A nonbehavioral theory of saving

ABSTRACT

The article presents a demand-driven model, where the saving rate of households at the bottom of the income distribution becomes the endogenous variable that adjusts for full employment to be maintained over time. An increase in income inequality and the current account deficit and a consolidation of the government budget lead to a decrease in the saving rate of the household sector. Such a process is unsustainable because it leads to an increase in the debt-to-income ratio of the households and its maintenance depends on some kind of asset bubble. This framework allows us to better understand the factors that led to the Great Recession in the United States and the dilemma of the present and the future regarding a repeat of this unsustainable process or secular stagnation.

KEYWORDS:

• Closure
• financial balances
• inequality
• saving
• secular stagnation

JEL CLASSIFICATIONS:

• E12
• E21
• E32
• E60

Notes

¹In addition to Kaldor, an early exploration of the concept of closure (without calling it that) is provided by Sen (1963). More complete treatments can be found in Marglin (1984), Foley and Michl (1999), and Taylor (2004).

²Because of the assumption of full employment, Sen (1963) calls this theory “Neo-Keynesian” as opposed to the original Keynesian theory where employment is endogenous.

³Recent contributions along “Kaldorian” lines (e.g., Ryoo, 2015) explain the changes of the past three decades in the distribution of income as the result of the exogenous changes in the long-run growth rate and the saving rate. As we discuss, this kind of explanation is not consistent with Kaldor’s justification of his theory of distribution.

⁴Kaldor was one of the major influences of Godley’s economics. Their paths crossed when they were both working for the Treasury and it was Kaldor who invited Godley to join the University of Cambridge. Godley famously said that Kaldor was touched by genius.

⁵The choice of the 90th percentile as the demarcation point of income distribution of income is common in the literature, but at the same time somewhat arbitrary. One could make the same arguments and arrive at the same conclusions with the bottom 99 percent against the top 1 percent or the bottom 95 percent against the top 5 percent. Further insights could be gained at lower levels of abstraction by looking, for example, at the top 1 percent of the top 1 percent or—as Palma (2011) would suggest—by looking at the top 10 percent against the bottom 40 percent and the middle class (the 50 percent in between). These are important issues, but they are beyond the scope of the present study.

⁶The importance of the role of interest rates, together with the growth rate of income and the rate of inflation for the path of the debt-to-income ratio of U.S. households has been emphasized by Mason and Jayadev (2014).

⁷As Godley mentions in the report, Alan Blinder compared the U.S. economy to one of its mighty rivers—it would “just keep rolling along,” while President Bill Clinton in the economic report of the president for that year was writing that “there are no limits to the world we can create, together, in the century to come.”

⁸Another way to approach this change in income distribution is provided by Tcherneva (2014, 2015), who—using the same data—shows that the top 10 percent has managed to appropriate an increasingly large share of income growth in each of the business cycles of the past four decades.

⁹A more detailed exposition of this argument is provided by Palma (2009, pp. 842–843).

¹⁰As Figure 4 reveals, the calculation of the saving rate is sensitive to the definition of expenditure and disposable income. It is important to keep in mind that the definitions of Saez and Zucman (2014) as well as those of Cynamon and Fazzari (2015b) are different from those of Cynamon and Fazzari (2015a) in Figure 4.

¹¹A summary of recent views on secular stagnation is provided in Teulings and Baldwin (2014). The idea is much older and goes back at least to the 1930s. Significant related contributions include Steindl (1952), Baran and Sweezy (1966), and Hansen (1939).

¹²The relative income hypothesis was meant to be a critique of the standard consumer theory but also of the Keynesian consumption function. The latter can explain the differential saving rate for different income classes but not the stationarity of the saving rate as income grows. It has been sidelined by the so-called life-cycle hypothesis (Modigliani and Brumberg, 1954) and permanent income hypothesis (Friedman, 1957), which are now the standard neoclassical theories of consumption and income, despite their inability to explain the patterns of saving and income.

¹³Smith is quoted in Luttmer (2005, p. 963) and Marx in Easterlin (1995, p. 36).

Additional information

Notes on contributors

Michalis Nikiforos

Michalis Nikiforos is a Research Scholar at Levy Economics Institute of Bard College. The author is grateful to Leilla Davis, Willi Semmler, Mark Setterfield, Anwar Shaikh, Lance Taylor, and an anonymous referee, as well as the participants in the Forty-First Eastern Economic Association Annual Conference in New York, the macroeconomic theory workshop at the New School for Social Research, the Nineteenth Conference of the Research Network Macroeconomics and Macroeconomic Policies (FMM) in Berlin, and the Thirteenth International Post Keynesian Conference in Kansas City for useful comments and suggestions. The author thanks Steven Fazzari and Barry Cynamon for generously providing the time series of the household sector saving rates. The usual disclaimer applies.