5. Employee Practices and Firm Characteristics
Ichniowski and Shaw (2003) report that innovative employee practices occur most in greenfields and reconstituted sites, while traditional systems are most common in brownfields sites. Consistent with the latter feature, firms that face large transactions costs in shifting systems may become stuck with more traditional HR approaches. The nature of the industry itself may also be important; firms are more likely to adopt innovative HR systems where there is greater potential for workers to work "smarter" rather than just harder (e.g. in industries that produce high-quality or highly complex products or that employ complex equipment for which labour is a complement). Thus uptake of new HR systems may be sector as well as age-specific.
In Canada, Leckie et al (2001) find that variable pay usage increases with firm size but then drops off sharply for firms with over 500 employees. This drop-off is consistent with the brownfields effect reported by Ichniowski and Shaw. Further, Leckie et al report that firms in industries facing rapid technological change and stiff competition are most likely to adopt performance pay. Separately, Leckie et al find that larger establishments tend to support more employee training than do smaller establishments, and training is most prevalent in industries considered to be high-tech. In New Zealand, Gobbi (1998) found that employee training was most prevalent in the services sector; by occupation, it was most prevalent amongst workers in skilled occupations.
Here, we examine typical traits of firms that adopt "high performance" employment practices. We concentrate on SFEP as our measure of employment practices given the results in the previous section. Before doing so, however, we note that only 5.8% of firms adopt all three of the individual HR practices that we have examined, while 33.8% of firms adopt none of them. Individually, 36.6% measure employee satisfaction at least bi-annually, 17.5% have performance pay for many or all staff, and 46.0% invest in innovation-related employee training. These figures indicate a wide disparity in adoption of high performance employee practices across firms.
We regress SFEP on three types of firm demographic variables distinguishing between three sizes,18 five ages,19 and fourteen sectors.20 We omit one of each category from the equation; the results indicate the propensity to adopt high performance HR practices relative to a small, old, agricultural firm. Firms with each of these characteristics generally have the "worst" HR practices across each category. The results are presented in (4):
R2 = 0.107
p-values in brackets
SFEP mean = -0.03;
SFEP std.dev.= 0.83
The results in (4) are consistent with results cited above. Practices improve with firm size: large firms have greater high performance HR practices than do medium sized firms (the difference between the two is significant at 1%), which in turn have better HR practices than do small firms (significant at 1%). Other than start-up firms, firms of all ages up to 10 years perform better than do old firms (although the difference is not significant at 5% for middle-aged firms), consistent with the brownfields results in the cited literature.
By sector, there is no statistically significant difference (at 5%) in HR practices between agricultural firms and those in mining, manufacturing, construction and "low-tech" services (WHT, RET, ACR). However HR practices for all other services sectors (excluding FRT21) are highly differentiated from the low-HR sectors. In these "high-tech" services sectors, firms are much more likely to adopt high performance HR practices.
The differences in HR practices between different types of firm is illustrated in Figure 1. We plot the kernel density for SFEP for all firms, and also for small, old, agricultural and manufacturing firms (grouped together), and for medium-sized and large, very young and young, high-tech services firms22 (grouped together). The rightward shift in the distribution of the latter relative to the former (and to the total) is readily apparent.
Back to Top