Robots and Mitigation of the Risk of Covid-19 Contagion in the Workplace: Some Evidence From Italy

The virus SARS-CoV-2 quickly spread from China all over the world during the first quarter of 2020. Europe was hit soon, and Italy, Spain and the UK suffered tremendously. No policy recipe book was available and, facing the emergency, national authorities had to implement novel and drastic measures to curb the transmission of the infection. Social distancing measures, ranging from the mandatory quarantine of those exposed to infected people to the recommendations to wear surgical masks and to keep a distance of two meters from other individuals, were meant to preserve public health and slow down the Covid-19 disease. Consistently with this goal, the authorities decided to tackle the risk of contagion at the workplace by temporarily discontinuing certain economic activities.

In Italy, the only sectors that were exempted from the general lockdown were those considered as “essential” by the government. No clear definition of “essentiality” was provided, but implicit reference was made to those sectors that are necessary either to the survival of the population (e.g., the food value chain) or to the full operation of the healthcare sector. Accordingly, the essential activities were identified by means of a broad sectoral classification.

Given the dramatic impact of such measures on the economy, a debate started as to whether other approaches should have been adopted to prevent the virus from spreading, while keeping the economy going. Some observers, in Italy and abroad, suggested to focus on preserving the functioning of the most intertwined sectors of the economy with a view to reducing bottlenecks along the structure of the production network, while others stressed the primacy of protecting those workers that are most exposed to social and physical proximity from the risk of being infected (for instance, by incentivising working at home). According to the latter approach, the actual ability of individuals and firms to comply with safety protocols should be used to decide whether an activity needs to be discontinued or not.

In late April 2020, under the auspices of a dedicated task force appointed by the Italian Government, INAIL (National Institute for Insurance against Accidents at Work) developed a measure of the risk of contagion in the workplace, a measure calculated for each industry on the basis of the exposure, the proximity and the aggregation of its occupations and tasks. This taxonomy was meant, in particular, to informing the selective lifting of the restrictive measures for the upcoming recovery period. Notably, the taxonomy exploits the INAPP-ISTAT classification of 800 Italian occupations, similar to the O*Net Classification used in the US, and accordingly it adheres closely to the specific features of the Italian labour market and industrial structure, providing a reliable source of information of the workers' actual tasks and working practices.

Intuitively, the INAIL taxonomy suggests that, given that the risk of contagion is associated with physical proximity and personal interactions, those firms and sectors where these latter are lower should, in principle, be less exposed to the risk of contagion. In this context, a more intense resort to automation and robots has been advocated to limit workers’ exposure to the risk of contagion and to reduce the likelihood that companies might be suspended. This is indeed one of the suggestions put forward by Abele-Brehm, Dreier, Fuest, Grimm, Krausslich, Krause, Leonhard, Lohse, Lohse, Mansky, Peichl, Schmid, Wess, and Woopen in the very recent monograph “Making the fight against the coronavirus pandemic sustainable”, published by the ifo Institute to contribute to the debate on a safe recovery.

Notwithstanding an intuitive relationship between robotization and contagion, no empirical evidence has been provided yet in the literature. To fill this gap, in a recent note on Covid Economics: Vetted and Real-Time Papers (Issue 17, 13 May 2020), we exploited the taxonomy of risk produced by INAIL and tested empirically whether the diffusion of robots across industries in Italy is associated with the industry-level risk of contagion. We show that this is indeed the case: after controlling for a number of confounding factors, industries where robots have been more intensively adopted tend to be attributed a lower risk of contagion by INAIL. In addition, exploiting the data available for other countries, we study the potential for further robotization in Italy and its potential effect on the risk of contagion. We concluded that the risk of contagion could be further reduced in Italy by pushing robotization closer to the technological frontier in those few sectors where Italy lags behind.

Given the relatively low level of interest rates at the moment and the need to reduce the risk of contagion at work, one would be tempted to derive from this evidence a recommendation to increase investment in robots and automation, in line with anecdotal accounts that a similar trend has already started in China after the suspension of the lockdown. These findings, thus, could have meaningful implications in designing the incentives and the state aid measures that European policy-makers will develop to support the recovery.

However, this scenario opens up a number of relevant issues, in particular in those European countries and regions mostly affected by Covid-19. Robotization, indeed, raises (at least in the short-term) a potential trade-off between safety and employment in the workplace. Some observers, such as Dalia Marin in a contribution on Project Syndicate, have already expressed their concerns that the Covid-19 pandemic and the associated recession might eventually create the incentives to introduce labour-replacing automation. Even trade unions would be in a difficult position as the decision of substituting robots for workers in certain phases of the production process would be motivated by the goal of reducing the tasks requiring physical proximity among workers. This trade-off, moreover, may be more or less relevant according to the system of industrial relations and to whether workers’ representatives participate in co-decision procedures in their own companies.

Ultimately, these considerations bear on the policy response that European policy-makers will design to shape a pandemic-proof economy, as well as on the approaches adopted to suspend only the activities and firms that are truly exposed to the risk of contagion. Hoping, of course, that no further lockdown of economic activities will be necessary in the future.