James Harrigan, Ariell Reshef, Farid Toubal23 January 2019
Economists have studied the nexus between labour demand, globalisation, and technology adoption for decades, but quantifying the relative importance of these factors is challenging. Using firm-level data from France, this column proposes a new measure of productivity based on the number of workers in technology-related occupations. It finds large effects of importing, ICT, and R&D on the relative demand for skilled workers through their effects on productivity. Interestingly, the demand for both skilled and unskilled workers rises when firms hire 'techies' or engage in offshoring
Since 1980, three trends have characterised the global economy: rising wage inequality, globalisation, and the growth of information and communication technology (ICT). Economists have been studying the nexus between these trends for decades, with Helpman (2018) and Acemoglu and Autor (2011) offering insightful surveys of this research. Theory and anecdote suggest that some combination of technology and globalisation has raised the relative demand for more skilled and educated workers, thus leading to rising wage inequality, but quantifying the relative importance of these forces has proved challenging. There is a consensus among researchers that technological change is 'skill-biased', which is to say that it raises the relative demand for more skilled workers, but direct firm-level evidence on the drivers of skill-biased technological change is remarkably sparse. In a recent paper (Harrigan et al. 2018), we are the first to estimate the separate firm-level effects of research and development (R&D), ICT adoption, exporting and importing on productivity and its bias towards skilled workers. Using employer-employee matched data from France between 2009 and 2013, we find large effects of importing, ICT and R&D on the relative demand for skill through their effects on productivity.
Measurement of ICT and R&D at the firm level
In order to identify the effects of ICT and R&D on firm-level productivity we use data on workers in technology-related occupations. We call these workers 'techies'. We view techies – engineers and technicians with skills and experience in science, technology, engineering, and maths – as essential to productivity growth, by creating new products and processes as well as by mediating technology. For example, consider ICT managers and support staff in a firm. These techies are central in creating, planning, installing, and maintaining ICT, as well as in training and assisting other workers in the use of ICT. Techies (for example, engineers in manufacturing firms) are also involved in R&D. They design new products and lead R&D teams, and ensure the transfer of new knowledge to other workers in the firm.
We measure techies by their share of hours worked. This measure of firm-level technological sophistication can be compared to R&D expenditures, a common metric for technology adoption in the literature. Firm-level R&D is a useful measure, but it excludes much of the ongoing expenditure and managerial attention that firms devote to technology adoption and ICT use. In fact, reported R&D is not even a necessary condition for technology adoption and innovation, and firms employ many scientists and engineers in non-R&D occupations. Conversely, R&D is likely to be impossible without the employment of techies, who are needed to install, maintain and manage the ICT used in R&D departments. Thus, the techie share is a more comprehensive measure of firm-level effort devoted to technology adoption than R&D expenditure. In our data, ICT and R&D techies together account for 3.6% of hours worked and 5.2% of the aggregate wage bill. About 44% of techies are in ICT-related occupations.
Causes of skill-biased technological change at the firm level
In our paper, we apply and extend new econometric techniques for the estimation of production functions and inferring productivity at the firm level. Our methodology allows us to separately identify two dimensions of productivity. The first dimension raises the productivity of all of a firm's inputs (capital, materials, and all types of labour) equally. In addition to this 'Hicks-neutral' measure of technology, we also identify an element of technological progress that directly raises the productivity of the firm's more skilled and educated workers. This 'skill augmenting' technology measure turns out to be central to our findings. In the final step of our methodology, we open up the black box of productivity and estimate the causal effects of firms' investment in ICT and R&D and of importing and exporting decisions on firm productivity.
We find that compared to firms that don't employ techies, firms with a lot of techies (in the 75th percentile) have skill-augmenting productivity which is 60% higher. We do not find that techies have a statistically significant effect on Hicks-neutral productivity. Turning to the effects of trade, we find that firms that import have skill-augmenting productivity which is 120% higher, and Hicks-neutral productivity which is 50% higher, compared to firms that do not import. The effect of importing on skill-augmenting productivity is mainly due to imports of intermediate inputs, which is consistent with the view that offshoring raises the productivity of skilled workers. We find no effect of exporting on productivity (i.e. no 'learning by exporting' effect), which is not surprising in the French context.
Putting all the pieces together, we are able to quantify the impact of these factors on the demand for skilled and unskilled labour. Again, compared to firms that don't employ techies, firms with a lot of techies have employment of skilled labour that is 60% higher, employment of unskilled labour that is 15% higher, and skill intensity that is 40% higher. The effects of importing are even larger. Firms that import have employment of skilled labour that is 115% higher, employment of unskilled labour that is 25% higher, and skill intensity that is 70% higher, compared to firms that do not import. When we aggregate across firms, we find that estimates can account for much of the overall increase in skill intensity in our sample from 2009 to 2013.
These results on the employment effects of technological change (mediated by techies) and trade are crucial to public policy debates. They show that unskilled workers are right to be wary of technology and trade, which we find do indeed favour employment of skilled workers. But this is a relative effect – because of the powerful productivity effects of technological change and trade, both skilled and unskilled workers see labour demand rise when the firms where they work hire techies and/or engage in offshoring. The reason is that both forces cause strong competitive effects that prompt firms engaging in these activities to expand. Another policy implication of our analysis is the importance of techies for productivity growth, both through R&D and ICT adoption.
All the results we find in our paper are within-firm effects of firm-level decisions. We do not consider why firms choose to employ techies or import, nor do we consider the effects of these firm-level decisions on industry- or economy-wide wages. These are limitations of the scope of our paper but do not impinge on the credibility of our research strategy. Furthermore, any credible analysis of the effects of technology adoption and globalisation on labour markets must be built on an understanding of what goes on within firms. This is where our contribution lies.
Acemoglu, D and D Autor (2011), "Skills, tasks and technologies: Implications for employment and earnings", in Ashenfelter, O and D Card (eds.), Handbook of Labour Economics, Volume 4b, Amsterdam: North Holland.