1. Introduction

The Church’s social doctrine looks at science and technology as a God’s gift enabling human beings to develop their freedom, autonomy and creativity and to contribute to the well-being of the humankind as a whole (Caritas in veritate, 2009, 69). In this view, technological progress must not be hindered but rather encouraged. However, some innovations may imply controversial impacts and consequences and they should therefore be targeted and tailored to maximize their social benefits (and so emphasizing their inner nature of public goods) and to minimize their possible negative effects on the society as a whole and on single individuals. In what follows, the relationship between technological change and employment will be investigated in detail, being this issue an important example of how innovation may entail possible social and individual losses.

Indeed, nowadays the increasing use of robots and algorithms able to perform tasks normally requiring human intelligence (such as speech recognition, visual perception, decision-making and translation of languages) has raised again a widespread fear of a new “technological unemployment” wave (see Brynjolfsson and McAfee, 2014). Moreover, not only employment in agriculture and manufacturing appears at risk, but employees in services – including jobs where cognitive skills are dominant – are no longer protected: see for instance how IBM Watson may displace the majority of legal advices and how the platform economy is gradually displacing the real economy (for example Uber is significantly crowding out taxi companies and Airbnb is rapidly becoming the biggest business model of “hotel company” in the world).

Indeed, the fear of technological unemployment has always emerged in ages characterised by radical technological changes. For instance, the striking response of the English workers to the first industrial revolution was the destruction of machines under the phantom leadership of Ned Ludd in the industrial areas and of Captain Swing in the countryside.

On the other hand, starting again at the time of the first industrial revolution, the economic theory has pointed out the existence of economic forces which can compensate for the reduction in employment due to technological progress. Since that time, two views have started to compete in dealing with the employment impact of technological progress: using David Ricardo’s words, the “working class opinion” was characterised by the fear of being dismissed because of innovation, whilst the academic and political debate was mainly dominated by an optimistic ex-ante confidence in the market forces.

2. The economic theory

In terms of economic theory, the clash between the two views mentioned above can be articulated in two steps. The first step states that labour-saving process innovations create technological unemployment, as a direct effect. The other view argues that product innovations and indirect (income and price) effects can counterbalance the direct effect of job destruction brought about by the process innovations incorporated in new machineries and equipment.

The first step is quite obvious: by definition, process innovations aim at producing the same amount of output with a lesser extent of production factors – especially labour – and, therefore, they destroy jobs. The second step can be articulated through different “compensation mechanisms”, able to positively impact on employment, compensating the immediate negative effect.

In the following presentation of the single compensation mechanisms, I also briefly discuss the potential existence of serious drawbacks often either ignored or mis-specified by the optimistic conventional economic wisdom.

a. “New machines”

The same process innovations which displace workers in the user-industries, create new jobs in the capital-sectors where the new machines are produced. However, labour-saving technologies can spread around in the capital goods sector, as well. In addition, new machines can be implemented either through additional investments or simply by substitution of the obsolete ones (scrapping). In this latter case – the most frequent one – there is no compensation at all.

b. “Decrease of prices”

To be economically acceptable, process innovation should lead to a decrease in the unit costs of production and – in a competitive market – this effect should be translated into decreasing prices. In turn, decreasing prices might stimulate an additional demand for products and induce supplementary production and employment. Nevertheless, this mechanism ignores that labour-saving technology initially decreases the aggregate demand due to the cancellation of the demand previously associated with dismissed workers. So, the compensation mechanism affecting prices deals with an initial shrunk demand that has to be more than counterbalanced. In addition, the presence of a competitive market cannot be taken as granted and the decrease in prices may be far less than proportional compared with the decrease in average costs.

c. “New investments”

In a world where the competitive convergence is not instantaneous, the gap between the decrease in costs – associated to technological progress – and the consequent fall in prices (see previous mechanism b), generates extra-profits. These may be accumulated by innovative entrepreneurs, which might invest them in new productions and new jobs. Nevertheless, pessimistic expectations may imply the decision to postpone investments even in presence of cumulated profits obtained by innovation. If a substantial delay in compensation happens, it may cause structural technological unemployment. Moreover, new investments involve the implementation of new labour-saving technologies that in turn originate additional unemployment.

d. “New products”

As clearly stated by Schumpeter, technological change is not only process innovation: the overall picture cannot ignore the potential job creating effect of product innovation which might have a “welfare effect” connected to the creation of entirely new branches of production, in turns creating additional employment. Nonetheless, this labour-friendly welfare effect has to be compared with a potential “substitution effect” caused by the displacement of mature products (for example, the MP3 and other music formats are product innovation currently displacing the compact disk which, in turn, displaced the vinyl). However, historically, the labour friendly nature of product innovation appears to be largely dominant.

Obviously enough, in different historical periods and institutional contexts, the balance between the direct labour-saving effect of process innovation and the counterbalancing impacts of compensation mechanisms and, partly, product innovation, can significantly vary.

Summing up, the economic theory does not provide a clear-cut answer about the quantitative employment effect of innovation, since this depends on institutional factors, crucial parameters such as price and income elasticities, demand and profit expectations and other contextual factors. Since the inner nature of innovation is being a public good (see the introduction), just relying on market forces may imply severe losses in terms of welfare. In more detail, compensation mechanisms can be hindered by the existence of severe market failures and employment compensation may be more or less effective.

3. The empirical evidence

Whilst theoretical economists have been developing stylized models about the employment impact of process and product innovations, applied economists had to identify proxies to measure them and their employment impact. The most common are:

– simple dichotomous variables capturing the presence of process and/or product innovation in a given time;

– the so-called “embodied technological change”, that is process innovation introduced through gross investments; this innovative input is particularly related to labour-saving process innovation;

– research and development expenditures; this innovative input is particularly correlated with labour-friendly product innovations.

– patents; since different patents may have dramatically different economic impacts, the most accurate studies use patents weighted by forward citations.

Together with the choice of a proper indicator of technological change, it is also crucial to clearly identify the level of investigation, whether macroeconomic, sectoral or firm-based.

Country-level studies explore the direct effects and compensation mechanisms at work in the aggregate. While they are attractive from a theoretical point of view (see above), on the minus side they are often severely constrained by the difficulty to find a proper aggregate proxy of technological change. In addition, the final employment national trends are co-determined by overwhelming institutional and macroeconomic determinants difficult to disentangle and to control for.

In contrast, microeconometric (firm-based) studies have the great advantage to allow a direct firm-level mapping of innovation variables, both in terms of innovative inputs and/or outputs. Indeed, only the microeconometric empirical analysis can grasp the very nature of firms’ innovative activities and their employment impact. However, the microeconomic approach cannot take fully into account the indirect compensation effects which operate at the sectoral and country levels (see above). Keeping these remarks in mind, let us turn our attention to the available empirical evidence (for recent extended surveys see Vivarelli, 2014 and Calvino and Virgillito, 2018).

In general, macroeconometric studies have tested the validity of the compensation mechanisms in a partial or general equilibrium framework. On the whole, the (few) aggregate studies available on the subject reveal – not surprising – that technological change can display its labour-friendly nature only when markets are characterized by competition and flexibility and a higher substitutability between production factors.

Switching to a more disaggregated level of analysis, sectoral studies are particularly important in investigating the overall employment impact of innovation connected to the secular shift from manufacturing to services. In manufacturing new technologies seem to be implemented mainly through labour-saving embodied technological change, only partially counterbalanced by market compensation mechanisms. However, this scenario may change if service sectors are considered. Summarizing the available sectoral evidence, a labour-saving tendency emerges in low- and medium- tech manufacturing, while a dominant labour-friendly impact has been detected in the high-tech manufacturing sectors and in the knowledge-intensive services, i.e. those sectors where product innovation is prevailing and demand is more dynamic. For instance, adopting an extended and revised Pavitt taxonomy, Bogliacino and Pianta (2010) found a moderate expansion of hours worked in the “Science-Based” industries driven by the dominant role of new products in those sectors; in the “Specialised Supplier” sectors, weaker positive effects of new products and stronger negative effects of labour saving process innovation were detected; the “Scale and Information Intensive” industries appeared dominated by a labour saving use of technology; and finally the “Supplier Dominated” sectors were clearly characterized by negative employment trends due to process innovation.

Moving to the firm-level studies (nowadays the most flourishing strand of literature), the vast majority of recent investigations provide evidence of a positive link between technological change and employment, especially when R&D and/or product innovation are adopted as proxies of technology and when high-tech sectors (both in manufacturing and services) are considered. However, some evidence of a labour-saving impact of process innovation is also detected by some studies, especially when the proposed analyses are able to take into account the so-called embodied technological change (that is innovation embedded in new capital goods); in this respect, major labour-saving effects can be detected in low-tech manufacturing and within the smaller firms.

4. An interpretative framework from the Church’s social doctrine

Given the theoretical and empirical picture sketched above, which are the suggestions we can derive from the Church’s social doctrine and the correlated implications for the economic policy? Indeed, the investigation of the very nature and profound meaning of labour is one of the main pillars of the Church’s social doctrine and different encyclicals deals with the issue of technological unemployment and its implications. The general framework here is the one depicted in the Introduction: in principle, technological change and innovation should contribute to the progress and prosperity of the humankind and as such should be encouraged; however, they may entail possible negative effects (in our context, born by the labour force) that should be taken into account and mitigated.

Turning the attention to our main issue, the main concern is that new technologies – instead of remaining allies of human beings – may become an enemy and a source of unemployment (Laborem exercens, 5). This may happen when, in a pure capitalistic society, the workforce – instead to be considered as a collection of unique human beings – is simply reduced to a production factor, that can be easily substituted by another production factor, namely by more technologically advanced capital (Laborem exercens, 7). In this context, profit maximization induces labour-saving technologies in order to decrease labour costs, implying an accelerating substitution of workers by machineries (Laudato si’, 128).

The consequent technological unemployment is not only a serious social problem (Centesimus annus, 15) but a deep depravation for a worker, since working is at the very heart of human nature and renders a worker even more human (Laborem exercens, 8). In more detail, in a market economy, entrepreneurs adopt technologies in order to increase productivity (Giovanni Paolo II, Discorso ai partecipanti all’assemblea generale della Pontificia Accademia delle Scienze Sociali: “Verso la riduzione della disoccupazione”, march, 6, 1999) and they do not care about the possible adverse employment consequences of their profit maximizing choices. As made clear above, the other side of the coin of productivity increase is an obvious labour-saving effect that can originate technological unemployment and social distress.

5. Implications for economic policy

Within the general framework shaped by the Church’s social doctrine and discussed in the previous section, I am going to raise some more specific issues that, on the basis of the theory and evidence discussed above, may be addressed by an economic policy able to take into account such a framework.

Firstly, what discussed above is a further confirmation that controlling technologies and their implementation is nowadays the very first source of power; in other words, nowadays the ownership of capital is as important as the ownership of knowledge and technology (Centesimus annus, 32). However, as any private ownership, also the ownership of knowledge and technology should be addressed to the common good (Laudato si’, 93). An obvious implication is that policy makers and the society in general should shape the new technologies in favour of the social inclusion and minimizing the social marginalisation due to the digital divide (Centesimus annus, 33). In terms of employment impact and taking into account the outcomes of the economic research (see above), the implementation of the new technologies should foster labour-friendly product innovation in high-tech sectors rather than labour-saving process innovation in manufacturing and traditional services.

Secondly, as made clear in a more general context (within a fierce critique of the “market idolatry”, see Centesimus annus, 40), the sole market mechanisms cannot assure welfare and a balanced social life. In our particular context, the sole price and income market mechanisms cannot guarantee compensation and full employment. Indeed, the theoretical arguments proposed above and the current empirical evidence show that the mainstream optimism about the full compensation of technological unemployment by market forces is unfounded. Indeed, product innovation is generally labour friendly but process innovation (for instance robots) may negatively affect employment, while compensation may be affected by several market failures. As a policy implication, on the one hand market failures should be taken into account and corrected by societal choices and intervention (Rerum novarum, 26), while safety nets should be framed for the possible victims of automation (this would mean to contrast the current trend of dismantling the welfare state, see Caritas in veritate, 25 and 32).

Thirdly, together with its possible overall labour-saving effect, innovation implies a substantial impact of the composition of the workforce with some skills and tasks that become obsolescent and some new competencies that become more and more requested (Laborem exercens, 5). In this context, the main societal challenge is to face the possible consequences of the implementation of the artificial intelligence: in fact, self-learning algorithms and human-imitating robots can perform tasks which so far have required human beings. In particular, innovation destructive of “routinized jobs” has emerged as a powerful driver of both technological unemployment and increasing polarization of jobs and wages, involving all the economic sectors and even highly cognitive tasks (see Frey and Osborne, 2017).This kind of scenario – potentially affecting each job and every task points out that policy makers and the society as a whole has the responsibility to provide proper answers in terms of labour market policies and education and training policies (Laborem exercens, 18). While the formers should design safety nets and labour mobility plans, the latter should envisage massive investments in education mainly targeted to those skills and tasks that appear more creative and less automatable.

Voce correlata: Tecnologia e lavoro al tempo dell'intelligenza artificiale