Does corporate social responsibility make over-educated workers more productive?

ABSTRACT

This article provides first evidence on whether corporate social responsibility (CSR) influences the productivity effects of over-education. By relying on detailed Belgian-linked employer–employee panel data covering the period 1999–2010, our empirical results exhibit a positive and significant impact of over-education on firm productivity. Moreover, they suggest that the effect of over-education is positively enhanced when the firm implements a CSR process, especially when it aims to have: (i) a good match between job requirements and workers’ educational level, (ii) a diverse workforce in terms of gender and age, and (iii) a long-term relationship with its workers. When focussing on required education and over-education, the results suggest that CSR, besides representing an innovative and proactive approach for the firms’ stakeholders, may also be beneficial for the firm itself through a bigger increase in productivity for each additional year of required education or over-education.

KEYWORDS:

• Educational mismatch
• productivity
• CSR
• linked panel data

JEL CLASSIFICATION:

• J28
• I2
• J24
• M50

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Under-education may also arise, representing an attained level of education that is lower than the level of education required to perform a job.

² A growing literature also focuses on the incidence and consequences of horizontal mismatch, i.e. the mismatch between the individual’s field of education and his/her occupation. Morgado et al. (2015), for instance, suggest that between 20% and 50% of workers are horizontally mismatched in EU countries. In terms of outcomes, the pioneering paper by Robst (2007), based on US data for college graduates, shows that individuals with a major subject that does not match their work have an annual income penalty of around 11% compared to their well-matched opposite numbers. Nordin, Persson, and Rooth (2010) report that this income penalty is even bigger in Sweden (12% and 20% for women and men, respectively). To sum up, horizontal mismatch is an important but still under-researched phenomenon (Verhaest, Sana, and Rolf 2015). In particular, it would be highly informative if we knew the effect of horizontal mismatch on productivity (rather than on wages or job satisfaction). Unfortunately, our data set provides no information on the worker’s field of education. Therefore, we have chosen in this article to focus on the productivity effects of vertical mismatch (over-education) in interaction with CSR.

³ The educational attainment of a worker is available in seven categories in our data set. This information, reported by firms’ human capital departments (on the basis of their registers), has been transformed into years of education. We then applied the following rule: (i) primary education: 6 years of education; (ii) lower secondary education: 9 years of education; (iii–iv) general, technical, and artistic upper secondary education: 12 years of education; (v) higher non-university education, short: 14 years of education; (vi) university and non-university education, long: 16 years of education; (vii) post-graduate education: 17 years of education.

⁴ Note that we also control for mean years of under-education.

⁵ The assumption of persistent productivity both at the industry and firm level is strongly supported by the literature (see e.g. Baily, Hulten, and Campbell 1992; Bartelsman and Doms 2000). Researchers ‘documented, virtually without exception, enormous and persistent measured productivity differences across producers, even within narrowly defined industries’ (Syverson 2011: 326). Large parts of these productivity differences are still hard to explain. This implies that productivity at time t in a given industry or firm is likely to depend significantly on its lagged value. This implies that there are strong arguments for modelling productivitty in a dynamic way, i.e. for including the lagged dependent variable among covariates in Equation (1).

⁶ Expected biases associated with OLS and the relatively poor performance and shortcomings of the FE estimator in the context of firm-level productivity regressions are reviewed by Van Beveren (2012).

⁷ The following sectors are included: Mining and quarrying (B), Manufacturing (C), Electricity, gas, steam and air conditioning supply (D), Water supply, sewerage, waste management, and remediation activities (E), Construction (F), Wholesale and retail trade, repair of motor vehicles and motorcycles (G), Transportation and storage (H), Accommodation and food service activities (I), Information and communication (J), Financial and insurance activities (K), Real estate activities (L), Professional scientific and technical activities (M), Administrative and support service activities (N).

⁸ The SES is a stratified sample. The stratification criteria refer to the region (NUTS-groups), principal economic activity (NACE-groups), and size of the firm. The sample size in each stratum depends on the size of the firm. The sampling percentages of firms are, respectively, equal to 10, 50, and 100% when the number of workers is below 50, between 50 and 99, and above 100. Within a firm, the sampling percentages of employees also depend on size. The sampling percentages of employees reach, respectively, 100, 50, 25, 14.3, and 10% when the number of workers is below 20, between 20 and 49, between 50 and 99, between 100 and 199, and between 200 and 299. Firms employing 300 workers or more have to report information for an absolute number of employees. This number ranges between 30 (for firms with 300 to 349 workers) and 200 (for firms with 12,000 workers or more). To guarantee that firms report information on a representative sample of their workers, they are asked to follow a specific procedure. First, they have to rank their employees in alphabetical order. Next, Statistics Belgium gives them a random letter (e.g. the letter O) from which they have to start when reporting information on their employees (following the alphabetical order of workers’ names in their list). If they reach the letter Z and still have to provide information on some of their employees, they have to continue from the letter A in their list. Moreover, firms that employ different categories of workers, namely managers, blue- and/or white-collar workers, have to set up a separate alphabetical list for each of these categories and to report information on a number of workers in these different groups that is proportional to their share in total firm employment. For example, a firm with 300 employees (namely, 60 managers, 180 white-collar workers, and 60 blue-collar workers) will have to report information on 30 workers (namely, 6 managers, 18 white-collar workers, and 6 blue-collar workers). For more details, see Demunter (2000).

⁹ For instance, we eliminate a (very small) number of firms for which the recorded value added was negative.

¹⁰ Some robustness tests have been done with the threshold fixed at 50 observations. Given that the number of data points per occupation at the ISCO 3-digit level is quite large, this alternative threshold has little effect on sample size and leaves results unaffected.

¹¹ As it leads to a very small drop in sample size, this restriction is unlikely to affect our results.

¹² Larger firms are likely to employ a lower share of over-educated workers because they generally have more sophisticated HRM procedures (notably in terms of recruitment) and a wider range of jobs (Dolton and Silles 2001). Moreover, the required level of education is probably better defined in bigger firms. As a result, the fact that medium and large firms are over-represented in our sample may under-estimate the incidence of over-education. Yet, caution is required. Indeed, empirical results provided by Karakaya, Plasman, and Rycx (2007) suggest that the impact of firm size on over-education is very weak in the Belgian private sector. Using matched employer–employee data for 1995, the authors suggest that the likelihood for a worker to be over-educated decreases by only 0.1% ceteris paribus if firm size increases by 100 extra workers.

¹³ The FE estimator only controls for the potential bias related to the time-invariant unobserved workplace characteristics. So, only GMM results are further reported. FE results are available on request.

¹⁴ Interestingly, the GMM coefficient on the lagged dependent variable falls between the OLS and FE estimates (available on request). As highlighted by Roodman (2009), this result supports the appropriateness of our dynamic system GMM specification.

¹⁵ Detailed dynamic system GMM estimates, including control variables, are presented in Appendix Table A1. Regression coefficients associated to the covariates are in line with earlier findings. Most sectoral dummies, for instance, are significant and they follow a similar pattern than that reported in the literature on inter-industry wage differentials (see e.g. du Caju, Rycx, and Tojerow 2012). Among the highly productivity sectors, we notably find the electricity, gas, and water supply industry (NACE D and E) and financial and insurance activities (NACE K). Not surprisingly, as shown by du Caju, Rycx, and Tojerow (2011), these sectors are also found at the top of the conditional wage distribution. The coefficient on part-time is found to be significantly negative. This corroborates estimates of Specchia and Vandenberghe (2013) and Devicienti, Grinza, and Vannoni (2015) showing that firms employing more part-timers are ceteris paribus less productive. An insignificant coefficient for blue-collar workers is also reported by Kampelmann and Rycx (2012). The authors find that occupations play different roles for remuneration and productivity in the Belgian private sector. While their estimations indicate a significant upward-sloping occupational wage-profile, they cannot reject the hypothesis of a flat productivity-profile. Finally, the insignificant coefficient associated to the share of women is in line with Garnero, Kampelmann, and Rycx (2014). The latter show that women are associated to economic rents. More precisely, their findings for the Belgian economy show that increasing the share of women within firms has no significant impact on productivity but decreases the average wage bill. Put differently, firms employing a larger proportion of women would be more profitable.

¹⁶ Note that we ran a test of differences between means in order to know whether a significant difference appears between the estimated parameters for the different subsamples, where the two parameters are not significantly different under the null hypothesis, while the two parameters are significantly different under the alternative. The results, showing that all coefficients are statistically different, are available on request.