Overhead labour and feedback effects between capacity utilization and income distribution: estimations for the USA economy

ABSTRACT

Empirical studies on the USA have not reached a consensus on whether its demand is wage- or profit-led, leading many scholars to scrutinize what drives the empirical results. This article tests two possible explanations for profit-led results which are related to the presence of overhead labour. To do so, a vector autoregression model is estimated for the USA from 1964 to 2010 and the wage share is split between supervisors/managers and direct workers. The results support the argument that the income redistribution away from workers and towards managers increased the likelihood of profit-led demand and suggest that an increase in the workers’ share of income would stimulate the economy. Also, increases in capacity utilization negatively affect the supervisors’ share, so that short-run profit-led results may be capturing the cyclical behaviour of the profit share, but the effect becomes positive as time goes by, suggesting a complex determination of functional income distribution, as capacity utilization affects it in ambiguous ways.

KEYWORDS:

• Capacity utilization
• functional income distribution
• within-wage distribution
• Kaleckian models

JEL CLASSIFICATION:

• E12
• E25
• E6

Acknowledgements

The author thanks Marc Lavoie, Maxime Gueuder, Eckhard Hein, Rosangela Ballini, Carolina Baltar, Rafael Ribeiro and two anonymous referees for their comments, as well as Simon Mohun for sharing the data for the supervisors’ and workers’ shares of wages. The usual disclaimers apply.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. The study closer to the present one is that by Carvalho and Rezai (2015), which includes personal income distribution but does not undertake a class perspective. Even if income groups reflect social classes to some extent, splitting wage earners into workers and supervisors shows how this particular income group encompasses very different social classes.

2. Indeed, Barrales and von Arnim (2017) find Granger causality between capacity utilization and the wage share in both directions, which suggests that the wage share is endogenous.

3. The model does not capture the restriction that the aggregate demand components add up to GDP (Onaran and Galanis 2014, 2495).

4. One may also point out that, while Barbosa-Filho and Taylor (2006, 392) argue that the main difference of their model to Goodwin’s model is that they ‘substitute the global rate of capacity utilization for the employment rate’ on the claim that the former would be a good proxy to the latter, Skott (2015) argues that there is no theoretical or empirical ground for this assumption. Also, while Barbosa-Filho and Taylor (2006) model is interpreted as a short-run model, with reasonable parameters Goodwin’s model would generate a cycle of 16–22 years (Atkinson 1969), which would hardly fit the short-run definition. I thank an anonymous referee for point out the shortcomings of the reference to Goodwin’s model.

5. The fallacy of composition argument pertains to the idea that even if all economies are individually profit-led due the positive impact of lower wages on net exports, if all economies reduce their real wages seeking to gain price competitiveness and increase their utilization rates, the result would be detrimental to the world economy (Blecker 1989; Lavoie 2014, ch. 1).

6. It is worthwhile mentioning that Petach (2019) uses the minimum wage at the state level as an instrumental variable to the labour share and estimates a non-parametric distributive curve that presents strong non-linearities. However, the estimated distributive curve has a different shape to that estimated by Nikiforos and Foley (2012).

7. In this case, labour productivity would also be procyclical. For other implications of the model with overhead labour, see Lavoie (2009, 2014, ch. 5).

8. I thank an anonymous referee for pointing this out.

9. In an open economy the mechanisms would be more complex, but, to some extent, the same rationale holds. See Blecker (1989, 2002) for instance.

10. The main restriction to the sample’s size derives from the wage income shares data. This leads to a rather limited number of observations, but empirical studies on the topic have similar sample sizes (Hein and Vogel 2008; Onaran and Galanis 2014; Onaran and Stockhammer 2005), as this restriction arises from the type of data required for these studies. While some authors opt to interpolate the data in order to achieve a higher number of observations (Carvalho and Rezai 2015), this requires additional assumptions on the series’ behaviour. Given the lack of information on the quarterly behaviour of the two wage income shares that would allow them to be interpolated in order to have quarterly series, I opt to use the annual series.

11. Despite avoiding some of the shortcomings derived from the use of a HP-filter, the Federal Reserve capacity utilization may also be only capturing short-run relations due to how it is constructed (Nikiforos 2016b).

12. Thus, it is assumed that public sector wages are split in the same way as in the private sector.

13. Following an alternative specification of the breakpoint unit root test, as reported in Table A1, the break would be in 1985. Both specifications were tested including a dummy which assumes zero until 1985 and one after 1986. The conclusions of the reported model hold in the specification with this alternative dummy variable.

14. The Eviews file is available from the author upon request.

Additional information

Funding

This work was supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (Brazilian National Council for Scientific and Technological Development) [131651/2015-3,140426/2018-3].