Financialization, earnings volatility, and firm turbulence in the risk society

ABSTRACT

This paper explores financialization through the lens of with profits accruing primarily through financial channels through risk commodification. While financialization ostensibly increases corporate profits, individual firms have been experiencing escalating volatility driven by heightened competition and technological disruptions. To the extent that firm instability is driven by something fundamental that cannot be reduced for shareholders through portfolio diversification, financialization simultaneously disperses risks to workers, fostering economic instability and exacerbating income inequality. The paper delineates a rise in earnings volatility among workers, attributing it to structural changes in labour markets marked by contingent work arrangements and diminished job security. This volatility reflects broader societal shifts towards a more precarious workforce, necessitating a re-evaluation of traditional measures of economic security. The confluence of these trends manifests in the intertwining of firm turbulence and wage volatility, particularly pronounced in the non-manufacturing sector. In parallel, financialization has also led to rising industrial concentration and market power in the U.S. economy that centralises profit. Superstar firms emerge as key actors, uniquely positioned to maximise their risk/return profile. By restricting profits to a small core of well-compensated workers, these firms are simultaneously offering higher wages while driving down the labour share, thus exacerbating income inequality.

KEYWORDS:

• Risk society
• financialization
• firm volatility
• earnings volatility
• superstar firms

JEL CLASSIFICATION:

• JEL: B50, P16

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. Some authors, for instance David Harvey, characterise neoliberalism as a ‘political project to re-establish the conditions for capital accumulation and to restore the power of economic elites’ – that is to achieve the restoration of class power in response to the economic crisis of the 1970s (Harvey 2005). Harvey points to the inequality of economic resources as a necessary functional characteristic of the ideal market system that comes about through the withdrawal of the state in many social areas that includes things like deregulation and privatisation. Others provide a more complexed and multilayered analysis of neoliberal ideology as something that must be constructed and will not come about ‘naturally’, thereby requiring some form of state intervention, a redeployed rather than receding state (Mirowski 2014).

2. Mirowski (2014).

3. See Peter and Moffitt (2009); Moffitt and Gottschalk (2012); Shin and Solon (2011). For a full summary of different studies, see Hacker and O’Leary (2008) in Table 1 and Dynan et al. (2012) on Table 3a.

4. This is based on a methodology originally outlined in Gottschalk et al. (1994).

5. Hacker and O’Leary (2008). Income volatility is measured by the cumulative percentage growth in the transitory variance of before-tax total family income, defined as the sum of labour income, asset income, and public and private cash transfers for all members of a family.

6. See Peter and Moffitt (2009).

7. Hacker and O’Leary (2008).

8. Dynan et al. (2012).

9. Dynan et al. (2012). Note that most studies look to income volatility from one year to the next, or over a two-year period, as intra-year income volatility is more difficult to assess (month-to-month income data is not widely available). Between 1970 and 2010, the standard deviation of annual hours for household heads rose 30%, and the frequency of very large declines in hours worked also increased. Over the same period, the standard deviation of earnings per hour increased 28%. See Dynan et al. (2012).

10. Hacker and O’Leary (2008).

11. Dynan et al. (2012). Over the same period, the share of households experiencing a 25% drop in income increased from 16% to 23%.

12. Dynan et al. (2012).

13. Dynan et al. (2012) point to an increasing volatility in government transfers. Hacker and O’Leary (2008) reject the ‘windfall’ years explanation, as there is no evidence that people are more likely today than in the 1970s to enjoy huge income gains in a single year and then go back to their ‘normal’ income.

14. Dynan et al. (2012).

15. Dynan et al. (2012) do find that dropping household heads with a financial interest in their business reduces the increase in the volatility of household heads’ earnings from 36% to 29% (which is still not negligible).

16. Jensen and Shore (2015).

17. This is the position of Dynan et al. (2012).

18. Mirowski (2014).

19. Fairlie (2005), Rissman (2003).

20. This said, the focus of the data used by Hacker and O’Leary (2008) and Dynan et al. (2012) is on working age adults so as to minimise the effect of schooling and retirement, with the understanding that these are more likely to be planned.

21. Peter and Moffitt (2009).

22. See Hacker and O’Leary (2008), Figure B.

23. See Hacker (2019), p. 19.

24. Katz and Krueger (2017, 2019).

25. See Hacker (2019) p.54.

26. See Hacker and O’Leary (2012)

27. See Hacker (2019) p.54.

28. See Peter and Moffitt (2009), who point out that while transitory variance rose through the late 1980s it did not fall back to its original levels, as the unemployment rate did in much of the late 1990s and early 2000s.

29. Comin and Mulani (2006); Mcconnell and Perez-Quiros (2000); Stock and Watson (2002)

30. Stock and Watson (2002).

31. Blanchard and Simon (2001). That is to say, excluding recessions does not eliminate the finding of decline in volatility.

32. Apart from a decrease in the volatility of government spending early in the post-war period.

33. There is a strong empirical correlation between output volatility and inflation volatility. See Blanchard and Simon (2001) and Stock and Watson (2002).

34. See Blanchard and Simon (2001).

35. See Kahn, McConnell, and Perez-Quiros (2002); McConnell and Perez-Quiros (2000).

36. See Stock and Watson (2002).

37. Comin and Mulani (2006), Comin and Mulani (2009); Comin and Philippon (2005).

38. This finding is based on the standard deviation of sales growth. Comin and Philippon (2005).

39. See Parker (2006) for the UK, and Thoenig and Thesmar (2004) for France.

40. Davis et al. (2006) suggest that this is not surprising given that in public markets older and larger firms that tend to be relatively stable account for a large share of economic activity.

41. See Davis et al (2006).

42. Irvine and Pontiff (2009).

43. Comin, Groshen, and Rabin (2009).

44. Comin and Mulani (2009); Comin and Philippon (2005). Institutional ownership and the level of financial market development may also play a role in amplifying product market competition and firm level volatility, the reason being that larger pools of investors allow entrepreneurs to pursue riskier business strategies if they can share the risk with diversified investor, see Thoenig and Thesmar (2004).

45. Comin and Mulani (2009); Decker et al. (2006).

46. Philippon (2003).

47. Comin and Mulani (2006, 2009).

48. Comin and Mulani (2009). Their model can account for 75% of the increase in the firm volatility of listed firms, and over 40% of the decline in aggregate volatility.

49. See Comin and Mulani (2009). Using a sectoral approach, aggregate volatility can be decomposed between the sector level variance of sales (variance component), and the correlation of the growth rates of sales across different sectors (covariance component). Comin and Philippon (2005) find that the decline in aggregate volatility is primarily driven by a decrease in this covariance component, rather than a decrease in the volatility of each sector. When looking at firm sales, for instance, the growth rate of aggregate sales exhibits a significant decline in volatility beginning in the 1980s, but the volatility of aggregate growth rates is entirely driven by the covariance of sales growth between firms – not the volatility of individual firm sales. See Comin and Mulani (2006).

50. See Grullon et al. (2019) for the period 2001–2014.

51. Irvine and Pontiff (2009).

52. The evidence of rising industrial concentration and market power in the U.S. economy is manifest in a number of other related trends: a systematic increase of the Herfindahl-Hirschman index (HHI) over the last two decades, with an average increase in concentration levels that has reached 90% (Grullon, Larkin, and Michaely 2019); a decline in business dynamism and entrepreneurship, as well as a lower rate of business start-ups and average number of firms by industry (Decker et al. 2014, 2016); a decrease in the number of IPOs and in the number of public firms, a U.S. phenomenon called the ‘U.S. listing gap’ (Doidge, Karolyi, and Stulz 2017; Kahle and Stulz 2017), particularly for small firms (Ritter, Gao Bakshi, and Zhu 2013); the steady rise in average mark-ups since the 1980s, from 18% to 67% above cost for publicly traded firms in the U.S (Loecker and Eeckhout 2017), as well as mark-ups and price increases resulting from mergers and acquisitions (Blonigen and Pierce 2016); a surge in mergers and acquisitions activity in the U.S. market (at its peak in 2000 total mergers and acquisitions deal value as a fraction of U.S. GDP reached an astounding 16.91% of GDP, see Gabriele, Megginson, and Sanati 2019), as well as an increase in value created during mergers and acquisitions, suggesting that the market recognises concentration as an important source of value (Grullon, Larkin, and Michaely 2019); changes in the distribution of firm market capitalisation, with a sustained increase in mean and median market capitalisation (Kahle and Stulz 2017).

53. Autor et al. (2017, 2020).

54. For example, in 2015 four firms alone (Apple, Google, Microsoft, and Amazon) accounted for a staggering 10% of the earnings of all public firms combined (Kahle and Stulz 2017).

55. Manyika et al. (2018).

56. Comin et al. (2009). This finding holds for volatility measures that capture very transitory and more persistent components of volatility. The transitory variance is the variance in the deviation of a variable over a given time interval (10 years). This can be decomposed into two components: the ‘very transitory’ variance (fluctuations over 5-year intervals); and the ‘persistent’ component, which is computed as the variance of two consecutive non-overlapping 5-year averages of the relevant variable.

57. Comin et al. (2009).

58. Azar et al. (2022) show that going from the 25th percentile to the 75th percentile in labour market concentration is associated with a 17% decline in posted wages.

59. Azar et al. (2022); Benmelech et al. (2018).

60. Comin et al. (2009).

61. Abraham and Taylor (2002).

62. Productivity (revenue per employee) rises massively as they have the ability to scale with relatively little employment. For this reason, sales and earnings rather than employment are the proper measures of concentration (Autor et al. 2017).

63. Karabarbounis and Neiman (2014); Elsby, Hobijn and Sahin (2013).

64. Autor et al., (2017, 2020); Barkai (2020).

65. Loecker and Eeckhout (2017).

66. Comin et al. (2009) show that firms with greater revenue pay their workers better – an increase in real sales of 1% point is associated with an increase in the average wage of about 0.73%. Critically this only holds for wage differentials between the top and the bottom of the distribution, and the top and the middle of the distribution, but not between the middle and the bottom of the distribution. Wages in medium and low-skill job categories are invariant to firm size. This suggests that though wages do increase with firm size on average, this result is exclusively driven by the upper tail of the skill distribution.

67. The mean real wages of firms that employ the top 1% of wage earners increased by 105% between 1982 and 2012, but only by 25% for firms that employ individuals in the middle of the distribution. Based on the distribution of earnings across U.S. firms between 1978 to 2012, the top one percent of the income distribution saw an increase in real wages of a staggering 94%, as compared with a mere 20% increase for the middle of the distribution. Yet during that period the wage gap between the most highly paid employees and average employees (within firms) has only moderately increased, while average wages paid by employers has increasingly diverged (Song et al. 2019).

68. Mueller et al. (2017).

69. Currie and McConnell (1992).

70. Katz and Krueger (2019).

Additional information

Funding

The work was supported by the Open Society Foundations.