Beyond Profit-Led versus Investment-Led Capitalism: Marx and Kalecki on “Capitalists’ Consumption” and “Workers’ Savings”

Abstract

Capitalist production and distribution processes change over time. Despite these changes, many heterodox theorists still view the economy largely in terms of profit-led or investment-led dynamics, which they attribute to the insights of Karl Marx and Michał Kalecki. This essay argues that these thinkers’ theoretical perspectives were more flexible than today’s theorists, adaptable to a changing capitalism. Marx and Kalecki explicitly entertained theoretical space not just for investment decisions but also for “capitalists’ consumption” and “workers’ savings,” broad categories that also participate in capitalism’s conditions of existence. By exploring this theoretical direction in the work of both Marx and Kalecki, by understanding these categories theoretically, and by tracking how these categories have changed empirically over time, this essay newly appreciates Marx’s and Kalecki’s theoretical contributions, enhancing our understanding of capitalism’s changes over the last fifty years.

Keywords:

• Capitalist Consumption
• Investment
• Michał Kalecki
• Karl Marx
• Profits

Acknowledgments

I am grateful for Judith Chien’s editorial assistance on early drafts of this essay. I also greatly appreciate Bruce Norton and an anonymous referee for their deep engagement with this essay’s content and their very helpful suggestions.

Notes

1 This debate emerges in a variety of theoretical debates and to various degrees. Most broadly debated is the role of the falling rate of profit versus the role of aggregate demand and profit realization: e.g., one form of Marxist debate surrounds the falling rate of profit due to the rising organic composition of capital versus a monopoly-capital perspective, often cast in terms of profit-led versus demand-investment-led dynamics.

2 Kalecki ([1939] 1990, 255) writes, “Marx does not pay attention to the problem of what happens if investment is inadequate to secure the moving equilibrium, and therefore does not approach the idea of the key position of investment in the determination of the level of total output and employment.”

3 For example, Marx (1976, 709) makes clear in volume 1 of Capital that he will address more varied distributions of surplus value in volume 3.

4 For example, Sheila Dow (2002, 323) argues that “the two primary founders of post-Keynesian economics are John Maynard Keynes and Michał Kalecki.”

5 Certainly, there are many other important contributors. However, most other influential heterodox economists tend to emerge out of a Marxian or Kaleckian tradition.

6 See Kliman (2017) for a useful starting point and overview of this perspective on Marx.

7 See Roberts (2015).

8 This equation is derived as follows: From the income approach,

Gross national product = gross profits net of direct taxes + wages and salaries net of (direct) taxes + taxes (direct and indirect).

From the expense approach,

Gross national product = gross investment + export surplus + government expenditures + capitalists’ consumption + workers’ consumption.

If you subtract taxes and wages and salaries from both the income and expenditure identities, you arrive at Kalecki’s profit identity. For a fuller explanation, see Kalecki ([1954] 1991, 242–3).

9 Based on Valeria Termini’s ([1981] 2011, 4) work concerning Keynes and Robinson, among others, logical time has three characteristics: “First, there is a unidirectional causality. Second, this scheme cannot cope with changing situations. Third, and most important, any temporal reference is absent from its laws.” For Termini, logical time suggests more of a “convention” to differentiate equilibrium states than denoting any effect of the passage of time. In contrast, “the core of ‘historical time’ is that past, present and future are qualitatively different, linked by expectations and plans” (10). Surely, the past may “leave traces,” but neither the functions’ parameters nor the functions themselves can be presumed as constant (11).

10 Peter Kriesler and Joseph Halevi (2020, 606) also make this connection.

11 This reading sharply contrasts with Trigg (1994), who argues that Marx’s falling rate of profit is a central component of Kalecki’s business-cycle theory.

12 Kalecki was employed as an economic journalist familiar with economic data. This willingness to use data continued throughout his career, including his employment at the Oxford University Institute of Statistics. His understanding of data’s role was careful, critical, and appropriate. “Kalecki was never a naïve empiricist who thought that sense data constituted the only reliable source of information about reality, but he did come to economic theory with a great deal of empirical knowledge, which he put to very good use” (Jefferson and King 2011, 962). Furthermore, “Kalecki’s experience showed that there is no ‘objective’ examination of data” (Toporowski 2013, 150).

13 The method used here for capturing class aspects by incorporating the BLS data on production and nonsupervisory wages to further inform the NIPA data and to distinguish capitalist from worker incomes is unique and valuable. Other investigations, such as by Duménil and Lévy (2011, 45), have relied on income brackets to approximate class aspects. However, despite this focus on class distributions, one still cannot cleanly disentangle capitalists’ consumption from workers’ savings, so we are left with the sum of the two. A few important points can be raised here. First, the difference of capitalists’ consumption minus workers’ savings has investigative value, as capitalists can either consume or save income. Consumption is accounted for directly, and capitalist savings may act as a pool of funds for workers’ dissavings, for government debt, or to be lent to corporations for investment. Hence, this difference captures much detail regarding capitalists’ role in realizing profits. Second, as Duménil and Lévy (2011) make clear, a significant shift has occurred in the distribution of income toward the upper income brackets, likely in favor of capitalists, managers, supervisory laborers, landowners, and equity holders. Therefore, the consumption and savings of these subsumed classes are likely rising. Third, estimates of savings rates by the bottom 90 percent of earners went from a local high of approximately 11 percent in 1983 to −8 percent in 2006. For the years 2009, 2010, and 2011, the savings rate of the bottom 90 percent hovered around zero (Saez and Zucman 2014).

While savings rates are likely very low for workers, moments during the 2008 financial crisis and the COVID-19 pandemic have produced increased savings levels across income brackets. Nevertheless, the rise in incomes for higher income brackets and the reduced savings of those in lower income brackets suggest both consumption out of the surplus and also a reduction of workers’ savings. More about the role of capitalists’ savings will appear later in this essay, but future research is certainly called for regarding the relative roles of capitalists’ consumption and workers’ savings.

14 Granger causality tests do not determine causation. Rather, they test whether a variable is more correlated to lagged values of itself or lagged values of another variable. It is a purely temporal and linear correlation that can be viewed as a necessary but insufficient condition to establish a linear causal connection.

15 Initial stationarity tests revealed nonstationary data series, which are not uncommon with this type of time-series data. All data were first-differenced and then retested to ensure stationarity. Decision on the number of lags was based purely on a “reasonable” sense of a fair time period for results to be felt. No testing of various other lags was involved.

16 For all F-statistics for Granger causality, F = ((SSE^R – SSE^U) / M) / (SSE^U / (n − k)). As the test is for four lags, M = 4; n = 193; and k = 9, the number of parameters estimated in the unrestricted model, including the intercept.

17 Please see Olsen (2015) for an excellent review and extension of the heterodox literature on the rise of unproductive labor.

18 See Wikipedia, s.v. “Paul Allen,” accessed 11 February 2020, https://en.wikipedia.org/wiki/Paul_Allen; Wikipedia, s.v. “Octopus (Yacht),” accessed 11 February 2020, https://en.wikipedia.org/wiki/Octopus_(yacht).