R&D, exports and employment growth: firm-level evidence using administrative data from South Africa

Abstract

This paper asks two main research questions. First, what is the relationship between firm-level innovation and export performance? Second, what is the direct effect of innovation on employment, while accounting for the indirect innovation-export link? I use a novel longitudinal administrative dataset of all firms registered to pay tax in South Africa, over 2010–2016. The first main result suggests a positive and significant relationship between innovation and export performance. Second, the fixed effects results show that direct innovation positively impacts employment growth in manufacturing firms, however, R&D-induced exports – the indirect channel – negatively impacts employment growth. This points to the notion that productivity enhancements necessary for South African manufacturing exporters to compete internationally might promote labor-saving innovations. Nonetheless, the overall innovation effect on firm employment growth is positive.

KEYWORDS:

• R&D
• exports
• firm growth
• employment
• South Africa

JEL Classifications:

• O3
• F16
• J23
• L25

Acknowledgments

The author would like to acknowledge UNU-WIDER, which provided funding for this project through the ‘South Africa – Towards Inclusive Economic Development’ (SA-TIED) project. Data access was made possible through support from the South African National Treasury and the South African Revenue Service. Special thanks to Amina Ebrahim, Junior Chiweza, and Treasury staff members who were helpful in answering pertinent questions about the data. The author is thankful for valuable feedback and comments on earlier versions of the paper from participants at the SA-TIED Work-in-Progress Meeting and the National Treasury Seminar Series, Léonce Ndikumana, James Heintz and Ina Ganguli.

Disclosure statement

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

Notes

1 Since the global financial crisis, unemployment has remained above 25% by the official definition, which estimates the proportion of the unemployed labor force actively seeking employment.

2 See Dosi, Grazzi, and Moschella (2015) for a comprehensive review of the literature.

3 Access to the dataset is through a secure data lab at the National Treasury office in South Africa.

4 The size classifications used in this paper are from the South African National Small Business Amendment Bill, 2003.

5 Or if R&D creates a stock of knowledge then the prior stock, which is often proxied by R&D expenditure itself would also enter the R&D and export decision.

6 For the regression where labor productivity is the dependent variable, the lagged dependent variable is also included. Labor productivity is measured as real value added per worker.

7 The Hansen test is significant at the 1% level, which suggests the problem of weak instruments for the export intensity regression and cautions the interpretation of this result.

8 The authors find that sectors which produce, adopt and disseminate ICTs are more labor absorbing than sectors that are ICT users.

9 The sample excludes retail and financial intermediation and is more highly weighted toward transport and communication.

Additional information

Funding

This work was supported by a research grant from UNU-WIDER, under the project Southern Africa – Towards Inclusive Economic Development (SA-TIED).