The Effects of Nominal Monetary Changes on Output, Employment and Inflation

Summary

In Chapter 12, the subject of the division of monetary induced expenditure changes into price, output and employment effects was largely ducked. This topic is addressed in the present chapter. A general theme, throughout the chapter, is the consideration of the imperfections that may allow a deviation of actual from potential (equilibrium) output to occur, leading to unused capacity and involuntary unemployment. In a world of certainty, there would be no undesired unemployment or disequilibria arising from incorrect and inconsistent decisions. In the real world, there is neither the information nor the marketing mechanisms (e.g., a Walrasian auctioneer) available to allow equilibrium to be achieved. In a complex, decentralised economy, in which individual decision-makers can only conjecture about the likely actions of others in the economy, there is much scope for inconsistent decision-making (e.g., by investors and savers) to lead to disequilibria. The institution of money provides the information network which enables a complex, decentralised economy to function at all. Since money is a necessary adjunct to such an economy, the disequilibria are sometimes regarded as monetary phenomena; it is, however, the inconsistency of decisions within the system, not the existence of the monetary framework, which is the proximate cause of disequilibria.

As already argued, price adjustments are not likely to be perfectly flexible and the resulting stickiness of prices will cause monetary policy changes, even when anticipated, to have real effects. Even if prices and wages did fall flexibly in response to excess capacity, this might not, however, stimulate a depressed economy if it generated expectations of further price falls. The process of sequential decision-making in modern economies tends to produce autocorrelated price movements. It may be more realistic to expect the rate of change (rather than the level) of prices to remain constant. The simple relationship between price inflation and the extent of spare capacity has thus to be adjusted to take the anticipated rate of inflation $(E\dot P)$ into account, as expressed in the following relationship:

$${\dot P_t} = {b_0} + {b_1}\frac{1}{{U{E_t}}} + {b_2}E\dot P$$

where UE = vunemployment level

In theory, b ₂ should equal unity. If $E\dot P = {\dot P_{t - 1}}$, then the rate of inflation will accelerate, stabilise or decline as the sum of

$${b_0} + {b_1}\frac{1}{{U{E_t}}}$$

is greater, equal or less than zero. This formulation suggests that only in the short run can an advantage in higher real output and real incomes be obtained from running the economy at an inflationary pressure of demand. In the longer run, inflation would accelerate without limit, the long-run Phillips curve being vertical, and no advantage in higher real output would remain to offset the worsened rate of inflation. Earlier experience had, however, suggested that wage — price adjustments, even in the longer run, were rather less volatile than this analysis would suggest; the value of b ₂ calculated in empirical studies using data from the 1950s and 1960s was usually well below unity, often in the region of 0.5. In the 1970s, however, inflation accelerated throughout the Western world, in spite of a higher average level of unemployment in many countries. This partly reflected the evaporation of the last vestiges of money illusion, a growing consciousness of inflation. The calculated value of b ₂ in empirical studies then rose to unity, or sometimes even higher.

We next turn in Section 2 to consider both recent historical experience and theoretical developments with the Phillips curve inter-relationships between nominal demand, output and inflation. The main practical problem has been that the estimated position of the vertical Phillips curve, or Non-Accelerating-Inflation Rate of Unemployment (NAIRU), has tended to shift rightward — i.e., to involve a higher level of unemployment over time; in fact, the estimated rightward shifts of the NAIRU have quite closely tracked the movements of actual unemployment. We examine various suggestions to explain why this might have happened — i.e., that those with existing jobs (the insiders) adjust their wage claims in response to potential changes in unemployment, not to its level, and that the experience of unemployment may impair the human capital of the unemployed (i.e., a hysteresis effect). Whatever the explanation may be, such variability in the NAIRU further undermines any case for seeking to run the economy at some pre-selected pressure of demand or level of unemployment and it also raises questions about the usefulness of the concept itself.

The main theoretical interest has been the application of rational-forward looking expectations to this analysis. So long as $(E\dot P)$ was based on backward-looking-adaptive expectations, then current monetary policy would affect real output, even if only temporarily. But if expectations are forward-looking, agents will adjust their price expectations to take account of anticipated monetary policy effects, leading to the Sargent-Wallace policy impotence theorem. This suggests that unanticipated monetary changes should affect output, whereas anticipated monetary changes should not do so; empirical tests of this hypothesis, begun by Barro, have had mixed effects.

Superimposing forward looking rational expectations onto the augmented Phillips curve leads on, however, to a Lucas-type functional relationship between deviations of actual from equilibrium output on the one hand and errors (surprises) between current inflation and the rate of inflation expected for the period on the other. In its simplest form, the theory suggests that such deviations should be serially uncorre-lated; this is implausible, however, because most real variables exhibit persistence in their cyclical fluctuations around their trends. Various amendments and supplements to the neo-classical model have been proposed to account for such persistence.

According to one neo-classical model, deviations from equilibrium are caused by imperfections in the transmission of information on certain key variables (e.g., national price trends or monetary growth). This seems implausible in a society where such information is rapidly and widely available at little cost to agents. In another version of the neo-classical model, the ‘real’ business cycle model, fluctuations are caused by stochastic technological shocks impacting on a system that remains in equilibrium: this view is not easy to square with observed unemployment. A much more likely source of market imperfection is to be found in price stickiness; such price stickiness would provide an alternative, and on this view more sensible, explanation of real-world macro-economic phenomena.

Finally, in the Appendix, we examine a case study of indexation, the UK experiment with the provision of indexed gilts from 1981. We start by considering the costs of unanticipated inflation and how indexation can theoretically reduce such costs. Against that background, we ask why practical politicians and Central Bankers have been so averse to the introduction of indexation. Then, in the second part of this Appendix, we record the various arguments that were deployed in the discussions leading up to the introduction of indexed gilts and briefly recount the resulting experience with the instrument.