The Comparative Statics of Effective Demand

ABSTRACT

Keynes introduces the term ‘effective demand’ in Chapter 3 of the General Theory as designating the point of intersection of two functions: the ‘aggregate demand function’ (D) and the ‘aggregate supply function’ (Z). For the first time in the literature, I here use specific functional forms for the D and Z functions and run numerical simulations which allow study of the comparative statics of the model in the face of various ‘shocks’. The demonstration of how the D/Z model actually works will hopefully prove useful for future students of the economics of Keynes.

KEYWORDS:

• Effective demand
• D/Z model
• Keynes

JEL CODES:

• B31
• E12

Acknowledgements

The article is based on KOF Working Paper No. 355, Zurich, March 2014. My contribution (in German) to the Festschrift for Jan Priewe (Hartwig 2014) also draws on that Working Paper and has thus some overlap with the present article. I would like to thank William Darity, Jr. and the reviewers of this journal for their excellent comments on earlier versions of this article. All remaining errors are mine.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1 See Weintraub (1958), Davidson and Smolensky (1964) and Davidson (1978, 1994, 2002). King (1994) scrutinizes the early stages of non-Post Keynesian aggregate supply and demand analysis. ‘To conclude that there was some confusion about aggregate supply and demand analysis in the early 1950s would be a grotesque understatement,’ he sums up (p. 14).

2 A recent symposium in the Review of Political Economy tried to find some common ground; see Allain, Hartwig, and Hayes (2013), Allain (2013), Hartwig (2013) and Hayes (2013).

3 There exists a literature (including Chick 1983; Galbraith and Darity 2005; Lawlor 2008) that has formalized Keynes’s D/Z model. But never before have specific functional forms for the D and Z functions been used and shocks to model equilibria been simulated.

4 This has been questioned by Lavoie (2003), for instance.

5 Hagemann (2010) recollects that the D/Z model—which he calls the ‘fish diagram’—had a very limited appeal for students on Paul Davidson’s course, ‘Money and Banking’, at the Graduate Faculty of the New School for Social Research in 1986 when Hagemann was a visiting professor there. Improving the didactics of that model is thus an important task.

6 Although Keynes subtracts what he calls ‘user cost’—the sum of intermediate consumption and depreciation allowances—from gross output in the aggregate (see Keynes 1936, pp. 23–24), for practical and pedagogical purposes gross value added is close enough to Keynes’s concept of income.

7 Some (post-Keynesian) economists, e.g., Shaikh (1974), Felipe and Fisher (2003) and Felipe and McCombie (2006), have criticized aggregate production functions, mainly for including ‘K’ as a measure of physical capital despite known measurement problems and the Cambridge (UK) ‘capital critique’. Keynes, who sympathized with ‘the pre-classical doctrine that everything is produced by labour’ (Keynes 1936, p. 213, emphasis in the original) and therefore regarded ‘labour …  as the sole factor of production’ (pp. 213–214), did not include physical capital in his production function. The latter is therefore not much affected by the above-mentioned critique.

8 Palley (1997) and others have pointed out that Keynes’s acceptance of the ‘first classical postulate’ (Keynes 1936, pp. 17–18) implies the adoption of the neo-classical supply-side assumptions of free competition, price-taking, profit-maximization and decreasing marginal returns to labor. I honor these assumptions in what follows.

9 With Π = aggregate profit, P^s = aggregate supply price level, Y = gross value added, N = employment, w = wage unit (= average nominal wage rate; see Keynes 1936, p. 41).

10 Davidson and Smolensky (1964, pp. 125, 134–135) discuss the connection between the wage share and the slope of Z and show that Z is linear for the production function Y = N^α, which I assume below.

11 Keynes’s notion of price-taking departs from the strict microeconomic theory of the small firm operating under perfect competition. That theory would not allow for entrepreneurs forming ex-ante expectations about demand. Keynes—who was concerned with the real world—did not have such firms in mind. In his theory, firms are not ‘atomistic’, but also not powerful enough to dictate the price. They have to form expectations about the price for their products the market will accept and about the market share that might be attributable to them (see Chick 1992).

12 In his essay ‘Relative Movements of Real Wages and Output’, Keynes (1939) relaxed the assumption of decreasing marginal returns to labor in the short period. In a companion paper (Hartwig 2017), I discuss possible reasons that might have prompted Keynes to relax his core assumption from the General Theory and analyze the consequences of assuming non-decreasing marginal returns to labor for the model of effective demand.

13 It was demonstrated in Section Two that the slope of Z in wage-units (Z_w) is equal to the inverse of the output elasticity (α), i.e., equal to 1/α. So the slope of Z is equal to w/α. Because of the normalization of the money wage rate to 1, we have the special case that the slope of Z is the same as that of Z_w.

14 The D/Z model refers to the Marshallian short period for which the impact of investment on the capital stock is disregarded.

15 It is the entrepreneurs in the consumption-goods sector who have to form expectations about the level of investment spending in order to calculate how much demand will be forthcoming to them through the multiplier mechanism (see Hartwig 2004). We can therefore regard the principle of effective demand as a model for the expectation formation of consumption-goods producing entrepreneurs who take a certain nominal investment level for granted. In what follows, I use the expression P^d for the expected price level of consumption goods.

16 Note that Keynes does not indicate that rising prices raise output and employment because they lower the real wage rate. They are a ‘symptom’ not a cause of rising output and employment.

17 In Figure 4 it is assumed that the elasticity of long-term expectations with respect to the nominal wage rate equals zero. If this elasticity is positive, the rise in the nominal wage rate by 10 percent (from 1 to 1.1) raises (expected) investment demand (above 20); and the D curve shifts further to the top. If the elasticity is negative, however, the D curve shifts downward.

18 The nominal wage rate is not explained by the model of effective demand; it is an exogenous variable (see Weintraub 1978/79).

19 An improved organization of the production process might cause such a shock.