The Taylor Rule and its Aftermath: An Interpretation Along Classical-Keynesian Lines

ABSTRACT

The aim of this paper is to assess to what extent the Taylor rule can be considered an appropriate representation of the tendency of central banks to react to inflation. After an overview of the origin and use of the Taylor rule, the paper stresses some difficulties in its implementation and the limits of its interpretation by the New Consensus models. Specifically, the inherent difficulties stemming from the notion and estimates of a benchmark interest rate determined by ‘productivity and thrift’ are pointed out. We then move on to advance an alternative interpretation of the Taylor rule along Classical-Keynesian lines. In this context, inflation is fuelled by conflicting claims on income distribution and the rule will be interpreted, as it is in actual fact, as a flexible and non-mechanical benchmark for monetary policies which will be seen to affect the division of product between wages and profits.

KEYWORDS:

• Monetary policy
• Taylor rule
• cost-push inflation

JEL CODES:

• E11
• E12
• E52
• E58

Acknowledgements

I wish to thank prof. Franklin Serrano and two anonymous referees for their suggestions on a previous version of this paper. The usual caveats applies.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Notes

1 Since in this case the constant term is given by $c=r^{\ast }+\left(1-{\vartheta }_{\pi }\right){\pi }^{\ast }$, when the value of the intercept of the estimated line with the ordinate axis is greater than 2% (and therefore, ${\vartheta }_{\pi }$ is lower than 1) there is bound to be increasing inflation since the real interest rate will not increase when the nominal interest rate increases.

2 In some models the interest sensitive element of aggregate demand is only consumption. In this case, $\sigma$is the intertemporal elasticity of substitution.

3 More precisely, it is ‘short run’ money endogeneity. In these models, the monetary policy instrument is the rate of interest and central banks accommodate the monetary base. However, in the long run, credit and money supply are bounded by the tendency of the interest rate towards its natural value. See Rochon (1999).

4 On this point, see Section 4 and Reifschneider, Wascher, and Wilcox (2013).

5 For instance, according to Curdia (2015), the natural rate became negative after 2008.

6 However, see Garegnani (1978–79) on some weaknesses in Keynes’s arguments due to the presence in his theory of traditional elements.

7 It should also be noted that multiple equilibria by itself would undermine the idea of the natural interest rate as a benchmark rate for monetary policy. In this case, in fact, central banks would not only be uncertain over its estimates; even if they knew the possible equilibria, they would also have to select the one the policy interest rate should be moved towards based on subjective ordering of these equilibria.

8 For an analysis of why these criticisms apply also to the New-Keynesian models although they refer to intertemporal o temporary equilibria, see Dvoskin and Petri (2017). Indeed, great ambiguity is present in these models with respect to the notion of capital. For example, Woodford (2003, pp. 167 and 353) refers to short run capital goods as specific inputs with different real rental rates. However, in these models, capital is often introduced into the production function as a single magnitude, firms are said to borrow at a rate equal to the value of the marginal product of capital, and Woodford himself treats capital as a single value magnitude when writing that there is a tendency of it to shift towards sectors with higher capital returns and that investments goods are perfect substitutes for savers (Woodford 2003, p. 166). As in the case of Marshall’s short run equilibrium or Hicks’s temporary equilibrium, a notion of capital as a single magnitude also reappears in the investment functions with adjustment costs that in the New-Keynesian models drive the out-of-equilibrium process of convergence of the ‘market’ interest rates towards the natural rate. It is also implicit in the assumption of a single benchmark natural interest rate independent of the numeraire chosen.

9 While the wealth channel is usually estimated as weak (see for instance Lettau, Ludvigson, and Steindel 2002), the housing market is an important channel of monetary policy.

10 The presence of the Gibson paradox in a fiat money economy has been widely recognised in the literature on the price puzzle. For a discussion of this literature, see Cucciniello, Deleidi, and Levrero (2022).

11 Rewriting equation [1] as a backward-looking Phillips curve ${\pi }_t=\frac{1}{\lambda }\left[q_t-q_L\right]+h{\pi }_{t-1}+{ϵ}_t$, assuming that cost shocks averaged out $\left(E\left({ϵ}_t\right)=0\right)$and h<1, we will have $\pi =\frac{\left[q_t-q_L\right]}{\lambda \left(1-h\right)}$ when the expected and actual inflation rates ${\pi }_{t-1}$ and ${\pi }_t$are equal. As we will see below, the possibility of accelerating inflation would thus stem only from an increasing cost-push inflation. In its absence, we may even have a flat Phillips curve. The possibility of a non-vertical long run Phillips Curve is nowadays recognised. See, for instance, Akerlof et al. (2000).

12 The normal profits of enterprise are taken as uniform but usually differ among industries. They are affected by top-managers’ remuneration, the degree of liquidity of productive investments in the various sectors and their ‘real or fancied risk’. They can also embody oligopolistic elements, as affected also by the period of validity of the patents rewarding the risk and trouble of developing and applying new knowledge (Pivetti 1991, p. 32).

13 Unlike the New-Keynesian models, here the real mark-up is the result of the nominal mark-up and the inflation rate.

14 We abstract here from the fact that the increase in money wages may not be uniform across sectors and that in the short-run the market prices may diverge from the natural ones for factors other than the initial adjustment to the historical costs of capital.

15 For the sake of simplicity, we use this formulation which derives from the accounting of the price as determined by the cost of labour per unit of product plus a margin that covers all other costs and profits on capital advanced in production.

16 Two cases correspond to $\dot{\mu }>0$. The first pertains to the adjustment in the nominal rate of interest needed to leave the real interest rate unchanged when a continuous increase in money wages occurs. The second case concerns “autonomous” changes in the nominal mark-up on prices. This will lead to one-off increases in prices that are greater than money wages, and therefore to a rise in the prices-to-money wage ratio.

17 The parameter h could also reflect changes in factors affecting the bargaining position of workers other than the unemployment rate. However, these factors already affect the ‘aspiration gap’ and therefore the value of the unemployment rate $U_{\mathrm{na}}.$This gap, together with the expected inflation rate by workers ${\pi }_{t-1}$, affects the increase in money wages during the time period t. In turn, the latter and the value of h affect the inflation rate in t.

18 The analysis carried out so far refers mainly to a closed economy but can easily be extended to an open economy taking into account the effects on domestic prices of changes in the exchange rates.

19 In this case, in fact, there will no need to further change the nominal interest rate in order to ensure a given targeted real interest rate. Therefore, with $\mathrm{h}=0,{\pi }_t={\pi }_{t-1}={\pi }^a$, where ${\pi }^a$ is the expected inflation rate.

20 The pursuit of a higher real interest rate may affect the workers' “aspiration gap” and therefore the same process towards the targeted inflation rate which ultimately rests on the direct and indirect effects of a higher unemployment rate on workers' bargaining power and on what is the final targeted real interest rate by the central bank.

21 Lavoie (2014) calls it the Post-Keynesian Phillips Curve.