Philippe Aghion, Antonin Bergeaud, Richard Blundell, Rachel Griffith 02 January 2020
The earnings of workers in the bottom part of the earnings distribution have performed poorly in recent decades across many countries. In the UK, the lowest-earning working households today earn little more than their counterparts in the mid-1990s (Joyce and Xu 2019: Figure 4). This has led to considerable interest in what policies might be effective at boosting earnings of these workers.
Improved access to and quality of administrative matched employee-employer data has meant that we are increasingly able to study the evolution of the wages of individual workers across substantial parts of their career. For example, in the UK we see that earnings inequalities are largely the result of differences in rates of pay progression (see Figure 1).
Figure 1 Pay progression by skill level
Notes: The figure shows mean hourly wage from Annual Survey of Hours and Earnings (ASHE) 2004-2016. ASHE is a random 1% sample of the UK working population. Wages include basic pay, incentive pay, overtime pay, additional premium payment for shifts that are not considered overtime and additional pay for other reasons. Skill levels are based on occupation classifications according to the National Qualification Framework (NQF), low-skill is NQF1-2 or no formal qualifications required, intermediate skills is NQF3-5, which corresponds to A-level or professional qualifications, high-skill is NQF6-8, which corresponds to a bachelor’s degree or higher.
What are the drivers of these differences? There is a growing literature that emphasises that firm heterogeneity plays a large role in explaining wage differences across workers (Gibbons and Katz 1992, Groshen 1991, Abowd et al. 1999, Bonhomme et al. 2019, among others). Productivity differences across firms typically playing an important role; Card et al. (2018) provide an excellent review. However, there is little consensus in explaining which features of the firm and what economic forces account for such variation in wages, and particularly how these feed through into the wages of workers in low-skilled occupations. In a recent paper (Aghion et al. 2019) we highlight one channel, namely, the interplay between a firm’s innovativeness and the complementarity between the (soft) skills of workers in low-skilled occupations and the firm’s other assets.
We use matched employee-employer data from the UK, augmented with information on R&D expenditures using UK administrative data, to analyse the relationship between wages and innovation. We show that more R&D-intensive firms pay higher wages on average, and in particular workers in some low-skilled occupations benefit considerably from working in more R&D-intensive firms (see Figure 2).
Figure 2 Pay progression by skill and R&D intensity of the firm
(a) workers in low-skilled occupations
(b) workers in intermediate-skilled occupations
(c) workers in high-skilled occupations
This is a surprising finding, and to our knowledge existing models do not provide an explanation. We develop a simple model of the firm which generates this finding. The basics of the model work like this. In some low-skilled occupations, workers are complementary to workers in high-skilled occupations. There are more of this type of occupation in more innovative firms. All workers productivity depends on both hard and soft skills. Hard skills are observable, whereas soft skills are difficult to observe and verify (both by the firm and the econometrician). For workers in low-skilled occupations, their soft skills form a larger proportion of their abilities, and so are important in determining their wages. For workers in high-skilled occupations, easily verifiable skills (such as a degree or publications) are more important in determining their wages.
In the model, what drives the returns to working in an R&D firm for workers in low-skilled occupations is that some have soft skills that are valuable to the firm because complementary with other assets. We are not claiming that the absolute importance of soft skills is greater for workers in low-skilled occupations than for high-skilled, but that soft skills are relatively more important for workers in low-skilled occupations. For example, think of a researcher and an administrative assistant. The researcher might have higher soft skills than the administrative assistant, but her income will be mostly determined by her track record of publications and inventions, which are verifiable. The administrative assistant might have lower soft skills than the researcher, but these will represent a higher share of her value to the researcher, and so play a more important role in determining the assistant’s wage.
The model implies that workers in low-skilled occupations with high soft skills command higher bargaining power in more innovative firms (than in less innovative firms). A worker whose value comes from difficult to observe soft skills is difficult to replace, because these soft skills are unknown at the point of hiring, or they require training to be developed by the firm; it is not a simple matching set-up. Tenure and training increase the wage premium of these workers. Workers in high-skilled occupations typically have observable qualifications; their wage is primarily determined by their education and outputs, which are easily observable and verifiable. A firm can replace a worker with observable hard skills by another similar worker with limited downside risk.
While we do not rule out other possible models or explanations for workers in some low-skilled occupations earning higher returns in more innovative firms, our model generates a number of additional predictions that find empirical support in the data. In particular: (i) returns to tenure are higher for workers in low-skilled occupations in innovative firms compared with non-innovative firms; (ii) workers in low-skilled occupations have longer tenure in more innovative firms than in less innovative firms, and more effort is spent on their training; and (iii) more innovative firms outsource a higher fraction of tasks where there is low complementarity between the workers in low-skilled and high-skilled occupations.
Our finding that the premium to working in more innovative firms is high for workers in some low-skilled occupations is not at odds with the view that technical change has become increasingly skill-biased over the past 35 years (Acemoglu 2002, Goldin 2010, Acemoglu and Autor 2011, Krusell et al. 2000). Indeed, we find that more innovative firms outsource a higher fraction of tasks performed by workers in low-skilled occupations, but presumably they keep those workers in low-skilled occupations with high soft skills and which are more essential to the firm. Thus, Akerman et al. (2015) study the impact of the adoption of broadband internet on wages and find that overall workers in low-skill occupations benefit less from the new technology even though the quality of some workers in low-skill occupations and the tasks they perform remain valuable.
The idea that innovation should play a key role in explaining cross-firm wage differences is in line with the endogenous growth literature (e.g. Romer 1990, Aghion and Howitt 1992), where innovation-led growth is motivated by the prospect of rents. For example, Aghion et al. (2018a) and Akcigit et al. (2017) look at the effects of innovation on income inequality, while Kline et al (2018) and Aghion et al. (2018b) find that returns to successful patenting accrue to the inventors and also to other employees or stakeholders within the inventor’s firm. We contribute to this literature by focusing on the returns to soft skills for some workers in low-skilled occupations in more versus less innovative firms, and on how innovativeness affects the degree of complementarity between these workers and the firm’s other assets (including workers in high-skill occupations).
Our analysis can be extended in several directions. It would be interesting to look at whether, as our model predicts, the (low-skilled) occupations that yield more return to innovativeness (i.e. for which wage increases more with innovativeness) are more relational. A second idea is to further explore whether more innovative firms provide more training to workers in low-skilled occupations. Third, our model predicts that our main effect – namely, that workers in low-skilled occupations benefit more from working in a more innovative firm – is stronger in more competitive sectors or in areas where potential replacements for incumbent workers in low-skilled occupations are of lower quality. Fourth, we used R&D investment as our measure of innovativeness, and it would be interesting to look at other measures, such as patenting.
Abowd, J M, F Kramarz, and D N Margolis (1999), “High Wage Workers and High Wage Firms,” Econometrica 67(2): 251–334.
Acemoglu, D (2002) “Technical Change, Inequality, and the Labor Market,” Journal of Economic Literature 40(1): 7–72.
Acemoglu, D and D H Autor (2011), “Skills, Tasks and Technologies: Implications for Employment and Earnings,” in O Ashenfelter and D E Card (eds), Handbook of Labor Economics Vol. 4, Elsevier, pp. 1043–1171.
Akcigit, U, J Grigsby, and T Nicholas (2017), “Immigration and the Rise of American Ingenuity,” NBER Working Paper 23127
Aghion, P, U Akcigit, A Bergeaud, R Blundell and D Hémous (2018a), “Innovation and top income inequality,” The Review of Economic Studies 86(1): 1–45.
Aghion, P, U Akcigit, A Bergeaud, R Blundell and R Griffith (2019), “The innovation premium to soft skills in low-skilled occupations“, CEPR Discussion Paper 14102
Aghion, P and P Howitt (1992), “A Model of Growth through Creative Destruction,” Econometrica 60(2): 323–351.
Aghion, P, A Hyytinen, and O Toivanen (2018b), “On the returns to invention within firms: Evidence from Finland,” in AEA Papers and Proceedings 108 pp. 208–12.
Akerman, A, I Gaarder, and M Mogstad (2015), “The skill complementarity of broadband internet,” The Quarterly Journal of Economics 130(4): 1781–1824.
Bonhomme, S, T Lamadon, and E Manresa (2019), “A Distributional Framework for Matched Employer Employee Data,” Econometrica 87(3): 699–739.
Card, D, A R Cardoso, J Heining and P Kline (2018), “Firms and Labor Market Inequality: Evidence and Some Theory,” Journal of Labor Economics 36(S1): S13-S70
Gibbons, R and L Katz (1992), “Does unmeasured ability explain interindustry wage differentials?,” The Review of Economic Studies 59(3): 515–535.
Goldin, C and L F Katz (2010), The Race between Education and Technology, Belknap Press.
Groshen, E L (1991), “Sources of intra-industry wage dispersion: How much do employers matter?,” The Quarterly Journal of Economics 106(3): 869–884.
Joyce, R and X Xu (2019), “Inequalities in the twenty-first century: introducing the IFS Deaton Review“, IFS, May 2019,
Kline, P, N Petkova, H Williams, and O Zidar (2018), “Who Proﬁts from Patents? Rent-Sharing at Innovative Firms,” NBER Working Paper 25245.
Krusell, P, L E Ohanian, J-V Ríos-Rull, and G L Violante (2000) “Capital-skill complementarity and inequality: A macroeconomic analysis,” Econometrica 68 (5): 1029–1053.
Romer, P M (1990), “Endogenous Technological Change,” Journal of Political Economy 98(5): 71–102.