Show Summary Details
This content is available to you

Wage-led versus profit-led demand regimes: the long and the short of it

Robert A. Blecker

Keywords: neo-Kaleckian models; growth and distribution; wage-led growth; profit-led growth; Bhaduri–Marglin

Empirical studies have found mixed results regarding whether various countries have wage-led or profit-led demand regimes. Most of the previous literature has paid little attention to the time dimension of this distinction, but most of the studies that have found profit-led results have used methodologies that emphasize short-run cyclical relationships. This paper argues that demand is more likely to be profit-led in the short run and more likely to be wage-led in the longer term, because the positive effects of lower labor costs on investment and net exports are likely to be strongest in the short run, while the positive effects of a higher wage share on consumption are likely to be strongest in the longer term, at least in the US case (other countries may differ).

Full Text

1 INTRODUCTION

The basic logic of the relationship between income distribution and aggregate demand in neo-Kaleckian macro models is now well understood. On the one hand, a redistribution of income toward wages boosts consumption demand because of the higher marginal propensity to consume (MPC) out of wages compared with profits. On the other hand, higher labor costs may diminish the competitiveness of national products, thereby reducing net exports, and can also lessen the profits that are one of the main incentives (or sources of finance) for private investment. Hence, demand is wage-led if the positive effect of a higher wage share on consumption dominates the potentially negative effects on investment and net exports, and is profit-led in the opposite case (Blecker 2002; Lavoie and Stockhammer 2013; Hein 2014). Even if demand is wage-led, economic growth (capital accumulation) can be either wage- or profit-led, depending on whether the positive effect of a higher wage share on capacity utilization is strong enough to outweigh the direct negative impact of lower profitability on investment. 1

Since the possibility that demand-driven economies may be either wage- or profit-led was first opened up by Blecker (1989), Bhaduri and Marglin (1990), and Marglin and Bhaduri (1990), an entire generation of empirical research has attempted to determine whether various countries have wage- or profit-led demand regimes. This now vast literature has yet to reach a consensus for many countries, including some of the largest ones such as the US, Japan, and various European Union (EU) members. Many studies have found that most countries have wage-led domestic demand and the larger economies have wage-led overall demand, although some smaller or more open economies have profit-led demand once foreign trade is taken into account (for example, Hein and Vogel 2008; Stockhammer et al. 2009; Stockhammer et al. 2011; Onaran and Galanis 2012; Stockhammer and Wildauer 2015). Nevertheless, several studies have found that the US has profit-led demand (for example, Fernandez 2005; Barbosa-Filho and Taylor 2006; Storm and Naastepad 2012), while Kiefer and Rada (2015) find that a panel of 13 OECD countries exhibits profit-led demand.

However, one should be cautious about seeking to determine whether particular countries can be unequivocally identified as either wage- or profit-led. The underlying theory implies that, on the contrary, the same economies can behave in either a wage-led or profit-led manner depending on various circumstances. Blecker (1989; 2011) has shown that an open economy with flexible mark-ups may respond differently depending on the source of a distributional shift: the impact is more likely to be wage-led if the shift is caused by a change in the monopoly power of domestic firms, and more likely to be profit-led if the shift is caused by a change in relative unit labor costs. Razmi (2014) shows that stock-flow portfolio balance adjustments in an open economy that is subject to external constraints can make such an economy behave in a profit-led fashion in the medium run, even if it would be wage-led in the short run. Onaran and Galanis (2012), von Arnim et al. (2014), and Obst and Onaran (2016) show that the impact of a redistribution between wages and profits can vary depending on whether the redistribution is limited to a single country or is global in nature. Palley (2016) shows that a redistribution of total labor income in favor of production workers’ wages (at the expense of capitalist-managers’ salaries) can be expansionary even in a system that is ‘profit-led’ overall. Similarly, models that incorporate ‘financialization’ (for example, Onaran et al. 2011; Hein 2012) stress that the distribution of capital income between ‘rentiers’ and firms can affect demand and growth as much as the underlying distribution between labor and capital. Carvalho and Rezai (2016) show that increased inequality in the personal distribution of income can affect the degree to which an economy is wage- or profit-led. Nikiforos and Foley (2012) and Palley (2014) show that, in the presence of non-linearities in either the demand or distributional relationship (or both), there may be multiple equilibria and the response of the system to exogenous shifts in distribution is not independent of initial conditions. Skott (2017) notes that, even in the absence of non-linearities, the slope of the aggregate demand relationship does not uniquely determine the impact of changes in exogenous factors that affect both the demand and distribution curves simultaneously.

This paper will focus on the divergent empirical estimates and will suggest another possible (and less explored) explanation for their conflicting findings. That is, the empirical evidence for profit-led demand regimes in many countries (especially the larger ones like the US) is likely to be relevant mainly to short-run behavior over the business cycle, not to longer-term economic performance. The relative magnitudes of the effects of income distribution on the components of aggregate demand (consumption, investment, and net exports) are likely to vary depending on the length of the time horizon considered. Specifically, it will be argued here that the positive effects of a higher profit share (or lower labor costs) on investment and net exports are mainly short-run phenomena, while the sensitivity of workers’ consumption to their wage income is, if anything, likely to be stronger in the longer term. As a result, holding other factors equal, national economies are more likely to exhibit profit-led (or more weakly wage-led) demand in the short run and more likely to exhibit wage-led (or more strongly wage-led) demand in the longer term.

Indeed, the idea that a higher average profit share would lead to chronically depressed aggregate demand, originally developed by Steindl (1952 [1976]), was not intended as an explanation of business cycles, but rather as a hypothesis about long-term secular stagnation (see Hein 2016). Similarly, the counterargument by Marglin and Bhaduri (1990) that a high profit share could cause secular ‘exhilaration’ if it has a strong positive effect on investment was also intended as an explanation of longer-term behavior. In contrast, most of the empirical literature based on the Goodwin cycle approach (for example, Barbosa-Filho and Taylor 2006; Carvalho and Rezai 2016), which has tended to find profit-led demand, only estimates the role of distribution in short-run business cycle fluctuations.

Only a few empirical studies have explicitly distinguished results over different time horizons. Stockhammer and Stehrer (2011) show that the estimated effects of the wage share on consumption and investment are highly sensitive to lag lengths. Kiefer and Rada (2015) find that demand is profit-led in terms of short-run cyclical dynamics, but also identify a longer-term decline in both the wage share and capacity utilization that is at least consistent with long-term wage-led behavior (even though their method does not allow them to test for the direction of causality between distribution and utilization in the longer term). Kiefer and Rada suggest that common factors, such as a ‘race to the bottom’ by various countries seeking to become more competitive by lowering their unit labor costs simultaneously, could potentially account for the long-run correlation of a lower wage share and depressed utilization. Vargas Sánchez and Luna (2014) find that the profit share (Marxian ‘rate of exploitation’) has a positive effect on output in the short run and a negative effect in the long run using vector autoregression (VAR) and vector error correction (VEC) methods for Mexico. 2 Barrales and von Arnim (2017) perform a ‘wavelet’ analysis that detects cyclical patterns in the data at different frequencies. Using US data with three alternative measures of demand (the income–capital ratio, output gap, and employment rate), these authors find evidence of Goodwin-type cycles reflecting profit-led demand and a ‘profit squeeze’ (that is, rising demand pressure increases the wage share) at all short-run and medium-run frequencies. However, the long-term dynamics show a similar pattern only up to about 1980. Barrales and von Arnim also find that the covariance of their demand measures and the wage share was negative before 1980 at all frequencies, but became positive after 1980 at longer horizons. 3

Before proceeding further, a few caveats are in order. First, when this paper refers to ‘longer-term’ behavior or effects, it does not mean a theoretical long run or steady state in the neoclassical sense (that is, all quantities grow at the same rate) or a theoretical long period in the classical sense (that is, capacity utilization necessarily returns to a ‘normal’ rate). The phrase ‘longer term’ is used here in the spirit of Kalecki's view that ‘the long-run trend is but a slowly changing component of a chain of short-period situations; it has no independent entity’ (Kalecki 1971, p. 165). 4 Second, income distribution is endogenous and there is likely to be simultaneous causality between distribution and demand. Because of space constraints, this paper will focus mainly on the ‘demand relationship’ (how distribution affects demand) and not the ‘distributional relationship’ (how demand affects distribution) or other feedbacks (for example, accumulation of financial assets and liabilities). Finally, the hypothesis proposed here is driven by the author's familiarity with the US case. It is entirely possible that the temporal behavior of consumption, investment, and net exports may differ in other countries that have different economic structures, implying that the short-run and long-term effects of income distribution could vary from those argued here. Some of these potential variations are noted below.

2 EMPIRICAL METHODOLOGIES AND FINDINGS

There are two main approaches to estimating the effects of income distribution on demand, which will be called the ‘structural’ and ‘aggregative’ approaches. To see this distinction, consider a standard version of aggregate demand in a neo-Kaleckian macro model, taken (with some modifications) from Stockhammer et al. (2011): 5

Y = A D = C ( Y , ψ , Z C )   + I ( Y , ψ , Z I )   + G + N X ( Y , P , Z X ,   Z M ) ,
where Y is output, AD is aggregate demand, ψ is the wage share, C is consumption, I is investment, G is government purchases, NX = XM represents net exports (X and M are exports and imports, respectively), P = P(ψ, ZP ) is the domestic price level, 6 and Zj is a vector of exogenous (control) variables affecting endogenous variable j (j = C, I, X, M, P). It is generally assumed that CY > 0, C ψ > 0, IY > 0, I ψ < 0, NXY < 0, P ψ > 0, and NXP < 0. 7 Then the effect of a change in the wage share on output, holding all the exogenous terms Zj constant, is given by
Y ψ = A D ψ 1 A D Y .
Assuming A D Y = C Y + I Y M Y < 1 for Keynesian (goods market) stability, the sign of ∂Y/∂ψ depends only on the sign of the numerator, ∂AD/∂ψ.

The structural approach estimates the individual components of AD using separate econometric equations for C, I, X, M, and P. In this approach, ∂AD/∂ψ is then calculated by summing the various partial derivatives for consumption, investment, and net exports with respect to the wage share, with the effects on net exports mediated by the impact of the wage share on the domestic price level P: 8

A D ψ = C ψ + I ψ + ( X P M P ) P ψ .
In contrast, the aggregative approach relies on estimation of the reduced form solution for output written as
Y = Y ( ψ , Z C , Z I , Z N X , Z P ) ,
and calculates the derivative ∂Y/∂ ψ directly by regressing output Y on (various lags of) the wage share ψ and any control variables that are included. 9

Regardless of which method is used to estimate ∂Y/∂ψ, both of these procedures only aim to estimate the slope of the aggregate demand relationship (1). Such estimates are subject to simultaneity bias if the wage share ψ is endogenous and is a function of output Y through some other channel such as wage- or price-setting behavior on the supply side (see, for example, Sasaki et al. 2013). To address this problem, equation (3) is sometimes combined with an equation representing the opposite direction of causality, that is, the distributional relationship

ψ = ψ ( Y , Z ψ ) ,
where Z ψ is a vector of exogenous factors affecting income distribution, and again lagged dependent variables may be included in the estimation. A model that combines equations (3) and (4) can be deemed a ‘systems’ approach in addition to being aggregative. A special case of the systems approach is the Goodwin cycle model developed by Barbosa-Filho and Taylor (2006), which consists of two simultaneous differential equations in the utilization rate u (defined as actual output relative to capacity or trend output) and the wage share ψ:
u · = f ( u , ψ )
ψ · = g ( u , ψ ) .
Barbosa-Filho and Taylor (2006) used a linearized version of the system (5)(6) written as a system of difference equations in discrete time with lags, which could be estimated as a VAR. The authors concluded that ∂u/∂ψ < 0 and ∂ψ/∂u > 0, that is, demand is profit-led and there is a profit squeeze on the distributive side, using US data. Later studies in this genre (for example, Kiefer and Rada 2015; Carvalho and Rezai 2016) tend to find the same qualitative results.

All econometric estimates using any of these methods are highly sensitive to various aspects of their specifications, including: data frequency and lag lengths; measurement or transformation of the variables (for example, logs, differences, normalizations, filtering, etc.); functional forms (for example, linear or non-linear); control variables included or omitted; and whether the methodology controls for the endogeneity of income distribution. The conflicting results in the literature discussed earlier undoubtedly result to some extent from different choices made by different researchers in regard to various of these issues. Although many of these differences are idiosyncratic to particular authors or papers, some general tendencies can be observed. Especially, the estimates of structural models generally treat the wage or profit share as exogenous, thus creating possible simultaneity bias, while many of the aggregative/systems estimates (especially those using the Goodwin cycles approach) often lack control variables, which suggests the likelihood of omitted variable bias. 10 Also, with a few exceptions, most studies using both approaches have paid scant attention to trying to identify common factors that could drive changes in both distribution and utilization or to testing whether different causes of variations in distribution have different impacts on aggregate demand. 11

Aside from these general pitfalls and caveats, each approach has its own advantages and disadvantages. The main advantage of the structural method is that it can identify the sign and magnitude of the effect of distribution on each component of demand, and thus allows for a distinction between domestic effects (measured by ∂C/∂ψ +I/∂ψ) and the total effect including net exports per equation (2). This method also allows for comparing the impact of distributional shifts within an individual country with the impact of global distributional shifts (Onaran and Galanis 2012). However, most estimates in this genre not only have ignored the potential endogeneity of income distribution, but also have neglected the systems aspect of their models (for example, common shocks) by estimating the individual equations using OLS. Moreover, the procedure of simply adding up the estimated partial effects of distribution on the various components of demand à la equation (2) potentially misses out on important dynamic interactions between the variables.

In contrast, perhaps the greatest advantage of the aggregative approach is that it easily addresses the simultaneity of demand and distribution, for example, by applying IV, VAR, or VEC methods to systems of equations like (3) and (4) or (5) and (6). The aggregative approach may also capture the dynamic interactions that the estimation of individual structural equations could miss. For example, if a rise in profitability stimulates investment and this in turn boosts consumption via the multiplier, this will be captured by an aggregative model as a positive effect of profits on demand, whereas in separate estimates of consumption and investment functions the effect on consumption would be picked up by the total income variable rather than the distributional variable. Similarly, if a rise in the wage share boosts consumer demand and this in turn stimulates investment via the accelerator effect, this would be incorporated in an aggregative model but might not be reflected in separate estimates of an investment function (in which the impact would be picked up by the utilization or accelerator term, not by the distributional variable). 12 However, a key drawback of the aggregative approach is that it cannot provide any insight into which components of aggregate demand (consumption, investment, or trade) are driving the results.

The greatest limitation of the aggregative/systems estimates to date is that they mostly provide information on short-term behavior over periods of business cycle length. This focus is seen clearly not only in the theoretical framework of Goodwin cycles, which most of these studies employ, but also in how they specify the variables in these models empirically. For example, Barbosa-Filho and Taylor (2006) and Carvalho and Rezai (2016) measure u by deviations of real GDP from a Hodrick–Prescott (HP) filtered trend. By this methodology, the mean u ¯ is forced to equal zero, so longer-term variations in utilization are ruled out by assumption. 13 In this approach, the findings of profit-led demand pertain only to short-run fluctuations in utilization relative to its mean or trend. 14

Figure 1 shows the type of information that can be lost by using an HP filter to measure capacity utilization. The graph compares the US utilization rate measured by the percentage deviation (in logs) of actual real GDP from its HP-filtered trend with the utilization rate as calculated by the US Federal Reserve (Fed) based on surveys of firms’ output and capacity. Of course, the two series generally rise and fall around the same times in recessions and recoveries, as one would expect. Nevertheless, the two series also reveal important differences. The deviations from HP-filtered GDP show, rather anomalously, that the Great Recession of 2008–2009 was milder than several earlier recessions. This results from the fact that the HP-filtered trend automatically adjusts to reflect the lower trend of real GDP since the financial crisis of 2008. 15 In contrast, the Fed index shows clearly that capacity utilization was lower in 2009 than at any point in the past 5 decades and also exhibits a downward long-term trend that cannot be seen in the HP-filter-based measure (which is constructed to have a constant zero mean).

Figure 1
Alternative measures of US capacity utilization, quarterly, 1967Q1 to 2015Q3

Citation: 4, 4; 10.4337/roke.2016.04.02

The studies using the structural approach have obtained more mixed results, but on the whole have been much more prone to find wage-led demand regimes (especially in the US case, which is the most studied). Of course, the structural studies have not been intended to identify long-run relationships any more than the aggregative ones have. Nevertheless, for various reasons the structural estimates may come closer to identifying long-run relationships than the aggregative studies. First, the fact that most structural estimates have not relied upon measures of capacity utilization based on HP filters makes them less susceptible to finding purely short-run, cyclical effects. Second, the most robust finding in the structural estimates is a strongly positive effect of the wage share on consumption; to the extent that this effect should dominate in the long run (as argued below), this would tend to make these studies more likely to obtain results that are relevant to the longer term. Third, some of the structural studies may underestimate the short-run effects of profitability on investment as a result of misspecification of the investment function, as discussed in the next section. Fourth, as noted earlier, the structural studies do not capture certain accelerator-multiplier interactions that are likely to dominate during short-run cycles, some of which lean in a profit-led direction (especially the nexus of profits → investment → income → consumption). Overall, the structural studies seem to pick up a mix of short-run and long-run relationships, with the exact mix depending on the precise specifications used in the econometric estimation, so it is not surprising that these studies come up with such varied results.

3 AGGREGATE DEMAND OVER DIFFERENT TIME HORIZONS

This section discusses the theoretical reasons why the negative effects of the wage share on investment and net exports are likely to prevail mainly in the short run, while the positive effects of the wage share on consumption are more likely to dominate in the longer term (with some notes on how other countries may differ from the US in this regard).

3.1 Investment

Investment is by far the most cyclically volatile component of aggregate demand, with fluctuations that are typically much larger than those of total GDP or the other components. In the US economy, profits are normally a leading variable driving investment up and down in expansions and recessions. Figure 2 compares the corporate profit share and the non-residential investment rate for the US economy from 1980Q1–2015Q2. These series exhibit a strong correlation in their short-run, cyclical fluctuations, with evidence of lagged effects of the former on the latter. However, Figure 2 also shows a notable divergence in the longer-term behavior of these two variables: the profit share trends upward (across cycles) from about 1995 to 2015, while the investment rate exhibits a secular decline that is especially notable after 2000. After the financial crisis and Great Recession of 2008–2009, the profit share reached record levels, while the investment rate remained below normal levels for an expansion period. Thus, the longer-term relationship between these variables differs markedly from their strongly positive cyclical correlation.

Figure 2
US corporate profit share and business fixed investment rate, quarterly, 1980Q1 to 2015Q2

Citation: 4, 4; 10.4337/roke.2016.04.02

The argument that profitability effects on investment are largely short-run in nature can be justified theoretically by reference to the accelerator model of investment. A modern version of the accelerator model can be represented by the following equation, which is adapted from Chirinko et al. (1999): 16

I t K t 1 = β 0 + i = 0 m β 1 i Δ Y t i Y t i 1 + i = 0 n β 2 i Δ U C t i U C t i 1 + i = 0 q β 3 i C F t i K t i 1 + ε t ,
where investment (I), output (Y), and capital (K) are measured in ‘real’ (deflated) terms, UC is the ‘user cost’ of capital, 17 CF is cash flow (also in real terms), the β ki are coefficients (k = variable, i = lag), ε t is the error term, and m, n, and q are the number of lags for each variable. CF is the variable posited by theories (both post-Keynesian and mainstream) of financial constraints on investment to be the key factor that relaxes those constraints and allows firms to carry out more of their desired investment plans (Stiglitz and Weiss 1981; Minsky 1986; Fazzari et al. 1988). CF is equivalent to firms’ gross retained profits (net profits minus corporate income taxes, net interest payments and dividend payouts, plus depreciation allowances), and thus represents the profit variable in this investment model.

In this approach, investment (normalized by the capital stock) depends on the rates of change of output and user cost, because these variables affect the level of the desired capital stock, while the flow of investment represents changes in the capital stock. In contrast, the cash flow variable, which is used to capture financial (liquidity) constraints, enters the equation in levels because it influences the current flow of investment spending rather than the desired capital stock. Therefore, profits (as reflected in cash flow) affect only short-run fluctuations in investment and are omitted when the long-run determinants of investment are estimated (Chirinko et al. 2011). This implies that the profit (cash-flow) variable should only matter in the short run; longer-term variations in investment should be driven primarily by accelerator effects (output growth ΔY/Y) and secondarily (if at all) by changes in the user cost of capital (ΔUC/UC). The vast majority of empirical studies using this approach finds that accelerator effects are strong and robust, user cost effects are sometimes statistically significant but usually relatively small, and cash flow is sometimes significant but only has short-run effects (for example, Fazzari 1993; Chirinko et al. 1999; 2011; Spatareanu 2008; Schoder 2013).

In contrast, most structural estimates of neo-Kaleckian macro models have used investment functions of the Marglin–Bhaduri variety:

I / K = f ( π , u ) ,
where π = 1 − ψ is the profit share and the dynamics are usually specified in an ad hoc way. This specification was justified by Bhaduri and Marglin with the claim that since π and u are underlying determinants of the profit rate, they influence the expected future profit rates that govern current investment decisions. But the accelerator theory of investment implies that the level or rate of investment should depend on changes in output rather than the level of output or the utilization rate, in which case the investment function (8) is misspecified. The accelerator approach also implies that longer-term trends in capital accumulation are driven mainly by output growth, while profitability only affects the short-run timing of investment and plays no independent role in the long run. 18

3.2 Net exports

A rise in the wage share can have a negative impact on net exports if it is caused by a rise in unit labor costs (ULC) for domestic goods relative to foreign goods (Blecker 1989; 2002; 2011). 19 Based on the standard Marshall–Lerner (ML) analysis, this negative impact will occur only if the sum of the price elasticities of export and import demand is sufficiently high. 20 Although there is much controversy about whether the ML elasticities condition is normally satisfied, for purposes of this discussion I will stipulate that it is satisfied in order to focus on whether the negative impact of a rise in the wage share on net exports in the short run is likely to persist in the longer term.

Suppose, then, that a country lowers its ULC relative to other nations (through some combination of reducing wages, boosting productivity, or depreciating its currency), thereby improving its external competitiveness and also increasing its net exports in the short run (after any J-curve lags are overcome). 21 There are several adjustment mechanisms that could come into play that would act to offset the country's improved competitive advantages in the longer term. First, if the rise in net exports leads to increased domestic employment, wages may eventually be bid up, thereby offsetting the initial decrease in ULC. Second, the country could experience a real appreciation of its currency, either via a nominal appreciation (if the exchange rate is flexible) or through increased inflation (if the nominal exchange rate is fixed or managed).

Of course, the operation of these or other adjustment mechanisms is far from automatic, as argued long ago by Robinson (1946–1947). For example, either a highly elastic labor supply in a dual economy or a repressive set of labor market institutions could prevent wages from rising in spite of a boom in net exports. Exchange-rate management can be used to prevent currency appreciation, and nominal devaluations can be used to offset higher domestic inflation. Financial inflows can sustain current-account imbalances for some time, albeit at the cost of creating international debt positions that may be unsustainable in the long run. Also, trade imbalances can be sustained for some period of time if Kaldorian cumulative causation sets in. That is, if faster export growth leads to faster output growth which in turn stimulates productivity growth via Verdoorn's law, then a country's international competitive advantages can be sustained and even amplified over time (Setterfield and Cornwall 2002; Blecker 2013). These sorts of mechanisms could account for why some hyper-competitive countries like China exhibit very strongly profit-led demand (Onaran and Galanis 2012).

Nevertheless, even if the standard adjustment mechanisms do not work or if Kaldorian cumulative causation starts to operate, reactions by other countries may also act to offset the home country's initial improvement in competitiveness. Most obviously, foreign countries can engage in competitive devaluations, wage cuts, or efforts to boost productivity, as analysed by Robinson (1947) in her critique of ‘beggar-my-neighbour remedies for unemployment.’ Technology transfers may lessen the competitive advantages of a home country by lowering costs for foreign producers. If any of these responses occur, a country that enjoys a period of profit-led growth will eventually lose its initial gains in net exports, but the negative consequences of the reduced wage share for domestic consumption will still remain. In a similar vein, even if some countries have profit-led demand as a result of strong negative effects of ULC on net exports, those same countries may end up behaving in a wage-led manner in response to a simultaneous change in ULC and wage shares globally (because then the international competitive effects roughly cancel out), with the implication that the world economy as a whole is likely to have wage-led demand even if some countries have profit-led demand individually (Onaran and Galanis 2012; von Arnim et al. 2014; Obst and Onaran 2016).

However, it may take several years for the competitive responses of other nations to cancel out the initial competitive gains of any particular country. Thus, a country may get a competitive advantage from lower ULC (and a lower wage share) that boosts its net exports in the short run, even if this advantage is eventually dissipated by the responses of other nations or other adjustments in the longer term.

3.3 Consumption

Consumption is the one part of aggregate demand for which the impact of income distribution is, if anything, likely to be greater in the longer term than in the short run. Consumption is expected to respond positively to the wage share because of the higher MPC out of labor income compared with capital income. The reasons for a lower MPC out of capital income are well-known. On the one hand, a portion of profits is retained by firms to finance investment or for other purposes (for example, stock buy-backs, mergers, and acquisitions), and all of these retained profits are counted as corporate saving in the national income accounts. On the other hand, pay-outs of dividends and interest go mainly to wealthier households (rentiers) who typically have a much higher propensity to save than middle-class or working-class families. Working-class households, in contrast, typically spend all or most of their wage income on current consumption (although some East Asian countries may be exceptions). The implication that the wage share has a positive effect on consumption is one of the most robust results in the structural estimates surveyed earlier. This same conclusion is supported by numerous cross-sectional or panel studies confirming that consumption–income ratios are inversely related to household income levels (for example, Bunting 1991; Jappelli and Pistaferri 2014), given that higher-income households receive much larger shares of their income from capital than middle- or lower-income households.

However, this evidence leaves open the question of how the responsiveness of consumption to labor or capital income varies over different time horizons. Some considerations from standard theories of the consumption function may help to illuminate this issue. A key ‘stylized fact’ about aggregate consumption is that the MPC is higher in the long run than in the short run. This difference was the basis for many of the classic theories of the consumption function, including Duesenberry's (1949) relative income hypothesis, Friedman's (1957) permanent income hypothesis, and Ando and Modigliani's (1963) ‘life-cycle’ hypothesis. A common foundation of all these theories is the idea that households attempt to maintain relatively steady consumption expenditures in the face of short-run fluctuations in income (albeit for different reasons) and are more constrained by income flows in the long run. One does not have to believe in neoclassical ultra-rationality or strict optimizing behavior to recognize that most households attempt to maintain some degree of stability in consumption in response to transitory fluctuations in income.

Recent work by Cynamon and Fazzari (2008; 2013), Palley (2010), and Kim et al. (2014), among others, has revived Duesenberry's institutionalist approach, in which ‘households will seek to maintain consumption relative to standards achieved in the past and contemporary standards established by others’ (Kim et al. 2015, p. 3). These studies emphasize the role of household borrowing and debt in the context of financialization: consumers are induced to increase consumption in the short run by the easy availability of borrowing, but remain constrained by income in the long run when increased debt levels may become unsustainable. Kim et al. (2015) estimate consumption functions that include variables for household wealth, borrowing, and debt burdens, along with disposable personal income and an index of consumer sentiment. They confirm that the MPC is notably higher in the long run (about 0.6 to 0.9 using two alternative estimation methods) than in the short run (about 0.4 for the whole sample 1952–2011, and 0.1 for the more recent period 1980–2011).

Given that most consumers are reliant primarily on labor income, we may infer from these results that the MPC out of wages is also likely to be higher in the long run than in the short run. In contrast, upper-class households who receive mostly capital income are likely to have MPCs that are not only low, but also do not vary as much between the short run and the long run. Because these households have very high saving rates and large accumulations of assets, their consumption is not closely tied to their income over any particular time horizon. Thus, the variations in MPCs out of wages between the short run and the long run are also likely to imply similar variations in the difference between the MPCs out of wages and profits, which is what matters for whether aggregate demand is wage- or profit-led (Blecker 2002). Nevertheless, this conclusion might not hold in countries where consumers have less access to credit, so that their consumption expenditures are more closely tied to their disposable income even in the short run. This could account for why, for example, López et al. (2011) find strongly negative effects of a real exchange-rate depreciation in the short run (using detrended data) for Mexico: such a depreciation usually results in a steep fall in the real wage in Mexico, where workers lack the ability to sustain consumption levels via borrowing. In such countries, demand could be strongly wage-led in the short run.

4 CONCLUSIONS

Of course, it is not unusual for econometric studies to find conflicting results. Nevertheless, it is disconcerting that empirical researchers in the neo-Kaleckian tradition have not obtained more consistent findings about an issue as important as the links between distribution, demand, and growth, especially in an era in which rising inequality is increasingly seen as linked to long-term economic stagnation in many countries. What is especially striking is the fact that the empirical studies have largely bifurcated into two camps: one that finds almost exclusively profit-led demand regimes using the Goodwin cycles version of the aggregative approach; and another that finds mainly wage-led results (especially for larger countries or groups of countries) using the structural approach.

This paper has argued that these conflicting estimates may result, at least in part, from the different time dimensions of the various distributional effects on aggregate demand, which (implicitly or explicitly) receive different degrees of emphasis in the alternative econometric methodologies. In particular, the studies using the Goodwin cycle approach, which most explicitly emphasize short-run, cyclical effects, have been the most prone to find that utilization is profit-led. I have argued here that (aside from some concerns about potential econometric biases, such as omitted variables or measurement error) this finding is not accidental, at least in the US case, because the positive effects of a higher profit share on investment and net exports are likely to be felt mainly in the short run, while the negative effects on consumption are likely to be felt more strongly in the longer term. As a result, the demand regime is more likely to be profit-led (or more weakly wage-led) in the short run and wage-led (or more strongly wage-led) in the longer term.

If this argument is correct, then the policy implications of the studies that have found that demand is profit-led in the short run need to be reconsidered. It is entirely possible that, in a typical business cycle, profits drive investment both up in the recovery and down in the recession, but a sustained higher profit share of income will not lead to higher investment or growth in the longer term – as we have seen, for example, in the sluggish US recovery since the Great Recession of 2008–2009 (Blecker 2016). As a result, a significant long-term fall in the wage share could potentially depress utilization and growth in the longer term, even if rising profits may be associated with typical short-run cyclical recoveries.

This argument is subject to certain qualifications and limitations. First, as noted earlier, the timing of these different effects of distribution on demand may vary in other countries that have different economic structures from the US and can also vary for different sources of distributional shifts. Second, this argument only concerns the direct effects of changes in distribution on the aggregate demand relationship; it does not address the complex, dynamic feedbacks that can result if other macroeconomic relationships (for example, wage and price setting, financial markets) are taken into account. Third, this argument is more of a hypothesis than a conclusion. To test this hypothesis, it will be essential for future empirical research on this topic to more explicitly compare short-run and long-term effects of distributional shifts on output or utilization as well as economic growth using appropriate econometric techniques, while also addressing the other econometric issues (for example, simultaneity bias, measurement error, and omitted variables) that have plagued the empirical literature on this topic to date.

Finally, it is important to stress two caveats for future research. First, the last few decades may have been qualitatively different from earlier periods. Historically, it was a stylized fact of capitalism that relative shares of wages and profits in national income varied only cyclically and had no long-run trends. Yet, the data shown in Figures 1 and 2 in this paper and several of the most recent empirical studies (for example, Kiefer and Rada 2015; Stockhammer and Wildauer 2015; Barrales and von Arnim 2017) verify that, since roughly the late 1990s, we have witnessed an unprecedented secular decline in the wage share in the US and various other countries coupled with weaker performance in terms of any measure of utilization or growth – a combination often referred to as ‘secular stagnation’ (Blecker 2016; Hein 2016). Thus, it may well be that the potentially wage-led nature of demand in the long run (at least for some causes of distributional shifts) will only become observable after we are able to obtain some historical hindsight on the present era, in which wage shares have sunk to unusually low levels on a sustained basis. Second, it is also clearly time to rethink the simple wage-led versus profit-led dichotomy and to develop a more complex understanding of the dynamic linkages between income distribution and economic performance over various time horizons and under different structural conditions.

  • 1

    If demand (utilization) is profit-led, then growth (accumulation) is also profit-led. In an intermediate case, aggregate demand or capacity utilization can be weakly wage-led while growth is profit-led, a case called ‘conflictual stagnationist’ by Marglin and Bhaduri (1990) and ‘conflictive’ by Palley (2016). In this paper, I use the more descriptive terms ‘wage-led’ and ‘profit-led’ instead of the more colorful but sometimes confusing terminology of ‘stagnationism’ and ‘exhilarationism’ employed by Bhaduri and Marglin.

  • 2

    There is a typographical error in the published cointegrating equation for output, but I have verified (in email from Gustavo Vargas Sánchez, 14 October 2014) that the sign on the exploitation rate (profit share) should be negative.

  • 3

    Earlier, Pérez Caldentey and Vernengo (2013) found that the real wage has a greater ‘coherence’ and ‘dynamic correlation’ with output and investment in several countries using low frequency data compared with medium or high frequency data, but they did not test for the wage share (that is, the real wage adjusted for productivity) and their methods cannot identify the direction of causation between the real wage and the other variables.

  • 4

    Therefore, I do not enter here into the long-running theoretical debate about whether capacity utilization must revert to a given ‘normal’ rate in the long run (see Lavoie 2014, pp. 387–405, for a survey).

  • 5

    For simplicity, I omit the role of intermediate imports and cyclical responses of G from the model of Stockhammer et al. (2011).

  • 6

    In practice, empirical researchers often use different price indexes for exported and imported goods, but I simplify here for expositional purposes.

  • 7

    The last partial derivative, NXP < 0, assumes that the Marshall–Lerner condition holds.

  • 8

    If these effects are estimated as elasticities, then it is necessary to weight them by the shares of the various components of GDP in total GDP.

  • 9

    If autoregressive distributed lag (ARDL) or VAR/VEC methods are used (and they usually are), then lags of the dependent variable (Y) are also included on the right-hand side of the regression equation.

  • 10

    Several economists have argued that debt and other financial indicators could be key omitted variables in this literature. Stockhammer and Michell (2016) demonstrate theoretically that Minskyan debt dynamics based on financial fragility can foster cycles in which utilization appears to be profit-led (because profitability is squeezed at the same time as output falls) even when no causal linkage between distribution and demand is assumed, or if demand is actually wage-led. On the empirical side, Palley (1994) and Kim (2013) provide econometric evidence demonstrating how debt variables affect output in the US economy, while Stockhammer and Wildauer (2015) estimate a structural model (in the sense defined above) that controls for financial factors using international panel data. In regard to simultaneity, Fernandez (2005) controlled for endogeneity of the profit share in an aggregative approach by using instrumental variables (IV) methods applied to estimates of simultaneous equations for the US profit share and utilization rate. He found that changes in the profit share were largely explained by a variable reflecting international competitiveness (relative unit labor costs). Barrales and von Arnim (2017) find bidirectional causality between each of their three alternative measures of demand (discussed earlier) and the wage share, suggesting that any estimates that treat the wage (or profit) share as exogenous are subject to simultaneity bias.

  • 11

    One exception is López et al. (2011), who did not attempt to estimate the effects of the wage share on demand and instead estimated the impact of underlying, exogenous determinants of both output and distribution. Another exception is Kiefer and Rada (2015), who found evidence for a ‘race to the bottom’ of many countries simultaneously seeking to drive their labor costs lower.

  • 12

    Obst and Onaran (2016) address the latter type of interaction by calculating the direct and indirect effects of the wage share on investment using a structural model, and find that the indirect effects (via consumption) make investment (growth) wage-led in many European countries. They do not, however, address the former type.

  • 13

    As noted earlier, Kiefer and Rada (2015) allow for longer-term trends in (or shifts in the means of) these variables, but their methodology cannot determine the longer-term causal relationships between those variables.

  • 14

    Some economists argue that the finding of profit-led demand may be misleading even in the short run. Lavoie (1995; 2014) observes that the profit share varies pro-cyclically because firms hoard overhead labor (and therefore experience a fall in measured labor productivity) during recessions, which implies that a positive correlation of the profit share and the utilization rate does not necessarily imply causality flowing from the former to the latter.

  • 15

    The HP filter was applied using the conventional value of 1600 for the lambda smoothing parameter with quarterly data. The fact that HP filters tend to put too much of the cycle into the trend is noted by Gordon and Krenn (2010), among others. Cogley and Nason (1995) show that HP filters can also generate spurious cycles where none exist. I am indebted to Gabriel Mathy for these references.

  • 16

    Chirinko et al. (1999) include additional subscripts for firms, but I suppress these given my macroeconomic focus.

  • 17

    The user cost of capital is a measure that combines the costs of external funds (interest rates on bonds and dividend pay-outs to stockholders), depreciation costs, and relative prices of capital goods, adjusted for tax policies that affect incentives for investment. See Chirinko et al. (1999, p. 57) for details.

  • 18

    This characterization might not apply in a structurally different situation, such as a developing or emerging market country in which capital markets are weak and investment is typically internally financed. Also, in countries where industries are highly export-oriented and a large portion of investment is dependent on international competitiveness in regard to attracting foreign capital inflows, domestic profit margins compared with those of rival nations could matter to long-term investment decisions.

  • 19

    However, Blecker also shows that a rise in the wage share caused by a reduction in the monopoly power of firms will not have this consequence, and could have a positive impact on net exports if mark-ups are reduced on internationally traded goods.

  • 20

    The standard ML condition is that the sum of these elasticities must exceed unity in absolute value under certain simplifying assumptions, including initially balanced trade. Modified ML conditions can be derived under more general assumptions.

  • 21

    The price elasticities of demand for exports and imports are likely to be low in the very short run (when firms are locked into contracts and goods have already been ordered or are being shipped) and to increase over time (as ordering and delivery lags are overcome, and firms can recontract with new suppliers). The result is that the trade balance (net exports) often worsens in the immediate aftermath of a currency devaluation, when the main impact is to raise import costs, but then subsequently improves, leading to a trajectory that roughly follows a ‘J’ shape. Most analyses show that the J-curve turns upward within about 1–2 years, which is still within the short run for present purposes.

References

  • Ando A. Modigliani F. , '‘The “life-cycle” hypothesis of saving: aggregate implications and tests,’' ( 1963 ) 53 ( 1 ) American Economic Review : 55 - 84 .

  • Barbosa-Filho N.H. Taylor L. , '‘Distributive and demand cycles in the US economy: a structuralist Goodwin model,’' ( 2006 ) 57 ( 3 ) Metroeconomica : 389 - 411 .

  • Barrales J. von Arnim R. , '‘Longer run distributive cycles: wavelet decompositions for the US, 1948–2011,’' ( 2017 ) Review of Keynesian Economics forthcoming. .

  • Bhaduri A. Marglin S.A. , '‘Unemployment and the real wage: the economic basis for contesting political ideologies,’' ( 1990 ) 14 ( 4 ) Cambridge Journal of Economics : 375 - 393 .

  • Blecker R.A. , '‘International competition, income distribution and economic growth,’' ( 1989 ) 13 ( 3 ) Cambridge Journal of Economics : 395 - 412 .

  • Blecker R.A. , '‘Demand, distribution, and growth in neo-Kaleckian macro models,’', in M. Setterfield (ed), The Economics of Demand-Led Growth: Challenging the Supply-Side Vision of the Long Run , (Edward Elgar, Cheltenham, UK and Northampton, MA 2002 ) 129 - 152 .

  • Blecker R.A. , '‘Open economy models of growth and distribution,’', in E. Hein and E. Stockhammer (eds), A Modern Guide to Keynesian Macroeconomics and Economic Policies , (Edward Elgar, Cheltenham, UK and Northampton, MA 2011 ) 215 - 239 .

  • Blecker R.A. , '‘Long-run growth in open economies: export-led cumulative causation or a balance-of-payments constraint?’', in G.C. Harcourt and P. Kriesler (eds), The Oxford Handbook of Post-Keynesian Economics vol. I Theory and Origins , (Oxford University Press, Oxford 2013 ) 390 - 414 .

  • Blecker R.A. , '‘The US economy since the crisis: slow recovery and secular stagnation,’' ( 2016 ) 13 ( 2 ) European Journal of Economics and Economic Policies: Intervention : 203 - 214 .

  • Bunting D. , '‘Savings and the distribution of income,’' ( 1991 ) 14 ( 1 ) Journal of Post Keynesian Economics : 3 - 22 .

  • Carvalho L. Rezai A. , '‘Personal income inequality and aggregate demand,’' ( 2016 ) 40 ( 2 ) Cambridge Journal of Economics : 491 - 505 .

  • Chirinko R.S. Fazzari S.M. Meyer A.P. , '‘How responsive is business capital formation to its user cost? An exploration with micro data,’' ( 1999 ) 74 ( 1 ) Journal of Public Economics : 53 - 80 .

  • Chirinko R.S. Fazzari S.M. Meyer A.P. , '‘A new approach to estimating production function parameters: the elusive capital–labor substitution elasticity,’' ( 2011 ) 29 ( 4 ) Journal of Business and Economic Statistics : 587 - 594 .

  • Cogley T. Nason J.M. , '‘Effects of the Hodrick–Prescott filter on trend and difference stationary time series: implications for business cycle research,’' ( 1995 ) 19 Journal of Economic Dynamics and Control : 253 - 278 .

  • Cynamon B.Z. Fazzari S.M. , '‘Household debt in the consumer age: source of growth – risk of collapse,’' ( 2008 ) 3 ( 2 ) Capitalism and Society article 3 .

  • Cynamon B.Z. Fazzari S.M. , '‘The end of the consumer age,’', in B.Z. Cynamon, S.M. Fazzari and M. Setterfield (eds), After the Great Recession: The Struggle for Economic Recovery and Growth , (Cambridge University Press, New York 2013 ) 129 - 157 .

  • Duesenberry J.S. , Income Saving and the Theory of Consumer Behavior , (Harvard University Press, Cambridge, MA 1949 ).

  • Fazzari S.M. , '‘Monetary policy, financial structure, and investment,’', in G.A. Dymski, G. Epstein and R. Pollin (eds), Transforming the U.S. Financial System: Equity and Efficiency for the 21st Century , (M.E. Sharpe, Armonk, NY 1993 ) 35 - 63 .

  • Fazzari S.M. Hubbard R.G. Peterson B.C. , '‘Financing constraints and corporate investment,’' ( 1988 ) 1988 ( 1 ) Brookings Papers on Economic Activity : 141 - 206 .

  • Fernandez J. , Income Distribution and Aggregate Demand in the U.S. Economy: An Empirical Investigation , (unpublished PhD dissertation, American University, Washington, DC 2005 ).

  • Friedman M. , A Theory of the Consumption Function , (Princeton University Press, Princeton, NJ 1957 ).

  • Gordon R.J. and R. Krenn (2010) ‘The end of the great depression 1939–41: policy contributions and fiscal multipliers’ National Bureau of Economic Research (NBER) Working Paper 16380 (September).

  • Hein E. , The Macroeconomics of Finance-Dominated Capitalism – and its Crisis , (Edward Elgar, Cheltenham, UK and Northampton, MA 2012 ).

  • Hein E. , Distribution and Growth After Keynes: A Post-Keynesian Guide , (Edward Elgar, Cheltenham, UK and Northampton, MA 2014 ).

  • Hein E. , '‘Secular stagnation or stagnation policy? Steindl after Summers,’' ( 2016 ) 69 ( 276 ) PSL Quarterly Review : 3 - 47 .

  • Hein E. Vogel L. , '‘Distribution and growth reconsidered: empirical results for six OECD countries,’' ( 2008 ) 32 ( 3 ) Cambridge Journal of Economics : 479 - 511 .

  • Jappelli T. Pistaferri L. , '‘Fiscal policy and MPC heterogeneity,’' ( 2014 ) 6 ( 4 ) American Economic Journal: Macroeconomics : 107 - 136 .

  • Kalecki M. , Selected Essays on the Dynamics of the Capitalist Economy 1933–1970 , (Cambridge University Press, Cambridge, UK 1971 ).

  • Kiefer D. Rada C. , '‘Profit maximising goes global: the race to the bottom,’' ( 2015 ) 39 ( 5 ) Cambridge Journal of Economics : 1333 - 1350 .

  • Kim Y.K. , '‘Household debt, financialization, and macroeconomic performance in the U.S., 1951–2009,’' ( 2013 ) 35 ( 4 ) Journal of Post Keynesian Economics : 675 - 694 .

  • Kim Y.K. Setterfield M. Mei Y. , '‘A theory of aggregate consumption,’' ( 2014 ) 11 ( 1 ) European Journal of Economics and Economic Policies: Intervention : 31 - 49 .

  • Kim Y.K. Setterfield M. Mei Y. , '‘Aggregate consumption and debt accumulation: an empirical examination of US household behaviour,’' ( 2015 ) 39 ( 1 ) Cambridge Journal of Economics : 93 - 112 .

  • Lavoie M. , '‘The Kaleckian model of growth and distribution and its neo-Ricardian and neo-Marxian critiques,’' ( 1995 ) 19 ( 6 ) Cambridge Journal of Economics : 789 - 818 .

  • Lavoie M. , Post-Keynesian Economics: New Foundations , (Edward Elgar, Cheltenham, UK and Northampton, MA 2014 ).

  • M. Lavoie and E. Stockhammer (eds), Wage-Led Growth: An Equitable Strategy for Economic Recovery , (Palgrave Macmillan, Basingstoke, UK 2013 ).

  • López J. Sanchez A. Spanos A. , '‘Macroeconomic linkages in Mexico,’' ( 2011 ) 62 ( 2 ) Metroeconomica : 356 - 385 .

  • Marglin S.A. Bhaduri A. , '‘Profit squeeze and Keynesian theory,’', in S.A. Marglin and J.B. Schor (eds), The Golden Age of Capitalism , (Oxford University Press, Oxford 1990 ) 153 - 186 .

  • Minsky H.P. , Stabilizing an Unstable Economy , (Yale University Press, New Haven, CT 1986 ).

  • Nikiforos M. Foley D.K. , '‘Distribution and capacity utilization: conceptual issues and empirical evidence,’' ( 2012 ) 63 ( 1 ) Metroeconomica : 200 - 229 .

  • Obst T. Onaran Ö. , '‘Wage-led growth in the EU15 member states: the effects of income distribution on growth, investment, trade balance and inflation,’' ( 2016 ) Cambridge Journal of Economics advance access 28 April doi: 10.1093/cje/bew009. .

  • Onaran Ö. and G. Galanis (2012) ‘Is aggregate demand wage-led or profit-led? National and global effects’ Conditions of Work and Employment Series No 40 Geneva: International Labour Office.

  • Onaran Ö. Stockhammer E. Grafl L. , '‘The finance-dominated growth regime, distribution, and aggregate demand in the US,’' ( 2011 ) 35 ( 4 ) Cambridge Journal of Economics : 637 - 661 .

  • Palley T.I. , '‘Debt, aggregate demand, and the business cycle: a model in the spirit of Kaldor and Minsky,’' ( 1994 ) 16 ( 3 ) Journal of Post Keynesian Economics : 371 - 390 .

  • Palley T.I. , '‘The relative permanent income theory of consumption: a synthetic Keynes–Duesenberry–Friedman model,’' ( 2010 ) 22 ( 1 ) Review of Political Economy : 44 - 56 .

  • Palley T.I. , '‘Enriching the neo-Kaleckian growth model: nonlinearities, political economy, and q theory,’', in S. Dullien, E. Hein and A. Truger (eds), Macroeconomics Development and Economic Policies: Festschrift for Jan Priewe , (Metropolis-Verlag, Marburg 2014 ) 71 - 84 .

  • Palley T.I. , '‘Wage- vs. profit-led growth: the role of the distribution of wages in determining regime character,’' ( 2016 ) Cambridge Journal of Economics advance access 28 March doi: 10.1093/cje/bew004. .

  • Pérez Caldentey E. and M. Vernengo (2013) ‘Wage and profit-led growth: the limits to Neo-Kaleckian models and a Kaldorian proposal’ Levy Economics Institute Working Paper No 775 September.

  • Razmi A. (2014) ‘Re-distribution aggregate demand and growth in an open economy: the crucial interaction of portfolio considerations and external account constraints’ University of Massachusetts Amherst Working Paper 2014-07 (November).

  • Robinson J. , '‘The pure theory of international trade,’' ( 1946–1947 ) 14 Review of Economic Studies : 98 - 112 .

  • Robinson J. , Essays in the Theory of Employment , (Basil Blackwell, Oxford 1947 ).

  • Sasaki H. Fujita S. Sonoda R. , '‘International competition and distributive class conflict in an open economy Kaleckian model,’' ( 2013 ) 64 ( 4 ) Metroeconomica : 683 - 715 .

  • Schoder C. , '‘Credit vs. demand constraints: the determinants of US firm-level investment over the business cycles from 1977 to 2011,’' ( 2013 ) 26 North American Journal of Economics and Finance : 1 - 27 .

  • Setterfield M. Cornwall J. , '‘A neo-Kaldorian perspective on the rise and decline of the golden age,’', in M. Setterfield (ed), The Economics of Demand-Led Growth: Challenging the Supply-Side Vision of the Long Run , (Edward Elgar, Cheltenham, UK and Northampton, MA 2002 ) 87 - 114 .

  • Skott Peter , '‘Weaknesses of “wage-led growth”,’', in Review of Keynesian Economics , ( 2017 ) forthcoming .

  • Spatareanu M. , '‘The cost of capital, finance and high-tech investment,’' ( 2008 ) 22 ( 6 ) International Review of Applied Economics : 693 - 705 .

  • Steindl J. , Maturity and Stagnation in American Capitalism , (Monthly Review Press, New York 1952 [1976] ).

  • Stiglitz J.E. Weiss A. , '‘Credit rationing in markets with imperfect information,’' ( 1981 ) 71 ( 3 ) American Economic Review : 393 - 410 .

  • Stockhammer E. Michell J. , '‘Pseudo-Goodwin cycles in a Minsky model,’' ( 2016 ) Cambridge Journal of Economics advance access 27 April doi: 10.1093/cje/bew008 .

  • Stockhammer E. Stehrer R. , '‘Goodwin or Kalecki in demand? Functional income distribution and aggregate demand in the short run,’' ( 2011 ) 43 ( 4 ) Review of Radical Political Economics : 506 - 522 .

  • Stockhammer E. Wildauer R. , '‘Debt-driven growth? Wealth, distribution and demand in OECD countries,’' ( 2015 ) Cambridge Journal of Economics advance access 27 November doi: 10.1093/cje/bev070 .

  • Stockhammer E. Onaran Ö. Ederer S. , '‘Functional income distribution and aggregate demand in the Euro Area,’' ( 2009 ) 33 ( 1 ) Cambridge Journal of Economics : 139 - 159 .

  • Stockhammer E. Hein E. Grafl L. , '‘Globalization and the effects of changes in the functional income distribution on aggregate demand in Germany,’' ( 2011 ) 25 ( 1 ) International Review of Applied Economics : 1 - 24 .

  • Storm S. Naastepad C.W.M. , Macroeconomics Beyond the NAIRU , (Harvard University Press, Cambridge, MA 2012 ).

  • Vargas Sánchez G. Luna A. , '‘Slow growth in the Mexican economy,’' ( 2014 ) 37 ( 1 ) Journal of Post Keynesian Economics : 115 - 133 .

  • Von Arnim R. Tavani D. Carvalho L. , '‘Redistribution in a neo-Kaleckian two-country model,’' ( 2014 ) 65 ( 3 ) Metroeconomica : 430 - 459 .

Affiliations

Blecker, Robert A. - Department of Economics, American University, Washington, DC, USA