores on each dimension to produce an overall continuous score, and some researchers do so, using OLS methods. The underlying measures are, 13 however, ordinal, and, along with other WERS researchers (e.g. Danford et al., 2007), we have chosen to retain the ordinal ranking and use an ordered response model (for review see Wooldridge, 2002, pp. 504-508). We also adopt a random effects estimator to allow for the fact that the employee level data are drawn from a number of workplaces. Since employees were not asked to evaluate their overall work experience, we computed average scores of the EEWE and then scaled them back to 0-4 categories. Doing this we found that only a very small proportion of replies were in the top (5.24%) and bottom (0.10%) categories. We, therefore, used a three-fold scale taking the value of 1 if the employee is dissatisfied (15.2%), 2 if satisfied (39.5%) and 3 if very satisfied (45.3%). In the same way, we created a three point index to reflect each of the three dimensions of the EEWE. v We also allowed for the possibility that the ‘very satisfied’ are qualitatively different from other workers, and for some analyses used a simple contrast between these two groups. 4 Formality and work experience by workplace size We first address Hypothesis 1 by examining the association between formality and size within workplaces. Table 2 shows the responses to the twelve questions on formality, together with the average for all twelve. So, for example, the right hand side of the table shows that 7% of small single-workplace firms have ‘a person mainly concerned with HR issues’. This proportion rises monotonically to 80% for single-workplace firms with more than 500 workers. For both multi- and single-workplace firms, it is the smallest sized workplaces which, on all twelve measures - and by implication also the average measure - that have the lowest formality. Second, formality rises with increases in workplace size for almost every measure. To address Hypothesis 2 we initially compare small workplaces that are part of a multi-site firm with small single-site workplaces. It shows that every measure of formality is significantly higher in 5-49 employee workplaces that are owned by a multisite firm than for single-workplace firms of this size. For example, in workplaces of 5-9 employees that are part of a multi-site firm, 29% report having an HR manager, compared with only 7% in single–workplace firms. These differences between multi and single workplaces diminish with increasing workplace size, however, such that there are 14 no significant differences (at the 5% level) between any workplaces with more than 500 employees. However, not all the SMMWs belong to a large firm. To address this Table 2 distinguishes between SMMWs that are part of a SME and those that are part of a large firm. It shows that formality is significantly greater in the latter than in the former, suggesting that the size of the firm, as well as workplace, matters. Second, it shows that SMMWs have greater formality than small and medium single-site workplaces (SMSWs). vi Table A1 in the Appendix provides more information and further evidence in support of Hypothesis 2. Hypothesis 3 addresses the link between workplace size and the EEWE. The findings are stark. Table 3 shows that workers in all small workplaces were more likely to have higher average EEWE than those in large workplaces. For the multi-site workplaces, the EEWE decreases significantly with increasing size until the 250-499 employee range. After that increases in workplace size seem unrelated to the EEWE. This effect, however, is not apparent in single-site workplaces, where the EEWE continues to decrease as workplace size increases. The finding that there is a significant difference in the means of the EEWE between those in small workplaces that are part of a multi-site firm and those in small single-workplace firms also provides support for Hypothesis 4. In particular, on 21 out of the 22 measures, workers in single workplace firms of 5-49 employees gave a higher evaluation than those in workplaces of the same size but owned by a multi-site firm. In contrast, for workplaces of more than 500 workers, there were significant differences at the 5% level on only six dimensions. Turning to SMMWs, two main findings are apparent: first, the means between SMMWs that are part of SME and those that are part of a larger firm were not statistically significant from each other on half of the EEWE measures. Second, the EEWE was significantly higher in the single-site SMEs than in SMMWs that are part either of a SME or a large firm, providing further support for Hypothesis 4 (Table A2 in the Appendix provides a more detailed picture). 15 Table 2: Employer Formality (0-1 with 1: Yes and 0: No; weighted estimates %) A) Workplaces that are part of a multi-site firm B) Single-workplace firms 5-49 50-99 100-249 250-499 500+ SMMW (5-249) Large SME firm 5-49 50-99 100-249 250-499 500+ SME Person mainly concerned with HR issues 28.85 38.70 61.23 69.22 81.73 26.50 44.78 6.69 14.46 42.18 52.39 80.07 11.92 Existence of a formal strategic plan 74.23 92.81 90.13 92.89 93.43 56.65 90.12 39.22 65.13 69.11 60.43 89.58 46.47 Investors in People (IiP) 43.59 50.71 48.04 54.36 48.95 20.27 53.47 12.12 22.19 15.08 27.18 43.08 13.94 Presence of tests at induction as part of recruitment 90.19 99.86 98.82 98.86 100 83.74 98.19 76.36 94.81 98.95 96.70 100 81.67 Any communication channels e.g. newsletters, internet , email 95.44 99.29 100 100 100 93.79 98.89 89.19 98.61 100 100 100 91.82 Any meeting between management and employee 92.00 97.18 95.57 92.41 98.1 90.43 95.56 80.84 85.43 93.26 97.38 90.26 82.94 Presence of a dispute procedure 42.73 46.97 54.10 67.42 71.23 36.43 50.06 22.48 34.48 47.48 37.44 58.02 27.13 Presence of an equal opportunity policy 81.46 92.66 94.03 97.56 95.37 61.33 94.70 39.77 70.34 90.79 89.58 96.88 50.14 Presence of a grievance policy 92.49 99.41 99.62 99.08 100 86.19 98.60 68.11 97.34 98.95 95.14 100 75.96 Presence of a performance appraisal programme 84.31 92.79 90.55 99.46 96.02 71.14 92.42 54.16 71.54 74.06 81.21 94.53 59.01 Formal target 87.64 95.77 92.75 98.49 94.12 77.86 94.69 62.01 92.60 91.41 95.10 93.73 69.85 Any non-payment benefits 83.77 89.62 98.91 97.46 98.55 79.56 92.59 65.56 70.60 73.73 65.61 88.46 67.24 Average formality 74.80 83.27 85.00 89.14 89.72 65.33 83.81 51.39 68.01 74.59 74.85 85.71 56.48 Workplace size Formality Measures: SMSW (5-249) Workplace size Notes: Within workplaces that are part of a multi-site firm and single-workplace firms, the null hypothesis that formality is the same between workplaces of different size was not rejected at the 5% level in the former sub-sample for "Investors in People". In most cases, formality increased significantly for workplaces above 5-49 employees. We test whether the proportions of workplaces that are part of a multi-site firm and single-workplace firms, with the same workplace size, are statistically different from each other. We also test whether the proportions of SMMWs that are part of a SME and those that are part of a large firm are statistically different from SMEs owned by a single-site firm. Bold black and underline values indicate where the difference was found to be statistically significant at the 5% and 10% level, respectively. 16 Table 3: Mean Employee Evaluation of Work Experience (EEWE) by type of workplace and organisation (weighted estimates) A) Workplaces which are part of a multi-site firm SMMW Workplace size (5-249) 550100250Large 49 99 249 499 500+ SME firm Job Satisfaction Measures: Factor 1 : EEWE regarding the nature of the job B) Single-workplace firms SMSW (5-249) Workplace size 549 5099 100249 250499 500+ SME 2.92 (0: very dissatisfied; 4: very satisfied) Sense of achievement 2.80 2.73 2.69 2.54 2.59 2.86 2.72 2.98 2.89 2.64 2.79 2.71 Scope for using initiative 2.85 2.82 2.77 2.61 2.66 2.93 2.80 3.04 2.93 2.72 2.98 2.71 2.98 The amount of influence on job 2.62 2.59 2.51 2.36 2.40 2.69 2.54 2.86 2.67 2.58 2.65 2.45 2.79 The training received 2.41 2.28 2.25 2.22 2.18 (2.34) (2.34) 2.52 2.38 2.27 2.44 2.23 2.47 The amount of pay received 1.83 1.78 1.89 1.74 1.81 1.97 1.81 2.14 2.00 1.91 1.94 1.92 2.09 The amount of job security 2.69 2.68 2.52 2.46 2.34 2.77 2.60 2.84 2.76 2.58 2.71 2.47 2.80 Satisfaction with the work itself 2.82 2.75 2.74 2.60 2.61 2.92 2.74 2.97 2.92 2.70 2.84 2.76 2.93 Amount of involvement in decision making at workplace 2.36 2.19 2.06 2.00 2.04 (2.29) (2.23) 2.60 2.30 2.22 2.26 1.97 2.51 Managers keeping employees informed about changes at work 2.49 2.38 2.21 2.11 2.20 (2.34) (2.41) 2.63 2.30 2.20 2.37 2.02 2.52 Managers keeping employees informed about staffing changes 2.45 2.23 2.04 1.94 2.04 (2.30) (2.29) 2.60 2.29 2.09 2.21 1.97 2.49 Managers keeping employees informed about changes in way job is done 2.57 2.40 2.23 2.16 2.18 (2.43) (2.44) 2.68 2.41 2.29 2.34 2.12 2.59 Managers keeping employees informed about financial matters 2.25 2.14 2.07 1.99 2.12 1.96 2.25 2.16 2.01 1.97 1.99 1.73 2.11 Managers seeking employee views 2.44 2.22 2.08 1.99 2.07 (2.28) (2.30) 2.62 2.31 2.06 2.10 1.88 2.50 Managers responding to employee suggestions 2.40 2.18 2.01 1.93 1.94 (2.25) (2.24) 2.61 2.26 2.08 2.10 1.85 2.48 Managers allowing employees/ employee reps to be involved in decision making 2.13 1.89 1.72 1.68 1.68 (1.97) (1.96) 2.34 1.89 1.80 1.89 1.55 2.20 Relationship between employee and management 2.88 2.65 2.48 2.36 2.38 2.84 2.69 3.14 2.74 2.59 2.53 2.34 3.00 Managers’ ability to keep promises 2.50 2.36 2.17 2.01 1.99 (2.41) (2.36) 2.74 2.37 2.30 2.22 2.05 2.63 Managers’ sincerity in attempting to understand employee needs 2.60 2.46 2.26 2.14 2.13 (2.50) (2.46) 2.82 2.50 2.39 2.39 2.13 2.71 Managers’ dealing with employees honestly 2.67 2.50 2.33 2.21 2.18 2.61 2.52 2.90 2.54 2.47 2.50 2.24 2.78 Managers’ understanding about having to meet responsibilities outside work 2.64 2.48 2.31 2.16 2.24 2.62 2.49 2.88 2.56 2.56 2.36 2.30 2.79 Managers’ encouraging employees to develop skills 2.63 2.50 2.38 2.28 2.33 (2.52) (2.54) 2.75 2.61 2.41 2.51 2.37 2.69 Managers’ treating employees fairly 2.63 2.47 2.29 2.18 2.21 2.62 2.47 2.87 2.62 2.47 2.52 2.17 2.78 Average score 2.52 2.36 2.26 2.15 2.18 (2.46) (2.40) 2.74 2.49 2.27 2.34 2.19 2.65 Factor 2: EEWE regarding information (0: very poor; 4 very good) Factor 3: EEWE regarding manager-worker experience (0: very poor; 4 very good) (0: strongly disagree; 4: strongly agree) Notes: We test whether the means of workplaces that are part of a multi-site firm and single-workplace firms, with the same workplace size, are statistically different from each other. We also test whether the means of SMMWs that are part of a SMEs and SMMWs that are part of a large firm are statistically different from single-site SMEs. Bold black and underline values indicate where the difference was found to be statistically significant at the 5% and 10% level, respectively. Parentheses show that the difference between SMMWs that are part of a SME and SMMWs that are part of a large organisation is statistically insignificant at the 5% or 10% level. 17 5 Determinants of employee evaluations of their work experience 5.1 The role of formality In order to complete our analysis, we now examine the association between the EEWE and management formality, taking into consideration a wide range of workforce and establishment characteristics. vii Table 4 presents the ordered probit results for the overall EEWE index, and also for the three disaggregated indexes (Model I). For comparison, we also include the results from the random effect ordered probit (Model II). We initially focus our discussion on the former. The results confirm our earlier finding that employees in single-workplace firms are more likely to evaluate their job experience highly. The coefficients for the workplace size measures also show strongly that the EEWE decreases as size increases. viii As for the formality indices, we are aware of the possible limitations of treating these as exogenous in the equation. One may argue, for example, that recruiting the wrong people and employees’ sense of grievance may generate job dissatisfaction and in response managers may introduce formal policy responses. Though this is an issue for further research, we would argue that formality is likely to reflect genuinely exogenous processes. It is well-established that the formalisation of discipline and dismissals procedures was driven by legal changes starting in the 1970s (Edwards, 1989). Similar points apply to equal opportunities policies. A second driver of formalisation is the general one discussed above, namely, the need for formal rules as organisations become more complex. A third is isomorphism (DiMaggio and Powell, 1983): firms grow alike because of coercive, normative and mimetic processes. This is not to deny that some formalisation may be driven by the state of employee opinion. At least among small firms, however, the evidence for growing formalisation in the handling of discipline and dismissals is weak (Earnshaw et al. 1998). We might attribute this to weaker isomorphic pressures than exist among larger firms. In short, we doubt whether endogeneity is a big problem. The results based on the overall EEWE measure are shown in the first column of Table 4. ix They show that ‘having a person mainly concerned with HR issues’, ‘communication channels’, ‘presence of dispute procedures’, and ‘any non payment benefits’ are negatively associated with the overall EEWE. We suggest that most of these 18 types of formal procedures may emphasise regulating the workplace through bureaucratic or documented means, and therefore, may impinge on employee autonomy or discretion and reduce the EEWE. Table 4, however, also shows some significantly positive relationships between formality and EEWE. ‘Investors in People’, ‘meetings between management and employee’ and the ‘presence of a performance appraisal programme’ were found to carry a positive and significant sign. One possible explanation is that these formal procedures are not purely rule-bound but allow employee concerns to be taken into account; IiP for example requires evidence of systematic training. When the model is estimated separately for the three disaggregated indexes of EEWE, three main conclusions emerge. First, formal procedures such as ‘having a person mainly concerned with HR issues’ and ‘presence of grievance policy’ are found not to have any significant association with evaluations of the job or the information provided by managers. However, both were found to be negatively and significantly associated with evaluations of the relationship between manager and employee. Second, the ‘presence of an equal opportunity policy’ and ‘formal targets’ were found to be positively associated with evaluations of information provided by managers. Finally, the role of formalisation seems to be less important in explaining attitudes to the job itself, since only one formality measure was found to be significant at the 5% level. This pattern complements our argument, for evaluations of the job itself are likely to reflect factors such as interest in the task itself and the quality of relations with work mates. We might expect formality to be less salient here than it is with views of managers - a result consistent with the work of Tsai et al. (2007). By comparing the results of the ordered probit analysis with the random effects ordered probit, we find: first, ‘having a person mainly concerned with HR issues’, ‘any communication channels’ and ‘any non payment benefits’ significantly lowers EEWE. However, ‘any meeting between management and employee’ is clearly associated with better evaluations. The role of other formal procedures such as ‘Investors in People’, ‘presence of dispute procedure’ and ‘presence of a performance appraisal programme’ becomes weak. Finally, the coefficient of rho (ρ), which indicates the magnitude of the workplace effects, is statistically significant and relatively large throughout the 19 estimations. It is clear therefore, that there is a high unobservable intra-workplace correlation in the determinants of the EEWE. Overall, the results suggest that the relationship between EEWE and management formality in workplaces is complex. We experimented further by aggregating the raw responses on formality to find an overall formality variable that ranges between 1 and 12. Since this variable appears in the right-hand side of the statistical model, we treated it, for simplicity, as continuous and found that its coefficient for the overall EEWE model was statistically insignificant (coeff.= -0.008 and std. err.= .0008). For the three disaggregated indices of job satisfaction, however, overall formality was found to have a significant and positive association with evaluations of information provided by managers (coeff.=0.023 and std. err.=0.007) but a significant and negative association with evaluation of the relationship between manager and employee model (coeff.=-0.019 and std. err.=0.007). x 20 Table 4: Ordered Probit (I) & Ordered Probit Random Effects (II) Estimates of Employee Evaluation of Work Experience Overall Measures of EEWE: Model: Variable Workplace size (5-49) 50-99 Nature of the job Information provided by managers Relationship between manager and employee I Coeff. II Coeff. I Coeff. II Coeff. I Coeff. II Coeff. I Coeff. II Coeff. -0.178** -0.207** -0.082** -0.082* -0.0981** -0.127** -0.170** -0.200** 0.039 0.060 0.036 0.048 0.035 0.058 0.037 0.059 100-249 -0.226** -0.271** -0.135** -0.146** -0.207** -0.262** -0.227** -0.270** 0.038 0.059 0.035 0.047 0.034 0.058 0.036 0.059 250-499 -0.259** -0.322** -0.145** -0.159** -0.244** -0.297** -0.311** -0.375** 0.045 0.071 0.041 0.057 0.040 0.069 0.043 0.070 500+ -0.259** -0.288** -0.243** -0.254** -0.158** -0.189** -0.278** -0.311** 0.043 0.068 0.040 0.055 0.039 0.067 0.042 0.068 0.127** 0.172** 0.141** 0.158** 0.076** 0.095* 0.130** 0.178** 0.034 0.051 0.031 0.041 0.030 0.049 0.032 0.050 -0.089** -0.080* -0.016 -0.008 -0.019 -0.009 -0.085** -0.082* 0.028 0.045 0.026 0.036 0.026 0.044 0.027 0.044 0.001 0.008 0.015 0.017 0.011 0.012 -0.050 -0.049 Establishment (W/pls that are part of a multi-site firm) Single workplace firm Formality Person mainly concerned with HR issues Existence of a formal strategic plan Investors in People (IiP) Presence of tests at induction as part of recruitment 0.038 0.057 0.034 0.046 0.034 0.055 0.036 0.057 0.053** 0.062 0.043* 0.041 0.076** 0.076* 0.031 0.039 0.026 0.041 0.024 0.033 0.024 0.040 0.025 0.041 -0.043 -0.052 0.001 -0.015 -0.060 -0.041 -0.052 -0.046 0.064 0.090 0.058 0.074 0.057 0.086 0.061 0.088 Any communication channels e.g. email -0.484** -0.513** -0.533** -0.572** -0.615** -0.661** -0.376** -0.403** 0.131 0.165 0.117 0.137 0.113 0.149 0.122 0.158 Any meeting between management and employee 0.122** 0.134* 0.021 0.012 0.233** 0.229** 0.151** 0.149** 0.051 0.076 0.046 0.061 0.045 0.072 0.048 0.074 Presence of a dispute procedure -0.078** -0.066 -0.040 -0.031 -0.075** -0.060 -0.055 -0.042 Presence of an equal opportunity policy 0.028 0.042 0.025 0.034 0.025 0.041 0.026 0.042 0.028 0.018 -0.004 -0.020 0.067* 0.061 0.035 0.025 21 0.042 0.063 0.038 0.051 0.037 0.061 0.040 0.063 Presence of a grievance policy -0.069 -0.098 0.018 0.001 0.028 0.008 -0.133* -0.140 0.073 0.102 0.065 0.082 0.063 0.095 0.068 0.099 Presence of a performance appraisal programme 0.074* 0.054 0.027 0.024 0.109** 0.079 0.134** 0.118** 0.044 0.066 0.040 0.053 0.040 0.064 0.042 0.065 0.034 0.046 -0.051 -0.047 0.072* 0.072 0.002 -0.003 0.046 0.068 0.042 0.055 0.041 0.065 0.043 0.067 -0.102** -0.118* -0.072* -0.074 -0.101** -0.102 -0.070 -0.092 Formal target Any non payment benefits Log likelihood Chi2 [degrees of freedom] Rho 0.049 0.069 0.044 0.056 0.043 -8,964.0 1,081.4[50] - -8,807.9 714.3[50] 0.144 -10,352.7 1,078.5[50] - -10,271.7 807.9[50] 0.079 -11,734.1 884.8[50] - 0.012 Observations 0.065 -11,474.1 533.4[50] 0.157 0.009 9,370 11,507 0.046 0.067 -10,154.3 1,273.4[50] - -9,958.1 760.1[50] 0.153 0.011 11,213 0.012 10,226 Note: All models control for personal and workplace characteristics. Full tables are available from authors on request. Standard errors are reported beneath parameter estimates (in italic). **Significant at the 5% level. *Significant at the 10% level. ` 22 We also estimated the overall EEWE model separately for workplaces which are a part of a multi-site firm and single-workplace firms. Additionally, for each of these two sub-samples the model is estimated separately for SMEs (5-429); and large firms (250+). For simplicity, we allow for possible distinct differences between those who feel very content and other workers, and use a binary variable taking the value of unity if the worker reports being very satisfied with his or her job and zero otherwise. Probit estimates (reported as marginal effects) of the probability of being very satisfied with job aspects are given in Table 5. xi We first examine workplaces that are part of a multi-site firm. The results (shown in column 1) clearly show that employees in SMMWs that are part of SME are more likely to be ‘very satisfied’ than those in workplaces that are part of a large multi-site firm. Having ‘a person mainly concerned with HR issues’, ‘existence of a formal strategic plan’ and ‘presence of a dispute procedure’ all lower the probability of high EEWE in large multi-site firms. The coefficient of ‘communication channels’ was found to be negative and statistically significant in both SMMWs that are part of a SME and workplaces that are part of a large multi-site firm. Employees, for example, may feel that communication should be more personal and direct, or it may be the case that employees simply dislike newsletters. In contrast, having a ‘formal target’ was associated with high level of EEWE. Not surprisingly, having ‘meetings between management and employees’ increased the probability of high EEWE in SMMWs that are part of a SME. Also, the formality variable that captures ‘any non-payment benefits’, was found to be negative and statistically significant. Finally, the coefficient of ‘presence of a grievance policy’ was found to be positive and statistically significant for the large firm sub-sample. The results for the single-workplace firms are also very revealing. We first notice that the coefficient of SMEs is insignificant. That is, the size of the firm does not make a difference as between single-workplace firms. However, a closer examination showed that SMEs and having ‘a person mainly concerned with HR issues’ have a high degree of negative correlation (corr.=-0.47). We re-estimate the model by including interactions between the two dummies. Only the interaction between SMEs and not having ‘a person mainly concerned with HR issues’ was found to increase the probability of reporting high EEWE when compared to large size firm that has ‘a person mainly concerned with HR 23 issues’. Our interpretation of this result is that HR ‘professionals’ seeks to change informal procedures that the workforce may have felt worked well in the past. It may also be influenced by the prior experience of the HR professional, so that someone coming into a SMEs and applying ‘large firm procedures’ is likely to encounter problems. It is important to note that, overall, the formality measures point in a single direction in the single-site, SMEs model. xii Here, the three formality variables that were found to be statistically significant all carry a negative sign. This is not the case for the large, single-site firms where the direction was mixed. We also find that some formality variables that are associated with the EEWE in large multi-site sub-sample are also found to be associated with the EEWE in large single-site firms, but in a different direction. For example, employees working in a large single-site firm that has a ‘formal strategic plan’ are more likely to be ‘very satisfied’ than those who work in a large single-site firm with no ‘formal strategic plan’. The opposite is found in multi-site large firm sub-sample. Perhaps in single, large workplaces this may proxy for the management having a clear vision for the firm and the ability to ensure this is shared by the workforce. 24 Table 5: Probit Estimates of Overall EEWE Sample: Workplaces that are part of a multi-site firm Single-workplace firms Model: I II III I II III IV1 Organisation size: Overall SMEs Large firms Overall Overall SMEs Large firms ME ME ME ME ME ME 0.059** - - 0.049 - - - 0.004 - - - - - - -0.116** 0.078 Variable Firm size (Large firms) SMEs 0.022 Firm size & HR (Large firm with a person mainly concerned with HR issues) SMEs with a person mainly concerned with HR issues 0.038 - - - - 0.053 SMEs without a person mainly concerned with HR issues - - - - 0.115** 0.044 Large firms without a person mainly concerned with HR issues - - - - 0.028 0.054 Formality Person mainly concerned with HR issues Existence of a formal strategic plan -0.042** -0.039 -0.041** -0.082** 0.014 0.051 0.014 0.033 -0.077** 0.000 -0.095** 0.055** 0.055** 0.040 0.078 0.031 0.291** 0.023 0.063 0.027 0.028 0.028 0.030 0.101 Investors in People (IiP) 0.026* 0.071 0.017 0.017 0.014 0.030 0.011 0.013 0.051 0.014 0.029 0.029 0.034 0.093 Presence of tests at induction as part of recruitment -0.022 -0.029 -0.095 -0.030 -0.028 -0.049 0.032 Any communication channels e.g. newsletters, internet , email Any meeting between management and employee 0.040 0.061 0.062 0.042 0.042 0.042 0.403 -0.337** -0.301** -0.333** -0.084 -0.084 -0.073 - 0.073 0.092 0.107 0.079 0.079 0.079 0.037 0.126* 0.004 0.013 0.017 0.068 25 -0.184 Presence of a dispute procedure Presence of an equal opportunity policy Presence of a grievance policy Presence of a performance appraisal programme Formal target 0.029 0.075 0.033 0.042 0.043 0.047 0.162 -0.050** -0.027 -0.051** -0.007 -0.007 -0.005 -0.017 0.014 0.045 0.015 0.026 0.026 0.029 0.090 0.031 0.056 0.032 -0.025 -0.020 -0.013 -0.032 0.025 0.053 0.032 0.029 0.030 0.031 0.142 0.123** 0.024 0.232** -0.089** -0.087** -0.094** -0.414 0.045 0.079 0.063 0.043 0.043 0.042 0.359 0.033 0.070 0.036 -0.012 -0.016 -0.013 -0.335** 0.026 0.060 0.031 0.033 0.033 0.035 0.134 0.078** 0.105* 0.055* -0.054* -0.052 -0.035 -0.472** 0.026 0.057 0.032 0.033 0.033 0.035 0.106 -0.037 -0.114* -0.002 -0.074** -0.078** -0.061** -0.440** 0.033 0.063 0.041 0.031 0.031 0.032 0.139 Formality dummies (p-value) p=0.007 p=0.007 p=0.000 p=0.001 p=0.006 p=0.001 Interaction dummies (p-value) - - - - p=0.006 - - -4,319.20 585.6[46] 0.42 6,753 -537.3 132.8[44] 0.52 872 -3,738.8 504.7[45] 0.41 5,881 -1,422.4 311.6[46] 0.54 2,285 -1421.6 313.1[47] 0.54 2,285 -1,143.6 245.2[45] 0.58 1,851 -242.7 98.27(41) 0.39 434 Any non payment benefits Log likelihood Chi2 [degrees of freedom] Predicted probability Observations Notes: All models control for personal and workplace characteristics. Full tables are available from authors on request. Standards errors are reported beneath parameter estimates (in italic). SMEs in multi-site firms include small and medium sized workplaces that belong to small and medium sized firms. Large firms in multi-site firms include small, medium sized and large workplaces that belong to large firms. 1 The variable "Any communication channels e.g. newsletters, internet , email" dropped due to colinearity. **Significant at the 5% level. *Significant at the 10% level. 26 p=0.003 6 Conclusions This paper has examined the link between management formality and employee evaluation of work experience in relation to size and ownership. These evaluations have been defined, not in terms of ‘satisfaction’, but as perceptions of certain dimensions of jobs. Overall the findings showed that size and ownership (whether workplaces are single-site or multi-site) are key predictors of formality and EEWE. In relation to size, four key findings emerge. First, employee evaluations of jobs are rated as higher in small, than in large, workplaces. Second, formality is greater in large than in small workplaces. Third, perceptions of jobs are better in SME single-site workplaces in comparison to similar size multi-site workplaces belonging to either an SME or a large firm. Fourth, single-site SMEs have lower formality than small and medium sized multi-workplaces that are part of either an SME or a large firm. There are two conclusions in relation to formality and employees’ experience of their jobs. First, different aspects of formality have different association with EEWE, and the character of the association differs between types of workplace ownership. Table 5, for example, shows contrasting results of some indices of formality between different types of workplace ownership and firm size. This result suggests that formality is not simply bad for employees and that much depends on context. There is no straightforward explanation for these results within a survey method, and the meaning of different kinds of formality needs to be grasped through more intensive methods. Nevertheless, in singlesite SMEs the negative association between formality and EEWE tend to predominate, as Table 5 underlines. This result is consistent with our expectations as to the impact of formality within such firms. The fact that size effects persist is consistent with the evidence of Kalleberg and Van Buren (1996) and Tsai et al. (2007), that there are features of workplace size that operate over and above such correlates of size as formality. The theoretical arguments for such effects reviewed in Section 2 are thus supported. One reason may be that small-firm informality is not perfectly captured in surveys such as WERS. For example, Taylor (2005) reports from four small firms that, although managers claimed to use formal recruitment and selection methods, employees stressed ‘being known’ and kinship links. In other words, over and above low levels of measured formality in small firms, there is 27 informality of practice. Now, as noted above, informality also occurs in large organisations, and it is an open question whether such informality-as-practice is any more common in small firms. But there are many reasons to expect that it is. In large firms, the rules determine the overall framework, and informality occurs in the interstices of the rules, as all the research on custom and practice demonstrated (Edwards, 1988). In small firms, managers tend to use personal knowledge and their own judgements, with formality being in the background (Edwards et al., 2004; Marlow, 2005). There thus remain ‘pure’ size effects on employee evaluations of work wherein small firms and small workplaces enjoy relatively good evaluations from their employees. Our findings in relation to formality are, as noted, more ambiguous. Some indicators of formality are associated with better EEWEs and others have the reverse outcome. Hence, we would not conclude that all formality is to be avoided in SMEs. Extreme informality can be a cover for autocracy (Rainnie, 1989). And there are instances where formalisation was consistent with the direction of the firm. Ram et al. (2001) for example, discuss a food manufacturing firm which had grown in size, and rationalised and modernised its production process. The result in HR terms was, as a manager memorably put it, a shift from being ‘very laid back’ to being ‘laid back’. There were now formal procedures for discipline, for example, and these were used, but the shopfloor atmosphere remained personal and workplace relations were generally good. In this case, therefore, a ‘fit’ was achieved. When firms grow, they are likely to need to formalise to a degree, but they need to address the way in which they do so, in particular seeking to minimise negative effects on employee attitudes. Nevertheless our finding that, in single-site SMEs there is a negative association between formality and EEWE perhaps imply that the case described by Ram et al can be viewed as an exemplar. We now turn to the possible public policy implications from our findings, particularly from a smaller firm perspective. Our view is that the pressure from government is to encourage smaller firms to embrace formality, perhaps on the (misguided) principle that this characterises HRM in large firms and so, therefore, it must be appropriate for SMEs which are viewed as less well managed. Three examples of this will suffice. First, funding from the Learning and Skills Council (LSC) is devoted to formal training, with informal training remaining broadly unsupported. This clearly 28 signals that government favours formality and that if small firms wish to receive support from public funds then they have to shift towards greater formality. Ideally we would like to have a measure of formal and informal training in WERS 2004 to examine its link with EEWE and other “performance” measures, but such data were not collected. Given the other associations identified in this paper our expectation is that a shift towards more formality would be unlikely to lead to an increase in EEWE amongst small firm employees. Second, the Department for Innovation, Universities and Skills continues to focus its attention on formal qualifications, to which small firms give a much lower priority than more subjective personal qualities such as ‘attitude’, ‘commitment’, and ‘reliability’. Whilst the more recent Leitch Review (2006) recognises the importance of ‘on-the job’ training it emphasises ‘focusing as much training as possible on qualifications, as long as qualifications reflect economically valuable skills, brings enormous benefits to both individuals and employers’ (Leitch Review, 2006: 50). Although there are recent initiatives aimed at delivering training flexibly, with smaller firms in mind, such as through ‘Train to Gain’, the emphasis remains on formality and qualifications (Leitch Review, 2006: 24) and hence continues the long-standing link between state sponsored training and formality (Kitching and Blackburn, 2002). Third, other evidence from examining the outcome from Employment Tribunals shows that small firms are more likely to win their cases if they have, but more importantly actually use, formal disciplinary procedures (Saridakis et al., 2006). If, as these examples illustrate, public policy seeks to encourage greater formality in small workplaces owned by single site SMEs then the impact upon EEWE is likely to be negative. 29 References: Acton Society Trust (1953), Size and Morale: A Preliminary Study of Attendance at Work in Large and Small Units (London, Acton Society Trust). Atkinson J. and Meager, N. (1994), ‘Running to Stand Still: The Small Firm in the Labour Market’, in J. Atkinson and D.J. Storey (eds.), Employment, the Small Firm and the Labour Market (London, Routledge). Bolton Committee (1971) Report of the Committee of Inquiry on Small Firms, Cmnd 4811 (London, HMSO). Brown, S., McNabb, R. and Taylor, K. (2006). Firm performance, Worker Commitment and Loyalty, Sheffield Economic Research Paper Series, 2006005. Chaplin, J., Mangla, J., Purdon, S. and Airey, C. (2005) The Workplace Employment Relations Survey 2004 (WERS 2004) Technical Report (Cross-section and Panel Surveys) (London, National Centre for Social Research). Curran, J and Stanworth, J (1981) ‘A New Look at Job Satisfaction in the Small Firm’, Human Relations, 34, 343-65. DiMaggio, P and Powell, W. (1983) ‘The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields’, American Sociological Review, 48, 147-60. Earnshaw, J, Goodman, J., Harrison, R., Marchington, M. (1998) Industrial Tribunals, Workplace Disciplinary Procedures and Employment Practic. Employment Relations Research Series 2 (London,Department of Trade and Industry). Edwards, P. (1988) ‘Patterns of Conflict and Accommodation’, in D. Gallie (ed.), Employment in Britain (Oxford, Blackwell). Edwards, P. (1989) ‘The Three Faces of Discipline’, in K. Sisson (ed.), Personnel Management in Britain (Oxford, Blackwell). Edwards, P., Rees, C. and Collinson, M. (1998) ‘The Determinants of Employee Responses to Total Quality Management: Six Case Studies’, Organization Studies, 19, 449-75. 30 Edwards, P. and Ram, M. (2006) ‘Surviving on the Margins of the Economy: Working Relationships in small, Low-wage Firms’, Journal of Management Studies, 43, 895-916. Edwards, P., Ram, M. and Black, J. (2004) ‘Why Does Employment Legislation not Damage Small Firms?’ Journal of Law and Society, 31, 245-65. Forth J., Bewley, H. and Bryson, A. (2006) ‘Small and Medium-sized Enterprises: Findings from the 2004 Workplace Employment Relations Survey’, London: Department of Trade and Industry. Gallie, D. (2003) ‘The Quality of Working Life: Is Scandinavia Different? European Sociological Review, 19, 61-79. Gallie, D., Felstead, A., and Green, F. (2004) ‘Changing Patterns of Task Discretion in Britain’, Work, Employment and Society, 18, 243-66. Golhar, D. and Deshpande, S. (1997) ‘HRM Practices of Large and Small Canadian Manufacturing Firms’, Journal of Small Business Management, 35, 30-38. Green, F. (2006) Demanding Work: The Paradox of Job Quality in the Affluent Economy (Princeton, Princeton University Press). Hill, S. (1974) ‘Norms, Groups and Power’, British Journal of Industrial Relations, 12, 213-35. Holliday, R. (1995) Investigating Small Firms (London. Routledge). Hoque, K. (2003) ‘All in All, it's Just Another Plaque on the Wall: The Incidence and Impact of the Investors in People Standard’, Journal of Management Studies, 40, 543–71. Idson, T. (1990) ‘Establishment Size, Job Satisfaction and the Structure of Work’, Applied Economics, 11, 606–28. Ingham, G. K. (1970) Size of Industrial Organisation and Worker Behaviour (Cambridge, CUP). Kaiser, H. F. (1970) ‘A Second Generation Little Jiffy’, Psychometrika, 35, 401-15. Kalleberg, A.L. (1977) ‘Work values and Job Rewards: A Theory of Job Satisfaction’, American Sociological Review, 42, 124-43. 31 Kalleberg, A. L. and Van Buren, M. E. (1996) ‘Is Bigger Better? Explaining the Relationship between Organization Size and Job Rewards’, American Sociological Review, 61, 47-66. Kaman, V., McCarthy, A.M., Gulbro, R.D., Tucker, M. (2001) ‘Bureaucratic and High Commitment Human Resource Practices in Small Service Firms’, Human Resource Planning, 24, 33-9. Kersley, B., Alpin, C., Forth, J., Bryson, A., Bewley, H., Dix, G. and Oxenbridge, S. (2006) Inside the Workplace: Findings from the 2004 Workplace Employee Relations Survey (Abingdon, Routledge). Kitching, J. and Blackburn, R. A. (2002) The Nature of Training and Motivation to Train in Small Firms. DfES Research Report RR 330 (Nottingham, DfES). Kotey, B. and Slade, P. (2005) ‘Formal Human Resource Management Practices in Growing Small Firms’, Journal of Small Business Management, 43, 16-40. Leitch Review (2006) Leitch Review of Skills: Prosperity for all in the Global Economy World Class Skills, Final Report, HM Treasury, December. (London, HMSO). Marginson, P. (1984) ‘The Distinctive Effects of Plant and Company Size on Workplace Industrial Relations’, British Journal of Industrial Relations, 22, 1-14. Marlow, S. (2005) ‘Introduction’, in S. Marlow, D. Patton and M. Ram (eds), Managing Labour in Small firms. Routledge, Abingdon. Mata, J. and Portugal, P. (2002) ‘The Survival of New Domestic and Foreign Owned Firms’, Strategic Management Journal, 23, 323-43. Millward, N. (2001). ‘Research note: the representativeness of WERS 98- a clarification’, Industrial Relations Journal, 32/4: 344-346. Purdon, S. and Pickering, K. (2001). The use of sampling weights in the analysis of the 1998 Workplace Employee Relations Survey, National Centre for Social Research. Ram, M., Edwards, P., Gilman, M. and Arrowsmith, J. (2001) ‘The Dynamics of Informality: Employment Relations in Small Firms and the Effects of Regulatory Change’, Work, Employment and Society, 15, 845-61. Ram, M., Edwards, P. and Jones, T. (2007) ‘Staying Underground: Informal Work, Small Firms, and Employment Regulation in the United Kingdom’, Work and Occupations, 34, 318-44. 32 Rainnie, A. (1989) Industrial Relations in Small Firms (London, Routledge). Ramsay, H., Scholarios, D. and Harley, B. (2000) Employees and High Performance Work Systems: Testing Inside the Black Box. British Journal of Industrial Relations, 38, (4), pp. 501-531. Saridakis, G., Sen-Gupta, S., Edwards, P. and Storey, D. J. (2006) ‘Employment Tribunal Cases: The Impact of Enterprise Size on Incidence and Outcomes’, CSME Working Paper, 89, Warwick Business School. Sen-Gupta, S, Whitfield, K & McNabb, B. (2007) ‘Employee Share Ownership and Performance: Golden Path or Golden Handcuffs?, International Journal of Human Resource Management, 18( 8): 1507-1538. Storey, D. J. and Sykes, N. S. (1996) ‘Uncertainty, Innovation and Management’. In P. Burns and J. Dewhurst (eds), Small Business and Entrepreneurship (London. Palgrave). Taylor, S. (2005) ‘The Hunting of the Snark’, in S. Marlow, D. Patton and M. Ram (eds.), Managing Labour in Small Firms (Abingdon, Routledge). Tsai, C-J., Sen-Gupta, S. and Edwards, P. (2007) ‘When and Why Is Small Beautiful? The Experience of Work in the Small Firm’, Human Relations, 60, 1779-1808. Watson, J. and Everett, J.E. (1998) ‘Small Business Failure and External Risk Factors’, Small Business Economics, 11, 371-90. Winship, Christopher, and Larry Radbill. (1994) ‘Sampling Weights and Regression Analysis’, Sociological Methods and Research 23(2):230-257. Wooldridge, J.M. (2002) Econometric Analysis of Cross Section and Panel Data, Cambridge, Mass.: MIT Press. 33 Appendix Table A1: Average formality by workplace and firm size (0-1 with 1: Yes and 0: No; weighted estimates %) Ho: proportions Workplace size Firm size Ownership are equal 5-49 50-99 100-249 250-499 500+ (p-value) Multi-workplace (51.64) 5-49 Single workplace 51.39 Multi-workplace 64.28 76.81 p=0.025 50-99 Single workplace 68.01 Multi-workplace 72.23 80.85 (76.00) p=0.243 100-249 Single workplace 74.59 Multi-workplace 71.61 (74.42) 84.15 (77.49) p=0.003 250-499 Single workplace 74.85 Multi-workplace 82.72 85.81 86.51 90.44 (89.91) p=0.000 500+ Single workplace 85.71 p=0.000 p=0.000 p=0.008 p=0.023 Ho: proportions Multi-workplace are equal Single workplace (p-value) Note: Within each workplace size, we test whether the proportions of single-site workplaces and multi-site workplaces, with the same or different firm size, are statistically different from each other. Parentheses show that the difference is statistically insignificant at the 5% level. 34 Table A2: Average EEWE by workplace and firm size (0: very dissatisfied; 4: very satisfied; weighted estimates) Ho: means Workplace size Firm size Ownership are equal 5-49 50-99 100-249 250-499 500+ (p-value) Multi-workplace 2.55 5-49 Single workplace 2.74 Multi-workplace 2.36 (2.55) p=0.112 50-99 Single workplace 2.49 Multi-workplace 2.50 (2.44) (2.23) p=0.008 100-249 Single workplace 2.27 Multi-workplace 2.17 2.17 (2.26) 2.08 p=0.322 250-499 Single workplace 2.34 Multi-workplace 2.56 2.36 (2.27) 2.15 p=0.000 (2.17) 500+ Single workplace 2.19 Multi-workplace p=0.000 p=0.001 p=0.852 p=0.417 Ho: means are equal (p-value) Single workplace Note: Within each workplace size, we test whether the means of single-site workplaces and multi-site workplaces, with the same or different firm size, are statistically different from each other. Parentheses show that the difference is statistically insignificant at the 5% level. 35 Table A3: Description of the Individual and Establishment Characteristics (N=9,370)1 Individual Characteristics Male Age (Less than 21) Age (22-49) Age (50 and over) White Single No child No qualification Job Contract (Permanent) Temporary Fixed Tenure (More than 10 years) Less than 1 year 1-2 years 2-5 years 5-10 years Occupational Status (Non managerial duties) Manager or supervisor Occupational Tasks (Computer is needed) No need to use computer Ln(Income) 1 Mean (%) Association 54.19 pos. 69.09 24.36 95.08 22.59 83.08 15.73 neg. insign. insign. insign. insign. pos. 3.38 2.22 pos. insign. 14.83 12.26 28.21 19.69 pos. pos. insign. insign. 38.37 pos. 23.48 neg. 5.74 pos. Ln(Working hours) Member of Trade Union (Never or not recently) Recent Member of Trade Union Establishment Characteristics Ln(Establishment age) Gender diversity (uneven) Equally by men and women % of Employees working part time Training (0 or less than 1 day over the past 12 months) 1 to less than 5 days) 5 days or more Sector (Public administration and education) Manufacturing Electricity, gas and water Construction Wholesale and retail Hotels and restaurants Transport and communication Financial services Other business services Health Other community services The figures that are reported are based on the overall EEWE model. 36 Mean (%) 3.58 Association neg. 24.88 neg. 3.21 insign. 36.69 20.44 pos. pos. 59.23 26.10 insign. pos. 21.61 2.66 6.49 14.24 3.67 7.45 8.08 16.56 8.55 6.30 neg. insign. pos. insign. insign. insign. insign. insign. pos. insign. Footnotes i Winship and Radbill (1994) stated that “Although sampling weights must generally be used to derive unbiased estimates of univariate population characteristics, the decision about their use in regression analysis is more complicated” (pp 230). Millward (2001) noticed that most of the published statistical analyses of WERS 98 uses weighting in descriptive analysis, but do not use weighting in multivariate analyses (some examples are Ramsay et al., 2000 and more recently Sen-Gupta et.al. 2007). Purdon and Pickering (2001) recommended that WERS researchers who use unweighted data in the regression model to run also a parallel weighted model and compare the coefficients. The authors suggested that if the coefficients do not differ by more than a small amount between the models then the unweighted data can be used, and advantage taken of the smaller standard errors. For this reason we also estimate our ordered probit and probit models, which we present later in the paper, using weighted data. We found that the changes in the magnitude of the coefficients were moderate. As it was expected weighted regression did generally inflate standards errors, but in most cases the coefficients did remain statistical significant at the 5% or 10% level. Winship and Radbill (1994) argue that if the parameters estimates are similar, then unweighted estimates are preferable because they are more efficient. Therefore, based on both Purdon and Pickering (2001) and Winship and Radbill (1994) our paper presents only the unweighted parameter estimates. It should be further noted that there is a lack of appropriate routines to allow weights for random effects ordered probit models. RE analysis, however, is important because allows for the fact that the employee level data are drawn from a number of workplaces. Therefore, following the recent work by Brown et al. (2006) our paper does considers a random effect estimator too. Since we found that the differences in the coefficient of the unweighted and weighted order probit models were moderate, we may argue that perhaps this will be the case when parameter estimates from the unweighted random effects ordered probit and weighted random effects ordered probit are compared. ii McNabb and Whitfield (2000) make a similar distinction in their examination of low pay. iii These measures are not the only possible ones. The most obvious other candidates include the use of standard induction schemes for new employees and the presence of job evaluation schemes. We already had measures related to recruitment and to appraisal, however, and did not wish to overburden the analysis. In addition, including these other measures might bias the index of formality towards a particular element. iv Using the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy test (see Kaiser 1970) we found that factor analysis can be used to group the variables with similar characteristics together (KM0=0.97). We treat the communalities - the proportion of variance that is due to the influence of common factors - as all 1 suggesting that are no unique factors. The results from the principal-component factor model are not presented here, but are available upon request. The results support using the 22 measures of job satisfaction to identify three factors, with each factor incorporating the items as discussed in the text. v Evaluations of the job: 1 if the employee is dissatisfied (10.3%); 2 if satisfied (40.4%) and 3 if very satisfied (49.3%); Evaluations of information: 1 if the employee is dissatisfied (26.2%); 2 if satisfied (34.1%) and 3 if very satisfied (39.7%); Evaluations of manager-worker relationships: 1 if the employee is dissatisfied (20.9%); 2 if satisfied (34.4%) and 3 if very satisfied (44.7%). vi Since SMMWs that are part of an SME and single-workplace SMEs are both same size firms, the differences solely reflect ownership differences. vii The workforce and establishment characteristics considered in the model are showed in Table A3 in the Appendix. We also provide a summary of the results for the overall EEWE model. viii A joint test for exclusion of the four size dummies takes a value of x 2 ( 4 ) = 54 . 05 , x 2 ( 4 ) = 38 . 22 , x 2 ( 4 ) = 50 . 54 , x 2 ( 4 ) = 70 . 37 , respectively. This suggests that the hypothesis of zero restrictions on the coefficients of size variable is rejected for each model. ix When the model is estimated without the size variable the estimated formality coefficient are larger in magnitude. x The full results are available upon request. 37 xi Similar conclusion can be drawn when an ordered probit model is used. We also experimented further by using the overall formality measure in this model. The results that are not reported here but are available upon request suggest that overall formality reduces EEWE in single-site SMEs. xii 38