a rope fastened between
critical reflections translations language games

Innovation, Employment, Inequality, Part I
(with Translation Notes)

March 29, 2021


This is my translation of an economics article published in La Vie des Idées, the online journal of the Institut du Monde contemporain of the Collège de France. The journal publishes essays, articles and book reviews by academics from various disciplines. This is the first part of my translation of this article. It is presented with translation notes, aimed at showing the reader some of the supporting structure, normally invisible, of a translation.

Publication details

French article by Mathilde Guergoat-Larivière & Malo Mofakhami
English translation by Urmila Nair.
Discipline: economics
Journal: https://laviedesidees.fr/
Date of publication: March 16, 2021.


The Translation


Is innovation going to make jobs disappear? While its effects vary considerably depending on the person and her skills, innovation generally tends to favor the highly skilled to the detriment of those less so.

The article

The question of the effects of technology on work is an old one1. The invention of the steam engine and the internal combustion engine forced industrial societies early on to confront the emergence of technological unemployment. The concern fueled many a social struggle, the most famous one being that of the Luddites at the start of the industrial era in England. Yet, the prophecy that jobs would disappear, with humans being replaced by machines on a large scale, has not been fulfilled—to this day. Now, however, digital technologies, driven by advances in big data and artificial intelligence, are reviving fears among some researchers, of a possible massive disappearance of jobs (Frey and Osborne 2017; McAfee and Brynjolfsson 2017).

While this troubling question certainly merits examination, it should not overshadow other concerns. Indeed, several recent studies have demonstrated that what our societies must confront, well in advance of any large-scale disappearance of jobs, are profound changes to the structure of employment2. The transformations in skill requirements induced by recent innovations raise questions about the evolution of job quality and the rise of new forms of inequality in the labor market3.

Drawing on a set of recent studies, this article shows that technological and organizational innovation has favorable effects on jobs and job quality in case of high-skilled occupations (indexed, for example, by an increase in the number of permanent positions). Low-skilled occupations and precarious jobs, however, often disappear (with various short- and fixed-term positions being eliminated). Low-skilled and precarious positions also often suffer a decline in job conditions when innovations are implemented, for example an increase in physical risk. A study of the mechanisms at work offers insights into measures that could help counteract these destabilizing effects. In the absence of a consideration of such measures, a steady pace of innovation could, in the medium term, contribute to a further increase in social inequalities which are already on the rise.

What we talk about when we talk about technological change and innovation

While there is much debate about the effects of technological change on employment, the question of how to measure such change is far from resolved.

In macroeconomic models aimed at explaining growth, “technical progress” was initially considered a residual, that is, a factor whose origin was not explained but which was observed to contribute to growth in addition to the usual factors of production (namely, labor and capital)4. Endogenous models of growth have subsequently proposed explanations of the origins of technical progress5 by foregrounding the role of human capital (education, training, health), public capital (infrastructure that generates positive externalities6 for private producers, etc.) and, of course, technological capital (R&D, patents, etc.).

Economic analyses of the effects of technological change on employment often use R&D expenditure or patent registrations as technological innovation indicators for countries and firms. Public policy at French as well as European levels also relies on these indicators to evaluate the progress of countries in terms of innovation. One of the targets of the Europe 2020 strategy is thus to have 3% of GDP devoted to R&D expenditure.7

Yet these innovation indicators are problematic on several counts. Firstly, R&D expenditure may be treated as an input that could eventually lead to an innovation. However this expenditure does not directly measure the production due to that innovation. Furthermore, this indicator excludes a lot of innovation-related expenditure apart from R&D expenditure, for example expenditure on product design, experimental production, training and investment in machinery related to the innovation. Finally, R&D expenditure is a good indicator with respect to certain aspects of technological change, particularly in sectors of highly capital-intensive activity such as industry8. However, it tends to restrict the notion of innovation itself: R&D expenditure as an indicator involves the risk of underestimating innovation in small firms where expenditure on non-R&D innovation is relatively higher.

Towards surmounting these problematic aspects, since 2005, the Organisation for Economic Co-operation and Development (OECD), in conjunction with the European Union, has developed a set of complementary innovation indicators. The aim is to measure directly the results yielded by innovations (their output) instead of merely considering the inputs that may lead to an innovation9. These indicators are described in the Oslo Manual (OECD 2010)10 and were collected via the Community Innovation Survey (CIS)11 of EU firms, conducted every two years. Two indicators in particular help identify technological innovations: first, the implementation of product innovation, which consists in the introduction of a new or significantly improved good or service, for example a new model of a car or a cellphone12. A second indicator is the implementation of a process innovation, consisting in the implementation of a new or significantly improved production or delivery method, for example an improved assembly line production system or internet sales delivery method. Non-technological innovations are also considered, distinguishing within them organizational and managerial innovations13. These indicators help to adopt a more multidimensional approach towards innovation: the focus is less on the role of R&D and science and technology, and more on the multiplicity of actors involved in the process, in particular on the role of employees and their skills development.

The different levels of analysis of the innovation-employment relationship: sector, firm, worker …

An analysis of the effects of innovation on employment raises the question of the ‘just’ measure of innovation. This question is in turn related to the analytical level under consideration: should one estimate the effects of innovation on employment at the level of a firm, a sector of economic activity or at the macroeconomic level of a country? In other words, on which level can and should one search for answers?14 For the expected effects of technological innovation on employment are not the same on these different levels of analysis. The degree of precision in the measurement of these effects also differs. Firm-level analyses can indicate the effects of innovation implementation on employment within a firm. Such analyses will not yield any information about the effects on employment in other firms. Macroeconomic analyses, by contrast, can give us information about the average, overall effect of innovation on jobs (the creation or destruction thereof). The measure here is, however, less precise since it is difficult to distinguish the effects of innovation from those of the set of co-occurring macroeconomic phenomena (Vivarelli, 2014).

Let us now consider the level of the firm in greater detail. Product innovations are supposed to favor employment since they can lead to the creation of new markets and to an increase in the range or quality of products. Process innovations, by contrast, are often aimed at reducing labor. This is termed their “labor-saving effect.” However, other factors may influence the effects of these two types of technological innovation. For example, in case of a product innovation, substitution effects (also known as cannibalization effects) may occur if consumers buy the new product and simultaneously abandon the older versions. This would entail a very slight effect, or even no effect at all, on the firm’s activity and employment. Furthermore, even if a firm markets a new product and manages to increase both its total sales and its employee numbers, the effects of the new product at the sectoral level could be very different: for instance, one might observe an expansion of the market with an increase in activity and employment within the innovative firm but no change in the demand addressed to other firms. Further, the sales of other firms could even drop if, for example, their clients turn to the competing firm’s new product. This is termed the “business stealing effect.” Thus, innovation could have a positive effect on employment in case of the firm concerned but no effect or even a negative effect at the level of the market or the economic sector concerned.

More generally, the effects of technological product and process innovation on employment will depend on a set of institutional factors, such as the degree of competition in goods markets, the elasticity of demand, the mechanisms of wage adjustment, the bargaining power of employees, and so forth (Calvino and Virgillito, 2017).

Consider next the macroeconomic level of analysis. It is particularly difficult to analyze the effects of innovation on employment at this level. The reason is that many factors change simultaneously. Historically, macroeconomic analyses have generally served to analyze the effects of technical progress. This brings us to Joseph Schumpeter’s analyses and the phenomenon of “creative destruction” that is peculiar to capitalism and is related notably to the role of entrepreneurs: the appearance of new products and processes destroys certain firms but also helps create other, more productive firms in their stead. This analysis bears a close resemblance to Alfred Sauvy’s hypothesis in his theory of “discharge” (déversement) as per which, with the introduction of technological innovations, economic activity and employment are gradually “discharged” from one sector to another—from a primary sector to a secondary sector, then to a tertiary sector. Viewed within this optic at the macroeconomic level, innovation engenders sectoral transformations and changes in skills requirements rather than an overall destruction of employment.

Empirically, macroeconomic analyses generally affirm a positive relationship between innovation and employment, and a hypothesis of complementarity between machines and workers. For instance, the widespread automation of production in the automobile sector has not entailed a disappearance of employment in the sector. The reason is that price decreases linked with increased productivity have fostered an increase in demand. At the aggregate level, price decreases in one sector, which help increase the disposable income of consumers, can also entail an increase in demand in other sectors, the service sector, for example. Empirical studies over the past several decades have confirmed this positive relationship between innovation and employment, which holds even if productivity gains lead to a feebler employment increase than the increase in production. These analyses, which mobilize R&D expenditure in particular as a measure of innovation, form the basis of the support for an increase in such expenditure at French and European levels.

These analyses—whether old or recent, at different levels—all confirm that innovation does not make employment disappear. They do nevertheless raise questions about the evolution of the nature of employment in the context of technological change.

The danger of an increase in inequality due to innovation

More than the disappearance of employment, it is the question of the nature and quality of jobs created and destroyed by innovation that merits consideration. The destruction and transformation of jobs can affect workers quite differently, particularly depending on their skill levels.

Two hypotheses have been advanced in the literature on the subject. The first assumes that technological progress is biased in favor of the highly skilled: jobs with lower skill requirements are lost while skilled workers benefit from technological innovation. This hypothesis is validated by numerous empirical studies conducted at the macroeconomic level as well as by studies at firm and sector levels. In the United States, R&D and IT investments thus seem to have favored a shift in the structure of employment towards the highly skilled (Autor et al. 1998). One sees this in various European countries too—France, the United Kingdom, Germany—even if here again, the indicators that serve to measure the role of innovation are relatively crude.

Another, more recent set of studies focuses on the alternative hypothesis of job polarization (Goos et al. 2014). As per this hypothesis, technological advances decrease the demand for “middling relative to high-skilled and low-skilled occupations” (Ibid.: 2509). In other words, middling occupations tend to be replaced by machines while jobs situated at the two extremes tend to be preserved. On the one hand, high-skilled jobs become more complex and important. On the other, nonroutine low-skilled jobs are created that, being nonroutine, cannot be performed by machines15. These jobs involve interactional, manual, non-repetitive, low-skilled tasks. Jobs at the low-skilled end of the spectrum are also favored by the development of services, for instance personal services in the context of an aging population, individual services, and so forth. This polarization of jobs is observed mainly at an aggregate macroeconomic level and is also related to factors other than technological change, in particular changes in labor market regulation that can favor the development of relatively low-skilled and poorly paid service jobs or of international specializations due to global trade (Ibid.).

These studies consider the manner in which technological change has modified not only the quantity of jobs but also their quality. Such job quality changes can accentuate certain inequalities in the labor market. Technological advances could thus degrade the quality of low-skilled jobs and increase that of high-skilled jobs.

To be continued in Innovation, Employment, Inequality, Part II

Bibliographic references mentioned in the French article

(Note: The bibliographic references cited in the translation notes are contained within the notes themselves and are not included here.)

Autor D.H., Katz L.F. and Krueger A.B., “Computing Inequality: Have Computers Changed the Labor Market?” The Quarterly Journal of Economics, 113, no. 4 (1998): 1169-1213.

Calvino F. and Virgillito M.E., “The Innovation-Employment Nexus: A Critical Survey of Theory and Empirics,” Journal of Economic Surveys, 32, no. 1 (2017): 83-117.

Casilli A.A. En attendant les robots — Enquête sur le travail du clic. Paris: Édition du Seuil, 2019.

Duhautois R., Erhel C., Guergoat-Larivière M. and Mofakhami M., “More and Better Jobs, But Not for Everyone: Effects of Innovation in French Firms,” Industrial and Labor Relations Review (2020).

Eurofound, Telework and ICT-based mobile work: Flexible working in the digital age. Luxembourg: Publications Office of the European Union, 2020.

Frey C.B. and Osborne M.A., “The Future of Employment: How Susceptible Are Jobs to Computerisation?” Technological Forecasting and Social Change, 114 (2017): 254–280.

Gallie D., “Quality of Work and Innovative Capacity: Implication for Social Equality,” QuInnE Working Paper, no. 8 (2018).

Goos M., Manning A. and Salomons A., “Explaining Job Polarization: Routine-Biased Technological Change and Offshoring,” American Economic Review, 104, no. 8 (2014): 2509–2526.

Jaehrling K., Ahlstrand R., Been W., Boethius S., Corchado L., Gautié J., Green A., Iléssy M., Keune M., Latniak E., Makó C., Martín F., Mathieu C., Perez C., Postels D., Rehnström F. and Wright S., “Virtuous Circles Between Innovations, Job Quality and Employment in Europe? Case Study Evidence from the Manufacturing Sector, Private and Public Service Sector,” QuInnE Working Paper, no. 6 (2018).

McAfee A. and Brynjolfsson E. Machine, platform, crowd: Harnessing our digital future. New York: W. W. Norton & Company, (2017).

Mofakhami M., Étude des interactions entre dynamiques d’innovation et qualité de l’emploi : une relation déterminante au cœur des mutations du travail à l’œuvre au sein de l’Union Européenne (PhD thesis, Université Panthéon-Sorbonne — Paris I, 2019)

OECD. Measuring Innovation: A New Perspective. Paris, OECD Publishing (OECD innovation strategy), 2010.

Vivarelli M., “Innovation, Employment and Skills in Advanced and Developing Countries: A Survey of Economic Literature,” Journal of Economic Issues, 48, no.1 (2014): 123–154.

Unless otherwise stated, the notes below are translations of the French source text’s footnotes.

  1. This article is based on recent research in the social sciences conducted within the scope of the QuInnE project (Quality of Jobs and Innovation Generated Employment Outcomes). QuInnE was a European research project conducted between 2015 and 2018 and funded by the Horizon 2020 program. The project’s aim was to analyze the relationships between innovation and job quality in Europe. For more information, visit http://quinne.eu/. ↩︎

    One finds “structure of employment” as the equivalent for “structure des emplois” in one of the references cited by the French article, viz. Goos et al. 2014. ↩︎

    Several articles cited by the French text use the term “labor market” which is therefore the equivalent selected for “marché du travail” (for example, Autor et al. 1998). ↩︎

    The terms “residue” (résidu) and “technical progress” (progrès technique) derive from the work of American economist Robert Solow, on which see the references below.

    References in French:

    • Dans le modèle de Solow, l’augmentation des facteurs de production (travail et capital) explique une part de la croissance. […] Toutefois, la plus grande part de la croissance n’est pas expliquée par ces deux facteurs, mais est due à un « facteur résiduel ». Il s’agit du progrès technique […]” (emphases added; https://wp.unil.ch/bases/2013/05/la-croissance-et-le-modele-de-solow/)

    • on attribue la croissance observée aux facteurs de production et à un résidu appelé progrès technique"(emphases added; Davenport, Paul. Investissement, progrès technique et croissance économique. L’Actualité économique. 1982, 58 (1–2), 167: https://www.erudit.org/fr/revues/ae/1982-v58-n1-2-ae2563/601018ar.pdf)

    References in English:

    • “Robert Solow received a Nobel Prize for his work on the concepts of technical progress function also known as the Solow Residual and total factor productivity (TFP).” (emphases added; https://www.investopedia.com/terms/t/technical-progress-function.asp)

    • “The Solow residual is the portion of an economy’s output growth that cannot be attributed to the accumulation of capital and labor, the factors of production. The Solow residual represents output growth that happens beyond the simple growth of inputs. As such, the Solow residual is often described as a measure of productivity growth due to technological innovation. The Solow residual is also referred to as total factor productivity (TFP).” (emphases added; https://www.investopedia.com/terms/s/solow-residual.asp)

    References on the meaning of “endogenous” versus “exogenous” in the context of economic theories of growth:

    • “Endogenous growth theory maintains that economic growth is primarily the result of internal forces rather than external ones.” (emphases added; https://www.investopedia.com/terms/e/endogenousgrowththeory.asp)

    • “Exogenous growth theory states that economic growth arises due to influences outside the economy. The underlying assumption is that economic prosperity is primarily determined by external, independent factors as opposed to internal, interdependent factors. […] While both the exogenous and endogenous growth models stress the role of technological progress in achieving sustained economic growth, the former posits that technological progress alone, outside of the economic system, is the key determinant in maximizing productivity, whereas the latter suggests that an economy's long-term growth is a byproduct of the activities within that economic system that result in technological progress.” (emphases added; https://www.investopedia.com/terms/e/exogenous-growth.asp#:~:text)

    References on “des externalités positives” (Fr) and the English equivalent “positive externalities”:

    • En économie, on appelle externalité (ou effet externe), les effets ou l’influence directe que peut avoir l’activité d’un agent économique sur son environnement qu’il soit humain, naturel ou économique. L’externalité est caractérisée par le fait que :

      • c’est un effet secondaire de l’activité principale,
      • le bénéficiaire ou la victime de cet effet n’a pas de lien direct avec l’activité d’origine
      • il n’y a pas de contrepartie financière.” (http://toupie.org/Dictionnaire/Externalite.htm)
    • les externalités positives […] sont bénéfiques ou favorables […] apportent un avantage gratuit.” (http://toupie.org/Dictionnaire/Externalite.htm)

    • “An externality is a cost or benefit caused by a producer that is not financially incurred or received by that producer. An externality can be both positive or negative.” (https://www.investopedia.com/terms/e/externality.asp)

    The following equivalences were found here in the context of the “Europe 2020 strategy”:

    On sector (secteur) and industry (industrie), the following references were found:

    • A French government website cites the work of British economist Colin Clark and his three-sector model as the basis of the French usage of secteur in this context:
      Le secteur primaire regroupe l’ensemble des activités dont la finalité consiste en une exploitation des ressources naturelles : agriculture, pêche, forêts, mines, gisements. Toutefois, selon le point de vue, les industries extractives peuvent aussi être classées dans le secteur secondaire.Le secteur secondaire regroupe l’ensemble des activités consistant en une transformation plus ou moins élaborée des matières premières (industries manufacturières, construction). Le secteur tertiaire se définit par complémentarité avec les activités agricoles et industrielles (secteurs primaire et secondaire).” (https://www.vie-publique.fr/fiches/269995-les-grands-secteurs-de-production-primaire-secondaire-et-tertiaire )

    • As per Investopedia, the difference between industry and sector in English in the context of economics is:
      “the terms industry and sector have slightly different meanings. Industry refers to a much more specific group of companies or businesses, while the term sector describes a large segment of the economy.” (https://www.investopedia.com/ask/answers/05/industrysector.asp )

    On innovation indicators and their output, we see for example the following discussion in an OECD document:
    “Indicator and related econometric research must move forward from innovation inputs and activities to include the outputs and impacts of innovation.”
    (Box 1, page 12, Towards a measurement agenda for innovation, 2010. pp. 11–17: https://www.oecd.org/site/innovationstrategy/45392693.pdf↩︎

    Undated version of the Oslo Manual: https://www.oecd.org/science/inno/2367614.pdf ↩︎

    “Community Innovation Survey (CIS)” found on page 5 of the Oslo Manual (op.cit.). ↩︎

    The following equivalences are used here:

    • innovation dite de « produit »” (Fr.) = “product innovation” (Eng.)

    • innovation de procédé […] la mise en œuvre d’une méthode de production ou de distribution” (Fr.) = “process innovation […] production or delivery methods” (Eng.)

    The above English equivalences are based on page 9 of the Oslo Manual (op.cit.):

    “A technological product innovation is the implementation/commercialisation of a product with improved performance characteristics such as to deliver objectively new or improved services to the consumer. A technological process innovation is the implementation/adoption of new or significantly improved production or delivery methods. It may involve changes in equipment, human resources, working methods or a combination of these.” ↩︎

    The following equivalences are used here:

    les innovations « organisationnelles » et les innovations de « marketing »” (Fr.) = “organisational and managerial innovations” (Eng.)

    The Oslo Manual discusses these two types of “non-technological innovations” on p. 88: “The major types of non-technological innovation are likely to be organisational and managerial innovations.”

    Further, note that in the English, “marketing” is used in relation to technological product and process innovation, and not non-technological innovation (pp. 8-9 of the Oslo Manual, op. cit.). ↩︎

    The French terms “niveau [d’analyse]” and “entreprise” are translated as “level” and “firm” here, drawing on the terms employed in the main reference cited in this section of the article, viz. Calvino and Virgillito 2017. ↩︎

    The French “moins routiniers” is translated as “nonroutine,” which is the term used in the article referenced here by the French source text, viz. Goos et al. 2014. ↩︎