RDP 9007: Operating Objectives for Monetary Policy 1. Introduction

The shift away from monetary targeting in the 1980s reflected a belief that demand for the monetary aggregates had become highly unstable. This was unfortunate, because simple monetary growth rules could be claimed to have a number of attractive features, at least in theory. In particular, it was argued that such rules could provide an anchor for price expectations in the medium to long term, as well as providing a reasonable degree of short-term stabilisation against exogenous shocks (excluding shocks to money demand). The abandonment of strict monetary targeting in many countries has raised the question as to whether alternative operating principles can be devised which have similar properties but are not subject to the same problems of short-term instability.

Goodhart (1989) notes that two main alternatives to monetary targeting have been pursued by different groups of countries. The first, applicable mainly to the smaller European countries, has been to move towards fixed exchange rates, effectively bypassing domestic problems of monetary instability by fixing to a suitably stable external currency. For many countries, however, this is not an attractive option because of their vulnerability to external shocks, such as shocks to the terms of trade. As argued recently by Blundell-Wignall and Gregory (1989) Australia falls fairly clearly within this group. The second alternative described by Goodhart is the use of an interest rate as the operating instrument of policy, with policy being aimed directly at stabilisation of ultimate objectives, rather than being defined in terms of a monetary target. Although this formulation is somewhat imprecise, Freedman (1989) has suggested that, for countries with floating exchange rates, this would become the dominant model for thinking about monetary policy in the 1990s. This view is endorsed by B. Friedman (1989) who notes that although such an approach is a logical consequence of money-demand instability, it has so far been given relatively little formal analysis.

The aim of this paper is to provide a systematic discussion of issues arising under this second approach to policy, using a formal theoretical model intended to capture some of the important features of the Australian economy. In particular, it is assumed that money demand is not sufficiently stable to serve as a basis for policy, and that the economy is subject to real shocks arising both domestically and externally. Operating rules for policy are designed so as to satisfy two criteria:

• long-run anchoring of inflation at a targeted rate;
• satisfactory short-run properties with respect to the stabilisation of prices and output.

Rather than attempt to derive general expressions for optimal policy, which would inevitably result in a rather uninformative “look at everything” conclusion, the paper proposes three simple policy rules and compares their properties using the performance criteria mentioned above.^[1] The three rules are nominal income targeting, targeting of consumer prices and, as a benchmark for comparison, a fixed exchange rate.^[2]

A number of arguments could be given in defence of this focus on simple rules, and two in particular seem worth emphasising. First, simple rules are likely to be more widely understood than complex rules, and may therefore promote accountability and credibility of targets.^[3] Secondly, simple rules have the advantage of imposing relatively small information requirements on the monetary authorities and are less likely to be dependent on detailed knowledge of the dynamics of the economic system. These arguments were important in earlier advocacy of simple money growth rules, for example by Friedman (1968).

The discussion that follows is divided into three parts:

• the need for a nominal anchor;
• short-run stabilisation properties of alternative rules;
• the information requirements of monetary policy.

Footnotes

The analysis here extends earlier work on the closed economy case by Edey (1989). It is also related to work by Blundell-Wignall and Gregory (1989), who examined optimal policy reactions to terms-of-trade and money-demand shocks, while retaining the assumption that the instrument of policy is a monetary quantity. Other related work by Aizenman and Frenkel (1986) and by Alogoskoufis (1989) uses a similar method to that employed here, but does not specifically cover terms-of-trade and money-demand shocks. [1]

Nominal income targeting has been studied by Bean (1983) and Taylor (1985), amongst others. Adoption of a formal price level target was recently advocated by Gavin (1990). [2]

This argument has been formalised in the literature on reputational equilibria, for example, by Barro and Gordon (1983). [3]