Working Paper Series no. 785: The Impact of ICTs and Digitalization on Productivity and Labor Share: Evidence from French firms

Over the last decades, productivity growth has declined in most developed countries, regardless of their distance to the technological frontier. If these rates were to continue, such low productivity gains would make it very challenging to finance the changes needed to tackle the large headwinds we currently face, such as population ageing, rising inequalities, environmental risks or public indebtedness. Digitalization is often seen as the potential source of a huge and possibly long productivity revival able to address these challenges. Hence, it seems important to better understand how ICTs and digitalization affect firms, in particular their productivity and labor share.

An abundant literature has been devoted in the last decades to the quantification of the impact of ICTs on productivity. However, few papers focus on the impact of digitalization on productivity, and those that do mainly concern specific digital technologies, such as broadband access. Moreover, very few papers have analyzed the impact of ICTs and digital technologies on labor share. The goal of this paper is to shed new light on the potential impact of the employment of ICT specialists (in-house and external) and the use of digital technologies (cloud and big data) on productivity and labor share in the case of France. France is indeed a very good candidate for such analysis as it is an important economy, close to the technological frontier and for which individual firm-level data on ICT employment and digital technology use are available.

Our empirical analysis exploits the richness of two Banque de France (BdF hereafter) firm-level datasets. The first BdF database, called FIBEN, contains firm-level annual financial statements and makes it possible to calculate the levels of labor productivity, total factor productivity and labor share. The second database is the BdF survey on factor utilization degrees (FUD hereafter), which has been carried out yearly since 1989 and targets French firms belonging to the manufacturing sector and with at least 20 employees. In addition to collecting original and unique FUD data, this survey also gathers, each year, information on topics of specific interest for policy makers. Interestingly for our purpose, in the survey conducted in 2018, firms were asked whether or not they employed in-house or external ICT specialists and used cloud and big data technologies, and if so for how long. Our empirical approach is definitely a between-firm one. The information on how long these technologies have been used in each firm in our dataset also allows us to estimate learning-by-doing and second-mover advantage mechanisms, meaning that short-term effects could differ from medium-term ones.

The empirical analysis aims at assessing the impact of the employment of ICT specialists and the use of digital technologies on productivity and labor share. However this empirical exercise is far from being trivial. On the one hand, productivity and labor share might also affect the employment of ICT specialists and the use of digital technologies, since most productive firms or firms that have lower labor share are also more likely to adopt such technologies. On the other hand, there might be confounding factors linked to the employment of ICT specialists and the use of digital technologies also influencing productivity and labor share. For instance, firms with a better management quality have a greater propensity to use such technologies and are also potentially more productive for this same reason.

Our empirical results confirm that the employment of ICT specialists and the use of digital technologies have a large impact on productivity. Ceteris paribus, the employment of ICT specialists (through internal and external employment) and the use of digital technologies (through the cloud and big data) could improve a firm’s labor and total factor productivity by about 23% and 17% respectively. Our results also point to the existence of a learning by doing effect concerning the employment of external ICT specialists and the use of cloud, their effect starting to be beneficial for firm’s productivity after five years. In contrast, the use of big data and the employment of in-house ICT specialists seem rather to be associated with a second mover advantage: the early adopters of such technologies face these appropriation costs and late adopters indirectly benefit from their experience. However, the employment of in-house ICT specialists and the use of big data both have a detrimental impact on labor share, of about 2.5 percentage points respectively. While various explanations for this phenomenon have been provided in the literature, one novelty of our study is that our estimations are between firms and not within firms. They suggest that ICTs and digitalization decrease workers’ bargaining power. If confirmed by further firm-level studies, these results would suggest that the employment of ICT specialists and the use of digital technologies might correspond to a third industrial and technological revolution.

In the context of the COVID-19 threat, national lockdown strategies have in all likelihood boosted the use of digital technologies by firms and households. This could be the starting point for an acceleration of ICT and digital diffusion. We could therefore expect significant productivity improvements after the exit of the lockdown periods. This is one possible non negative impact, if any, of this dramatic pandemic event.