Abstract

Inflation targets have become popular among monetary theorists and central bank practitioners leading to their adoption in a number of cases where the alternatives of constraining central bank behaviour through targets for monetary aggregates and exchange-rate pegging have appeared unappealing. The paper focusses on the latter alternative and on a combined strategy, particularly for countries which have the option of joining a monetary union. It also looks at the role of an inflation target for the joint monetary policy of the future European Central Bank.

JEL classification: E 42, 52, F 33

Key words: Monetary policy, inflation targets, exchange-rate policy.

Author’s e-mail address: okonth@pc.ibt.dk
 
 

1. Introduction

Inflation targeting as a framework for monetary policy has become popular among theoretical economists as well as central bank practitioners over the past decade.¹ Among the industrial countries four countries pioneered in the practice of inflation targets in the late 1980s and early 1990s (New Zealand, Canada, the United Kingdom and Sweden) and they have been joined by Finland, Australia and Spain and Italy more recently. In preparing for European monetary union (EMU) the central bankers of the likely participating countries have also examined the case for a joint monetary strategy based primarily, or at least partly, on an inflation target.

While the general arguments in favour of or against the adoption of such a strategy have been well rehearsed in the academic and central banking literature, balancing them against one another has usually been inconclusive. This is no surprise, since conclusions depend on the alternative strategy as well as on the particular variant of an inflation target considered. In the United States the academic discussion has focussed on a comparison of inflation targets and a strategy combining inflation and output variability, the so-called Taylor Rule, concluding in some recent contributions that attention in practice to stable, low inflation may have led to undesirably large fluctuations in real output, see notably Ball (1997) and Fuhrer (1997). In European countries where the choice has appeared to be largely between sticking to one (or more) monetary aggregate(s) as an intermediate target and the adoption of an explicit inflation target, the debate has resumed that of the 1970’s when the former strategy was adopted in a number of industrial countries; and some recent research, notably Bernanke and Mihov (1997), has argued that even the central bank which has most strongly continued to argue in favour of a monetary aggregate target, the Bundesbank, has come close to de facto targeting inflation. Finally, in countries participating in the European Union, the debate has centred on comparing inflation targeting to another form of intermediate targeting, viz. of the exchange rate. Such a comparison is natural, because monetary policy in most EU countries outside Germany was for a decade from 1983 (and again recently) directed primarily at defending a fixed exchange rate vis-á-vis the Deutsche Mark (DEM). Although the United Kingdom and Sweden broke out of their - much more short-lived - commitments to a stable exchange rate in 1992 and have been floating for the past five years, the ambition of all other EU countries remains that central rates vis-á-vis the euro for EU currencies which do not participate in the monetary union scheduled to start on 1 January 1999 should be declared no later than the date.

The present paper has a deliberately restricted focus relative to this wide- ranging set of issues. It asks, in particular, whether the pressure on the United Kingdom and Sweden, which have abandoned exchange-rate targeting in favour of inflation targeting - and have done well in this "new" regime - to return to the former strategy is justified. It further asks - still more superficially - what role an inflation target for the future monetary union could play; here the choice is not between this strategy and an exchange-rate target, but rather whether an explicit inflation target could stand alone.

The paper does not present new results, but aims to survey, with the help of recent academic contributions, answers to one important issue where monetary theory does have a contribution to make in designing monetary policy. However, that contribution is often not itself sufficient to arrive at conclusions; political economy considerations are very important as should become evident.

The rest of the paper is structured as follows: Section 2 looks at the remarkable conversion to nominal targets for monetary policy in Europe over the past two decades. Section 3 reviews the relative merits of inflation targets normally stressed by their advocates over other types of target-constrained policies, notably pegged exchange rates: they are less vulnerable to speculative crises in the foreign exchange market and to supply shocks in the area to whose currency the domestic currency is pegged. It further looks at the argument that an inflation target facilitates communication with the public and the related issue of how it may affect credibility. This section also looks at more operational issues. Given the long time lags with which monetary instruments influence inflation, policy will normally have to rely on supplementary indicators, one of which is the exchange rate, the more so the more open is the economy. Again the cases of Sweden and the United Kingdom can be used as illustrative examples, supplemented by a reference to the so-called monetary conditions index, pioneered by the Bank of Canada. Section 4 looks at the arguments in favour of prime reliance on an inflation target for the joint monetary policy to be conducted by the future European Central Bank. Section 5 brings together the conclusions.
 

2. The European shift to nominal objectives for monetary policy

There has been a profound and remarkable shift in European thinking about the appropriate strategy for monetary policy over the past 25 years. Prior to the high inflation of the early and mid-70s monetary policy was seen as a potentially important contributor to an active demand management policy; empirical research seemed to underpin that view, see e.g. OECD (1975). The stagflation of the mid-70s removed any confidence that monetary policy could be used for the purpose of shorter-term stabilization and led to the introduction at various times from 1974 onwards of so-called "target-constrained" monetary policies, initially formulated in terms of a selected monetary aggregate as an intermediate objective. The main purposes of this shift were to bring inflation down and to do so with the help of an objective closer to the influence of central banks than inflation itself.

Already at that time, or from the late 1970s, many European countries attached great importance to intra-European exchange rates. Until 1983 the ambition to preserve a regional system of fixed-but-adjustable exchange rates, after 1979 in the European Monetary System (EMS), could not be seen as a genuine part of a monetary strategy. The EMS was fairly lax, permitting realignments relatively frequently; hence it could not enforce stable, low inflation throughout its member-ship, although it may have prevented national inflation rates from diverging further. The EMS operated basically as a system for keeping the competitiveness of participants broadly in line, see Gros and Thygesen (1997). Since such a system , as is well known, leaves the long-run price level indeterminate, the need for national nominal anchors was still present.

From the French decision to stay in the EMS in 1983 the System entered a disciplinary phase. The aim became primarily to stabilize inflation at a low rate. Better convergence in price performance with Germany became a prime objective of policy and the fixed DM-exchange rate was the primary instrument for Germany’s partners. Domestic monetary aggregates were allowed to fade into the back-ground, and monetary policy was seen to "borrow credibility" from the Bundesbank. This policy of substituting an increasingly rigid peg to a low-inflation currency for a domestic monetary anchor worked well for nearly a decade. For the period up to 1989-90 Germany provided a demanding and stable anchor; and despite a rise in the German inflation following unification, most of Germany’s partners proved anxious to continue to peg rigidly to the DM until 1992-93.

The key notion in European monetary policies in the decade 1983-92 became credibility. The central bankers, not surprisingly, constituted a receptive audience for the new doctrine of the potential inflation bias in a discretionary monetary policy and for the notion of a commitment technology for locking policy-makers into a steady non-inflationary performance. Although there were few moves afoot already in the 1980’s to give European central banks more formal independence of their respective governments, the tighter EMS and increasing capital mobi-lity made each central bank more dependent on their colleagues in the system, requiring greater freedom of action in relation to their own governments. While this was granted, the limits to ex ante coordination of policies were also becoming very visible. Only a change in the constitutional basis for central bank cooperation, i.e. a change in the Rome Treaty, could modify this and eliminate the congenital fra-gility of a system of fixed-but-adjustable exchange rates. Such a change was ultima-tely proposed as an essential element in the Maastricht Treaty of 1991 and ratified two years later. However, this change will only take effect when the first group of participants in monetary union has been identified, presumably in the spring of 1998, enabling them to conduct a genuinely joint policy in the European Central Bank. But until that time a tight EMS was seen by the framers of the Treaty as an essential element in the strategy for achieving both monetary unification and low inflation throughout a group of prospective participants.

The weaknesses of this strategy were exposed by the 1992-93 turmoil in foreign exchange markets in Europe, illustrating in succession two dangers inherent in fixed-but-adjustable rate systems. One danger is the readiness of participating governments to defend even misaligned rates and then having to give them up; the other is the proclivity of market participants to seize upon any sign of tension even for currencies whose fundamentals are in good order. The two dangers are related; if government and central banks have to give up their defence and "allow" an EU-country to depreciate, future speculative attacks on other currencies in the system become more likely. The two cases can also be examined in the light of so-called "first"- and "second-generation" models of speculative attacks. In the former model there are grounds for suspecting divergent monetary and/or fiscal behaviour in the country concerned, but the exchange-rate continues to be defended by the central bank and its partners as long as the country’s international reserves permit. When the EMS failed ultimately in defending the central rates of the Italian lira and the British pound and accepted or even encouraged devaluations of several other currencies in the few months following September 1992 that was seen by financial markets as an invitation to attack the remaining currencies in the system. In August 1993 the margins of the fluctuations were dramatically widened and although central rates among the remaining currencies were maintained, the role of exchange-rate targets was reduced.

This shift was emphasized not least by countries who had left the EMS (Italy, the United Kingdom) or had been unable to sustain their unilateral link to it (Sweden, Finland). By adopting inflation targets as the main basis for their monetary policies these countries intended to indicate to sceptical financial markets and the domestic public that they were determined not to allow large devaluations to feed into domestic prices. They were remarkably successful - to an extent so that there has recently been no clear difference in inflation behaviour between those who stayed in the EMS and those who devalued in 1992-93 with or without leaving the system. It should be noted, however, that this remarkable inflation convergence has not yet been tested in most countries by a sustained long upswing taking actual GDP to or above its estimated long-run growth; except in the United Kingdom growth has been moderate in the 1995-97 period.

The revival of monetary policy targets in the form of quantified inflation objectives is encouraging from the viewpoint of inflation performance and conver-gence in Europe. Yet it is incomplete in the perspective of the plans for monetary union by 1999. The Maastricht Treaty, in establishing the criteria for admission to monetary union stressed not only an inflation objective - the inflation rate for the most recent year should be no more than 1½ percentage points above the average performance of the three countries with the lowest inflation rate - but also an exchange-rate objective: the candidate country’s currency should have participated in the EMS for a two year period, having managed their currency without tension and devaluation (at their own initiative). These criteria have been recently confirmed to apply also to countries who join monetary union later; they will be expected to join a successor arrangement to the EMS in which outside currencies will be pegged to the now single currency, the euro, within wide fluctuations on margins similar to those of to-day.

In this context there is, in other words, insistence on compliance with both an inflation and an exchange-rate target as entry requirements. Two countries, Sweden and the United Kingdom, have argued that the exchange-rate criterion should be regarded as superfluous after the erosion of the EMS rules in 1993 and that it might therefore be suspended. The 13 other present Member States have so far taken the line that the exchange-rate criterion remains valid and valuable. The issue is unlikely to come to a head in the near future, since neither the United Kingdom nor Sweden appears to be an early candidate for monetary union; and the two other countries which had given up their exchange-rate peg in 1992 and stuck to an inflation target while improving their price performance - Finland and Italy - significantly rejoined the EMS in the autumn of 1996 in order to revive their cre-dentials for monetary union.

The next section of this paper looks at the justification for maintaining an exchange-rate target in addition to an inflation target even when the latter is successfully implemented in a way that has assured price-level movements broadly parallel to those in partner countries. The particular reasons linked to the launching of monetary union will be kept separate from the more general theoretical considerations that would apply also to countries without any aspiration to join a monetary union.

Since this paper is written for a conference on what monetary policy can learn from monetary theory an aside on that perspective may be appropriate. Axel Leijonhufvud wrote in a thematic note on the conference: "... academic monetary theory and central bank operating doctrines seem further apart than at any time pre-viously". A proposition he goes on to illustrate by means of some references to apparent Federal Reserve policy. It is not obvious that he could easily have found European illustrations of his proposition. The conversion to monetary targeting in the 1970s was initially less complete in a number of European countries (except for Germany) than in the United States, but the orientation of monetary policy towards a purely nominal objective has become more unreserved in the course of the 1980’s, in the Maastricht Treaty on monetary union and the preparations for it in designing the monetary strategy and in giving the likely national central bank participants more independence of political authorities during the transition.

The concepts and main results of the credibility analysis, are thoroughly familiar to European central bankers. The latter have seized upon modern monetary and macroeconomic theory with some enthusiasm, because it provides a refined rationale for a purely nominal orientation of monetary policy, for more inde-pendence for central banks individually and ultimately for the ECB, and for re-solving some of their conflicts with national budgetary authorities in ways that constrain the latter so as to give further assurances to central bankers. There is hardly any great distance between what is said in graduate monetary classes and in the boardrooms of European central banks.² This is broadly how European central banks see themselves. There is, more particularly, little discussion of whether they could do better by putting more weight on output stabilization, as is currently the case in the United States. Orthodoxy has become more pervasive in Europe than in the United States. In that perspective, the discussion in the next three sections on certain residual disagrements on monetary strategy or its best implementation should not overshadow the high degree of consensus which has built up regarding the nominal orientation of monetary policy.
 

3. The relative merits of inflation targets

Three merits of direct inflation targets over other strategies for a monetary policy to influence the economy’s nominal variables are usually mentioned in the literature: An inflation target is less vulnerable than an exchange-rate target to speculative attacks and to supply shocks originating in the area to whose currency the exchange rate is pegged. Finally, the strategy does appear to be easier to communicate to the public. We shall examine these properties in turn, keeping in mind that the focus of the present paper is a comparison of inflation and exchange-rate targets, not least with a view to seing how two could coexist. Much of the academic literature appears to assume that they can not coexist, because obligations to intervene in the foreign-exchange market would undermine the efforts to maintain a stable inflation rate. Yet they do appear to have coexisted in some European countries, e.g. Spain. We leave out of the comparison the merits of inflation targets for the monetary aggregates or for nominal GDP.

Let us look first at the properties of an explicit inflation target as against an exchange-rate target. The main contributions are those of Persson and Tabellini (1996) and de Grauwe (1996). The former looks at the relative stabilization potential of the two procedures and concludes that the expected losses for the outside country are smaller in the inflation-targeting regime in the face of asymmetric supply and speculative shocks. De Grauwe’s paper reaches a similar conclusion by a different route. He argues that in terms of the traditional credibility model, inflation targets come out best because they avoid the cumulative build-up of a misalignment arising from the combination of a fixed nominal exchange rate with less than full convergence of inflation rates. As the degree of overvaluation rises it will at some point trigger a speculative crisis and a devaluation. De Grauwe also presents some empirical estimates of credibility which suggest that the advantage of inflation targets is not due to better credibility than is obtainable with exchange-rate targets, but to the less dramatic consequences of the former. But all of the empirical observations are from periods that are hardly representative of to-day’s world of high convergence of inflation and long-term interest rates. The problem is no longer whether a country can be more successful in converging its inflation rate to a hard core of low-inflation European countries by means of one or the other monetary strategy. Convergence of inflation rates is by now far advanced, making also exchange-rate targets much more credible than in even the recent past. Hence the cost referred to by de Grauwe are no longer unavoidable.

Canzoneri, Nolan and Yates (1996) in a related paper reach a conclusion more favourable to exchange-rate targets. These authors introduce the notion of political pressure on the central bank in the form of pressure for lower interest rates which tends to produce an inflation bias. The main factor in this model for choosing between inflation- and exchange-rate targets hinges on the flexibility of implementation, where an exchange-rate target has clear advantages.

In practice, the choice between inflation and exchange-rate targets is hardly ever of a pure either-or type. A well-designed monetary policy should find it possible to reconcile a high degree of exchange-rate stability in the short term with low and stable inflation in the longer term. The time horizon for implementing the two strategies is simply so different that they can usefully be regarded as complements and as mutually reinforcing rather than as necessarily conflicting. An inflation target can enhance the efficiency of interventions or of other short-term measures designed to stabilize the exchange rate. This interaction is neglected in models which analyse the policy-maker’s loss functions in an atemporal context.

Simply declaring an inflation target does not offer sufficient operational guidance for monetary management under flexible exchange rates over short time horizons. This is in part due to the difficulties of getting a precise understanding of how domestic policy changes impinge on the exchange rate. The power of monetary policy in influencing domestic demand may well increase with the degree of flexibility of the exchange-rate regime. But the transmission mechanism becomes more uncertain, because the strength of the linkages through the exchange rate depends strongly on how monetary actions are perceived by the markets.

Uncertainty about the transmission mechanism relates not only to the impact of monetary policy actions on the exchange rate, but also to the subsequent combined impact of interest rates and the exchange rate on nominal domestic demand and output and - ultimately - on the break-down of the latter between its real and monetary components. One of the first central banks to introduce inflation targets as a long-run framework for monetary policy, the Bank of Canada, also pioneered in the analysis of the combined impact of interest- and exchange-rate changes by means of a so-called Monetary Conditions Indicator (MCI). Using as weights the impact of each of the two transmissions channels on domestic activity, and hence indirectly on inflation, it becomes possible, in principle, to construct an interest-rate reaction function for the central bank showing how it could optimally respond to tensions between desired and forecast levels of the target variable - inflation over a control horizon of up to two years - while taking into account the change in the ex-change rate induced by the monetary actions. This exercise is, however, extremely difficult and involves in practice a heavy dose of judgement, see e.g. White (1996), even though the longer-run inflation objective facilitates the reduction of uncertainty.

European countries who have the prospect, if politically willing, to move to fixed exchange rates with most of their trading partners, will experience a con-siderable simplification of an admittedly weaker transmission mechanism for a joint monetary policy on which they can exercise some influence without diminishing the commitment to a medium-term inflation objective bound to be close to what they would have set individually. The smaller and the more open the domestic economy, the larger the weight of the exchange-rate transmission channel will be relative to that of the interest-rate channel. And the firmer the exchange-rate links among the country’s partners, the more important the external transmission channel will tend to become for a country remaining outside.

The discussion of optimal contracts for central bankers is also relevant to the decision whether it is desirable to encourage countries with an inflation target to also implement an exchange-rate target. Panizza (1997) has shown that if (1) the shocks hitting an economy - the supply and speculative shocks referred to above - are large, and (2) if the individual central bank studied and the larger area to which it may peg have the same type of contract, say, to aim for a similar inflation rate over the control horizon, and (3) if a weak form of rnd who both see the movements in the exchange rate vis-á-vis their European partners as providing important information on future inflation. In that case a central bank performance contract written in terms of the exchange rate could bring some of the benefits of the linear inflation contract, because the contract would exploit any linear relationship between current movements in the exchange rate and in future inflation. Such a contract would be easy to monitor on a current basis because the exchange rate is available continuously and without the ambiguity that necessarily surrounds the inflation target. Hence it is surely more readily enforceable than the inflation contract proposed by Walsh (1995). The external target is also readily understandable to economic agents.

To this one can, in the European context, add a more strategic consideration. If the country wishes to be sure to maintain the option to join a monetary union with its partners, and there is a risk that participation in an exchange-rate system for a transitional period before being admitted is going to be a formal requirement, longer-run considerations may swing the balance of the argument in favour of taking the minimal step of declaring a central rate and undertaking to keep the fluctuations in the market exchange rate within the wide margins under which the present EMS - and its successor arrangement recently outlined for those who do not join the first group in monetary union - functions.

Arguably, the most important part of joining such a transitional system has now become the declaration of a central rate, since the fluctuation margins are so wide that they appear unlikely to be tested. Some have concluded that with margins as wide as 15% between the strongest and the weakest currency, the EMS has come close to a system of flexible rates, but this is in my view a mistake. The central rate is the crucial element, since the obligation to keep movements around it manageable brings external considerations explicitly into the domestic policy debate; an inflation target alone can do that only imperfectly and indirectly. The former target is therefore a particularly useful precursor to participation in a monetary union where the ECB will manage a unified monetary policy targeted at nominal developments in the area as a whole, see section 4 below.

There is a further practical consideration in the European context.³ The clustering over the past year of market exchange rates in the EMS around the central rates - the spread between market rates has recently narrowed to less than 2% between the strongest and the weakest currency, except for the Irish punt which has fluctuated in a range 10-12% stronger than the weakest currency, the French franc - suggests that the rates at which currencies joining monetary union will be converted into the future single currency, the euro, on 1 January 1999 will be the central rates or rates close thereto. These rates have been the centre of gravity of the system for more than a decade for a large majority of the prospective participants; and any currency which has joined more recently or been devalued over this period have been admitted on the basis of careful negotiations on the appropriateness of the central rates currently at entry. It is valuable to both central banks and to market participants to have had extensive experience with central rates and their acceptability in the markets before the final locking of intra-European exchange rates takes place.

The large majority of present Member States in the European Union which insists on participation by all potential candidates for monetary union in an exchange-rate arrangement during a transitional period prior to full admission therefore have a point. Such participation is seen as a useful apprenticeship period for which even a highly creditable pursuit of an inflation target is not a fully satis-factory substitute.

Anyway the Maastricht Treaty already contains an inflation objective, formulated in relative terms. Pursuit of a tighter and absolute target as in Sweden and the United Kingdom is fine, but it has not protected the countries themselves and their European partners from substantial erratic movements in the exchange rates between the outsiders and the EMS participants, which can not fail to have an impact on the degree of goods market integration to which all Member States in the European Union in principle aspire. The most recent example of such fluctuations is the rise of the pound sterling in 1996-97 to a level which the UK authorities themselves consider clearly excessive and erratic, see e.g. Bank of England (1997). This rise may be reversed in the nearer-term future, as was the overcorrection in the opposite direction of the Italian lira and the Swedish krona in 1995. Fortunately the earlier slumps in these currencies proved so temporary that they had only a small influence on the macroeconomic performance of the two countries in general and on inflation in particular. Having now lasted longer, one can not be confident that the present overvaluation of sterling will be absorbed at similarly low costs. The anti-inflationary gains which appreciation has brought are reversible.

Most of the above discussion of the usefulness of inflation targets accepts that such targets are useful and mark a significant advance in policy formulation in the countries that have adopted them. The thrust of the argument is rather that, in individual smaller or medium - sized economies, they should be supplemented by exchange-rate targets, because the latter facilitate monitoring and save the initiating country the cost of residual exchange-rate variability which can remain substantial even in cases of convergent inflation behaviour. At least if the margins of fluctuations are wide, as is the case in the present and future EMS, having an exchange-rate objective in the form of a central rate can be helpful and should not unduly limit the element of discretion in monetary policy.

Finally, there is the issue of which targets best facilitate communication with financial markets and with the general public. Although inflation targeting can not be demonstrated to have significantly affected the way in which monetary policy is conducted, see i.e. Almeida and Goodhart (1996), Ranki (1997) and White (1996), the central banks that have adopted it see a major advantage in the opportunity it has provided for better communication and transparency. Some central banks, notably the Bank of England, Sveriges Riksbank and the Banco de España, have introduced new regular publications on inflation in which they present their inflation forecast along with an analysis of how it deviates from the target (or target range) see e.g. Bank of England (1997). These publications contain a useful clarification of the practical implementation issues, e.g. about the precise price index which is targeted. There is no doubt that these efforts have been helpful in improving public understanding and acceptance of the new framework for policy in a way that would hardly have been possible solely by means of an exchange-rate target, particularly in the aftermath of the traumatic experience of the early 1990s. In these circumstances, relying solely on reset exchange-rate targets by countries which have re-cently been obliged to withdraw from a futile defence of their earlier central rates, would have appeared arbitrary and hence hard to justify.

Some of the theoretical contributions, see Svensson (1995) and Green (1996), have seen a problem in the need for the central bank to announce a lower target for inflation - possibly to have such a target assigned to it by the government - than the underlying social preference for inflation in order to hit the latter. Since an inflation bias persists even in a system with explicit targeting, the latter has to allow for that and be more conservative in setting its objective than what it can truly expect to achieve. On average, the announced target can therefore be expected to be overshot, hence eroding future credibility to the extent that the inconsistency is reflected in private expectations.

The problem of setting a benchmark which is never quite met could be papered over by setting a target range for inflation and de facto aiming for the lower half of the range rather than announcing a single-valued target. Another method would be to define an enforceable target by means of a Walsh-contract, possibly indirectly through an exchange-rate target as mentioned above. Svensson (1995) also proposes a third method, viz. to announce both an inflation and an output target, the latter set equal to some natural, non-inflationary rate and hence consistent with the former, but that would meet with major difficulties in Europe where the 1990s have made estimates of capacity output even more hazardous than elsewhere.

In practice this problem of the self-destructive properties of a procedure which repeatedly fails to meet the announced inflation target has not had to be faced yet by European policy makers. If anything, inflation targets have tended to be undershot in the European countries that have used them, notably in Italy, Spain Sweden and Finland. This tendency for inflation to be reduced faster than announced has undoubtedly been due to the considerable slack in the European economies in the long drawn-out adjustment following the deep recession of the early 1990s. But the problem will arise in the future and may already be visible in the United Kingdom where the upswing in the economy has been stronger than on the Continent. It would be better to address it before it risks eroding the public acceptance of inflation targets.
 

4. An Inflation target for the European Central Bank?

The European Monetary Institute (EMI) which is preparing the strategy and the implementation of monetary policy after 1 January 1999 has devoted some attention to the proper choice of objective(s). In its main statement on the issue, EMI (1997), the central bankers mention five possible objectives for the joint monetary policy:

(1) the exchange rate, (2) interest rates, (3) nominal income, (4) monetary aggregate(s) and (5) inflation. But the list of candidates is immediately narrowed to the latter two and possible combinations of them. The EMI did not consider it posible, or necessary, to determine more specifically which strategy should be preferred by the ECB. The EMI Council did, however, identify a number of key elements as an indispensable part of any strategy, EMI (1997): p.III:
 

The main argument against reliance on an EU wide monetary aggregate as the main target is the obvious one that the relationship between any chosen target and nominal incomes - after a lapse prices - will be subject to particular uncertainty in the early stage after the major reform of launching monetary union. But it is also pointed out that inflation targeting too is complicated by this shift which may not leave the relationship between various leading economic and financial indicators and future inflation unchanged. In short, the prospective candidates for monetary union appear to have agreed that a mixed strategy with both inflation and monetary targets is preferable to emphasis on a pure strategy involving only one or the other. The non-committal language may, of course, also have been influenced by the well-known differences of opinion between the two schools of inflation-target advocates led by the United Kingdom and monetary-target advocates led by Germany. These two groups have not found it productive to push their discussion to a more definite conclusion at the present time when participation in monetary union is still surrounded by some uncertainty.

What is clear and well-justified is that exchange-rate targeting will play a very minor role in the future strategy, at least initially. Particularly if the monetary union is large, comprising, say, 10-11 Member States, the union will resemble the United States and Japan in its degree of openness to the outside world, i.e. with less than 10% of its GDP being traded with non-participants. In these circumstances, the exchange rate will not even provide much information as an indicator of future inflation and the ECB will be tempted, as have the monetary authorities in the United States and Japan, to treat it with a degree of "benign neglect", at least until serious policy conflicts with the other main economies in the world encourage a reevaluation.

The ECB will be well-placed to pursue with a relatively low degree of distraction whatever monetary strategy it ultimately settles for. The two factors which could upset it, large-scale exchange-market interventions and major imbalances in the sum of national public sector finances, have both been brought under tighter constraints than is presently the case in any of the world’s largest economies.

The Maastricht Treaty leaves exchange-rate management largely in the hands of the ECB, except in circumstances where the governments in the euro area unanimously agree to seek a formal agreement with other major countries on a new global exchange-rate system - a very unlikely event given attitudes in both Europe, the United States and Japan. In other circumstances a qualified majority of Finance Ministers can issue "general orientations" for exchange-rate policy (Art 109, 2 of the Treaty), but the ECB is not obliged to observe them if it deems them to be in conflict with its primary objective of price stability. EU officials are currently in the process of studying what forms such orientations might take, but nothing to substantially encroach upon of the ECB’s dominance is likely to result. With a large euro area, some countries will still remain outside monetary union, but here too, outright obligations to intervene in the remaining EU currencies are bound to be infrequent and subjected to the primary and internal objective.

As regards disturbances to any ECB stragegy from imbalances in public finance, considerable efforts have been made to assure through the entry requirements that participants will start out in monetary union with historically modest deficits and stable or falling debt ratios and through constraints on their subsequent behaviour in the form of a "Pact for Stability and Growth", elaborated at recent meetings of the European Council. If participants live up to the two main principles of these rules, viz. to aim for budget balance or a small surplus in normal years and to a precommitment to vote in favour of applying sanctions to countries which allow their budget deficits to rise above 3% of GDP in any year which cannot be seen as truly exceptional because of a severe recession, a major step towards achieving a better policy mix than in the past and hence protecting the independence of the ECB to pursue its chosen stategy will have been taken. There are obvious question marks concerning compliance with both of these principles, but at least the qustions have been squarely faced.
 

5. Conclusion

The present paper has focussed - without any pretence of generality - on some issues facing a subgroup of European countries in their choice of monetary policy strategy. These countries have had to define a proper role for inflation targets in a situation where an unusual degree of consensus has emerged in Europe that central banks should confine their vision to the achievement of a stable nominal framework for the economy, i.e. low inflation. Is an inflation target in itself the best way of achieving this objective, or should it be supplemented by an exchange-rate target for individual countries in the transition to participation in a monetary union - and thereafter, once the union has started? Could the European Central Bank in its efforts to sustain the low inflation presently reached throughout Western Europe usefully supplement an explicit target by other objectives?

Any answers to these questions must be tentative, but they point in the affirmative direction for both. Sweden and the United Kingdom have both relied on an inflation target as the dominant monetary policy objective for the past five yars. They have obtained unexpectedly good results (as have Canada and New Zealand outside Europe). It would be regarded as hazardous to recommend to them on the grounds of monetary theory to switch to a strategy of pegging once more their exchange rates. There are arguments in favour of sticking with the inflation target, but also for supplementing it with an exchange-rage target in the present, rather loose, EMS and later in its successor arrangement, in order to reduce erratic exchange-rate fluctuations and to retain the option to enter monetary union when that is thought to be politically feasible.

As far as the future joint monetary policy to be conducted by the European Central Bank is concerned, inflation targeting also has some appealing properties. The whole construction of monetary union should make it easier for the ECB than for any existing central bank to pursue such a strategy well protected from inteference from pressures arising either in the foreign exchange market or in unbalanced budgetary policies in the participating countries. But the ECB as well as those who have given it the mandate to pursue price stability have an interest in also using monetary targets as an intermediate objective in order to improve monitoring, communication to the public and hence accountability. Exchange-rate targeting would be inappropriate in this case, except in currently unforeseeable circumstances.
 
 

References