Is technological change really skills-biased? Firm-level evidence of the complementarities between ICT and workers' education

ABSTRACT

This paper extends and refines the concept of ICT-driven skills-biased technological change by disentangling the effects of information technologies (IT) and communication technologies (CT). Guided by the theory that IT and CT differently affect firms' production processes, we investigate the complementarities between these two distinct technologies and workers' levels of education in affecting firms' productivity. Exploiting within-firm variation between 2005 and 2017, we find that the use of IT – measured as use of business management tools – is particularly beneficial for workers with a tertiary vocational education. In contrast, CT – measured as workers' use of the intranet – is especially complementary to workers with a tertiary academic education. While consistent with the ICT-driven skills-biased technological change hypothesis, our results offer evidence on the necessity for differentiating between the effects of IT and CT on firm productivity when differently educated workers use these technologies.

KEYWORDS:

• Skills-biased technological change
• information technologies
• communication technologies

JEL CLASSIFICATIONS:

• J24
• O33

Acknowledgments

This study is partly funded by the Swiss State Secretariat for Education, Research, and Innovation. We are grateful for comments from Uschi Backes-Gellner, Stefan Wolter, and participants at the KOF-Leading House ‘Economics of Education’ joint Workshop, at the 4th CVER Conference, at the 2nd BIBB Conference on the Economics of Vocational Education and Training, and at 1st LISER-IAB Conference on Digital Transformation and the Future of Work.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Notes

1 We use the terms ‘upper-secondary VET’ and ‘tertiary VET’ for education programs that prepare their students for labor market entry in specific occupations. ‘Occupation’ refers to the profession for which a young person receives training and is synonymous with vocation or trade.

2 The questionnaires, which closely resembles the EU Community Innovation Survey, are available online in all Swiss official languages (German, French, and Italian) at www.kof.ethz.ch/en/surveys/structural-surveys/innovation-survey/

3 The unbalanced structure of the panel might raise concerns regarding a potential attrition bias. To rule out these concerns, we conduct the BGLW-test (see Becketti et al. 1988; Fitzgerald, Gottschalk, and Moffitt 1998) for attrition bias. Specifically, we find no correlation between the dependent variable and dummy variables indicating firms' entry and/or exit from the panel.

4 The Swiss education system has both an academic and a vocational track at the upper-secondary and tertiary levels. After finishing compulsory education, the vast majority of Swiss youngsters starts a vocational education (either dual-VET or full-time VET-school) and receive a nationally recognized VET-diploma that gives them access to vocational institutions at the tertiary level: Universities of Applied Sciences, Professional Education and Training Colleges, and (Advanced) Federal Professional Education and Training Exams. In contrast, the proportion of pupils opting for general education courses at upper-secondary level is relatively small (about 20%). See Wolter et al. (2018) for a detailed description of the Swiss education system.

5 See Bresnahan, Brynjolfsson, and Hitt (2002) for a related procedure.

6 To enhance the representativity in terms of industry and firm size, the KOF Swiss Economic Institutes conducted selective reminder calls after the initial data collection round to firms that were insufficiently represented.

7 Estimations including only linear terms provide qualitatively similar results. Similarly, estimation of complete translog production functions, i.e. estimations including also $K\ast L$ and $K\ast Tech$ interactions, provide qualitatively similar results.

8 This dummy variable for highly skilled firms is time-invariant. To derive it, we pool for every firm the percentage of Lower educated workers across waves. In doing so, we prevent firms from changing dummy from one wave to the next avoid to generate a baseline effect in the fixed effect estimation.