Economic indicators for monetary policy in a globalised economy

7The globalisation of financial markets does indeed complicate the analysis and interpretation of the economic indicators relevant for the conduct of monetary policy.

8The behaviour of monetary aggregates, for instance, may not be fully consistent with domestic developments as evidenced by conditions in the domestic money and capital markets. In particular, the sensitivity of domestic monetary aggregates to interest rate changes may vary according to the external conditions facing the domestic economy. Furthermore, the increased responsiveness of capital flows to economic conditions in different economic areas makes money demand less stable. It is therefore more difficult to disentangle short-term movements of capital across the borders from more structural and longer term flows. Credit developments may also be affected, as they may reflect the need to finance investments abroad, rather than domestic demand, and their leading indicator properties may therefore vary over time.

9Moreover, in an environment in which firms can easily allocate production to different countries in response to increasing demand in one country, the national or domestic output gap becomes more difficult to estimate.

10In the same way, exchange rate developments may increasingly reflect the prospective developments in the real economy in the different economic areas of the world, rather than simple nominal interest rate differentials. While nominal interest rates can explain exchange rate developments in a context of stable real economic growth across different geographical areas, this is no longer true when it is precisely the difference in the expected economic growth potential that is affecting the expected real exchange rate. Assessing the nature of an exchange rate depreciation is a fundamental part of any monetary policy strategy aimed at price stability, as different interpretations of the signals that a depreciation sends about domestic inflationary pressures can have quite different policy implications. If the exchange rate depreciates as markets discount excessive future domestic absorption, the optimal policy response would be to tighten domestic liquidity conditions. However, if the exchange rate depreciates because capital flows leave the country, having been attracted by better expected real returns abroad, the depreciation might not signal greater pressure on domestic inflation and may warrant a different policy response.

11The same sort of considerations apply to the interpretation of the yield curve and its information content concerning future price developments. Assets denominated in different currencies are held by international investors who are subject to different constraints and have different objectives. When deciding on their portfolio allocation, these investors must necessarily take into account the developments in the different regions of the world and in monetary policy in each of these regions. In this context, the long-term interest rate in a given economic area may not primarily reflect conditions which are specific to that area, but rather international capital market conditions. In this respect, the monetary policy strategy must be capable of correctly interpreting movements in long-term interest rates, as they may convey information concerning domestic inflationary developments as well as long-term real interest rate differentials and thus real economic growth differentials.

12More generally, the importance of international capital flows only complicates the existing difficulties inherent in the interpretation of asset prices. For instance, asset price developments can have different effects on future price developments depending on the nature and extent of the wealth effects linking asset prices to goods prices. In this respect, market capitalisation may be a misleading indicator of possible future price pressures if considered in isolation. Market-specific institutional features, such as the structure of asset holdings, should also be taken into account. Positive wealth effects are more likely to be significant when holdings of assets are concentrated in the household and corporate sectors, as any improvement in their balance sheets is reflected in higher expenditure (on consumption and investment, respectively). By contrast, when assets are concentrated in the balance sheets of a small number of institutional or foreign investors, such wealth effects might be far less relevant. Individuals holding pension fund accounts with such institutions might adjust their “permanent” income expectations according to asset price developments, but the aggregate effect on total private demand would be influenced by the number of people holding such pension fund accounts, their ages and the international distribution of their portfolios. Finally, wealth effects can, in certain circumstances, affect not only domestic demand, but also domestic supply for protracted periods. This may occur, for instance, when over-investment in working capital or in real estate leads to firms’, banks’ or households’ balance sheets being unsound for lengthy periods or when by contrast, wealth effects only reflect strong productivity increases in firms after some delay.

Monetary policy strategies in a globalised economy

13As a result, globalisation has two kinds of effects on the information available to central bankers around the world for the conduct of monetary policies. On the one hand, the continuing integration of the world economy implies that the information conveyed by many indicators no longer exhibits a stable relationship with domestic developments. On the other hand, because of the globalisation of the world economy, individual economies are behaving differently.

14For example, it is likely that the transmission mechanism of monetary policy in any country or region of the world is affected by the globalisation of the goods and the capital markets. This is why the structural reforms that were discussed mainly in view of their optimality in a closed context in the 1990s are today considered necessary in order to face the challenges of globalisation and to play a direct role in short-term market expectations. In addition, monetary developments abroad have an effect on the structure of the domestic economy and on how it adjusts because they may make the financing of domestic investment easier. Given the increasing importance of international capital flows in determining the economic environment in which monetary policy has to be conducted, issues such as transparency and the communication of monetary policy decisions have also come to be crucial. Monetary authorities must increasingly be able to communicate their assessment of the economic situation and the corresponding monetary policy decisions effectively and in a transparent manner. When uncertainty is reduced to the minimum and the credibility of monetary policy is maintained, private agents take decisions consistent with the policy decisions and contribute directly to achieving the objective of price stability in the medium term.

15In this context, monetary policy strategies must retain a high degree of flexibility. Neither strict monetary rules nor simple forecasting techniques are sufficient. Monetary indicators and forecasting exercises are important components of such strategies, but they cannot be considered a complete strategy in themselves. Monetary policy-makers simply cannot rely on just one or a few indicators in order to gauge the inflationary pressures in the economy. Nor can monetary authorities rely only on a single synthetic and more or less structural model when forecasting inflation or react mechanistically to any deviation of this forecast path from the inflation objective. Whatever the approach used, some discretion in the analysis is necessary to cope with the complications listed above. At the same time, and in line with the arguments set out above, the transparency of monetary policy and the ability to communicate policy decisions to the markets effectively are essential. The challenge for central bankers is therefore to design their monetary policy strategy in a way that combines rules and discretion and that reflects both the globalised context and the need for good communication.

16This need to combine rules and discretion is increasingly understood by central bankers around the world. Most of them now recognise the need to use a large set of indicators and to exercise a certain degree of discretion both in interpreting monetary aggregates and in producing forecasts.

17As far as the Eurosystem is concerned, its monetary policy strategy, based on a definition of price stability and on “two pillars”, is entirely consistent with the global financial environment described above. The definition of price stability, which is the primary objective of the Eurosystem, is of course consistent with the view that price stability is the best contribution that central bankers can make to long-lasting and sustainable growth. On the one hand, the Eurosystem has made it clear that price stability is to be maintained over the medium term, thereby reducing the emphasis on short-term fluctuations in price indices, in particular those due to terms-of-trade effects, which do not have an impact on medium-term developments in inflation. On the other hand, the ECB definition – an increase in the HICP for the euro area of below 2 % year on year – makes it obvious that deflation, as well as inflation, is considered inconsistent with price stability.

18By virtue of its two-pillars, the Eurosystem’s monetary policy strategy is able to address the type of uncertainties which characterise the economy at the present juncture. Although the analysis under each of these pillars can be affected by internationalisation and globalisation, the combination of the two makes the strategy robust. The first pillar is the analysis of the information content of developments in a broad monetary aggregate, such as M3, and in its counterparts with regard to future price developments. Although there is no model of the economy available which integrates money in a satisfactory manner, monetary developments are considered, at least in the euro area, relevant to price developments and have to be taken into account in their own right. Indeed, money demand appears at present to be stable. Money has shown good indicator properties for price developments in the past, in particular concerning robustness to asset price shocks. The second pillar is composed of a broad set of other indicators, ranging from the exchange rate to the yield curve, asset prices, confidence surveys, wage developments, various types of forecasts and analyses of fiscal policy. Overall, this choice of strategy reflects the predominant role attributed to monetary aggregates in explaining medium-term developments in inflation, while at the same time recognising that it is impossible to use a single model or statistic to summarise economic developments in the current context.

International monetary policy co-ordination in general

21In theory, international monetary policy co-ordination in the sense of common decision-making can improve welfare in the case of symmetric shocks hitting different economies in the same way (e.g. a supply shock or an oil price shock). In such circumstances, domestic monetary authorities are better off internalising the effects of domestic policies on foreign countries. When considering this general theoretical conclusion, however, it should not be forgotten that common decision-making in the field of monetary policy is very difficult to achieve. In practice, co-ordination is considerably more complex than standard theory would suggest.

22First, as evidenced in Section 2, the models and the economic analysis used on a continuous basis by central bankers are both more complex and less precisely defined than the models used in this literature. In particular, it remains difficult to develop a widely accepted model of the world economy and to define precisely the data-generating process for particular variables and ultimately to measure the size both of spillovers and of policy multipliers. The problem of “model uncertainty” does not necessarily lead to the sub-optimality of the co-ordination exercise. Nonetheless it makes its implementation very difficult. Second, the preferences of those engaging in the co-ordination are rarely defined ex ante and may vary over time. This complicates the co-ordination exercise, as each individual actor is not able to quantify the costs and benefits of co-ordination. Furthermore, changing objectives and preferences weaken the credibility of the monetary authorities’ commitment to specific policy rules. Therefore common decision-making on the basis of pre-agreed procedures is often difficult.

23Third, the optimality result depends on the participation of all relevant parties in the co-ordination exercises. [1] This requires the ability of policy-makers, among other things, not only to stick to their co-ordinated commitment, but also to enforce the policy decision even when domestic constituencies have conflicting interests. The credibility and realism of co-ordination procedures is critical to their success. Hence, a realistic approach to international co-ordination would require that the procedures be tailored in such a way as to enhance the commitment of, and the information available to, all interested parties.

24Fourth, the benefits of co-ordination may often derive more directly from the common analysis of international developments which central bankers can achieve together than from any commitment to act in a predetermined way. For instance, interpreting the nature of capital flows (and related asset price movements) is often better done by discussing both their sources and their destination. To the extent that such flows may, in the extreme, result from either underlying factors or self-fulfilling prophecies, sharing analysis on their interpretation is particularly useful and reduces the monetary policy dilemma that may arise from divergent interpretations.

25All of this suggests that ad hoc co-operation is more suitable for international co-ordination efforts between major economies than is strict co-ordination. First, ad hoc co-ordination intended as a continuous exchange of information among the participants is one practical way of ensuring that every player “learns” what the models of the other players are. Ghosh and Masson [1991] showed that, whenever Bayesian learning by policy-makers about each others’ models is allowed, co-ordination dominates autarchic regimes. Hence, ad hoc co-operation can help to resolve the model uncertainty issue, i.e. the fact that co-ordination is difficult because different governments use different models to calculate the pay-offs from co-ordination or autarchic behaviour. [2] This also implies that regimes other than such ad hoc co-operation have to pay a great deal of attention to the credibility of the commitments they make.

26Second, the sustainability of co-operative equilibria – in the game theoretic sense – depends not only on the independence and degree of “conservatism” of the policy-makers, but also on the credibility of the commitment of the policymakers to their own private sector. [1] As every player involved in the co-ordination exercise has imperfect information about the degree of commitment of the other players, the co-ordination procedure has to be structured in such a way as to facilitate the transmission of information among all players. In this sense, monetary policy-makers must be able simultaneously to achieve two things : first, to always put in place policy decisions that are consistent with their objectives and, second, to be able to communicate clearly to the other players the reasons for and scope of such decisions. As such credibility can only be established gradually, this often makes ad hoc co-operation preferable. At the same time, because ad hoc co-operation consists in the exchange of information, it limits the risk of losing credibility for any policy-maker involved in the co-operation exercise who cannot follow automatic policy rules owing to new domestic economic or political developments modifying the constraints within which that policy-maker has to act.

27Finally, as co-ordination takes place over time and actions are repeated, the theoretical advantages of pure co-ordination compared with repeated ad hoc co-operation are likely to be limited.

Exchange rate pegs and international co-ordination

28Owing to the practical difficulties involved in implementing “pure” monetary policy co-ordination exercises, it has been suggested that international policy co-ordination might focus on just one variable – the exchange rate. This approach would avoid the complexities linked to co-ordination based on monetary aggregates or interest rates, would leave the partners free to set other policies independently and would be more transparent, because the exchange rate is easier than other variables for private agents to monitor.

29However, this apparent simplification glosses over significant problems. Fixed exchange rate regimes may increase the credibility of domestic policy decisions when real divergence is limited or when discretionary decisions on the part of the participants in the exchange rate agreement are impossible. [2] However, fixed exchange rate regimes limit the ability of the country to adjust to real shocks when domestic prices are sluggish. Furthermore, if, in an exchange rate agreement, one country bears the entire burden of adjustment to real shocks, this might create incentives for the other country to abandon the agreement. In this respect, barring the case of monetary union, fixed exchange rate arrangements including currency boards are more likely to succeed in the long term when they are accompanied by real convergence and strong political commitments.

30These considerations are underlined by many episodes of economic history. Adjustable pegs such as the gold standard or the dollar standard proved to be viable as long as the other participants accepted the dominant position of a single economy in the determination of monetary policy. In the same way, the Exchange Rate Mechanism (ERM) crisis in 1992-93 and the more recent Asian crisis showed that a fixed exchange rate arrangement is only viable as long as the costs implied by adhering to such an arrangement are sustainable. Moreover, it demonstrated how overly rigid exchange rate targets may, by providing excessively cheap insurance to market operators, lead to major financial crises in some cases. The extent to which a given exchange rate arrangement is appropriate thus depends upon various institutional factors, including the degree of real convergence among the participating economies, the nature of their commitment to the arrangement and the degree of political support which this arrangement enjoys.

31It follows from the above that monetary policy co-ordination, even in the “weaker” form of exchange rate targeting, is not viable for large countries or economic areas. In such cases, political commitments to exchange rate stability are unlikely to supersede and determine other policies. Moreover, real divergence may be significant owing to the divergence of economic and structural policies. At the same time, the other extreme – a decision to allow the exchange rate to float freely – leaves the country concerned exposed to the variability of the financial markets, which is not always driven by fundamental factors and which may force the exchange rate out of an equilibrium path. As this may lead to the misallocation of resources, some form of “dirty” floating is usually preferred, in connection with ad hoc co-operation in the context of groups such as the G7.

32This may not apply to smaller countries. International monetary policy co-ordination may, in such cases, take the form of a “Stackelberg equilibrium” (i.e. a situation in which monetary policy is wholly delegated to a foreign country, which acts as the leader when setting the common monetary policy). This may have beneficial effects for the country in question, provided that the delegation is credible. Domestic monetary conditions align themselves with those in the foreign country, thus eventually helping to deflate the economy at a lower cost in terms of output loss. [1] In addition, the country concerned could have better access to global financial markets, as long as the pegging country is perceived to be committed to the exchange rate peg. It would then achieve economic integration with the country to the currency of which its exchange rate is pegged. However, as the issue here is credibility, such forms of monetary policy delegation are still subject to “institutional” risks. Even in the presence of a convergence of real business cycles, asymmetric shocks may happen. The commitment of the domestic monetary authority may then be challenged and the costs of sticking to the exchange rate arrangement might become too high. In this case, either credible adjustment mechanisms are at hand, or it might be envisaged, under appropriate circumstances, to renege completely on domestic monetary sovereignty by adopting a currency board or by immediately introducing the foreign currency as legal tender (“dollarisation” or “euroisation”). Devising such adjustment mechanisms is not easy. They need to ensure a sufficient degree of exchange rate stability in order to be credible. At the same time, they need to give timely responses to shocks. Finally, the way they function should not provide too much insurance for speculators, but should provide sufficient support to enable exchange rate pags to weather periods of excessive volatility. The ERM, despite some degree of success, experienced major crises up to August 1993, after which it became more stable. ERM II, which started in January 1999, probably provides a better example of such a mechanism. By giving a central role to the euro, it makes clear which currency is the Stackelberg leader. By ensuring that participant central banks, including the ECB, can call for a realignment, it tries to ensure that adjustment decisions are taken in a timely manner. By using a fluctuation band of ± 15 %, it limits the “one way” risk and makes it easier to agree on realignment ahead of an exchange rate crisis. In this way, adequate communication may be ensured between market participants, the authorities of the country which has an exchange rate target and the anchor country or currency area.

International policy co-ordination in a globalised world

33In view of the above, international co-ordination in a globalised world is likely to be achieved through a combination of ad hoc co-operation between large countries and a certain amount of exchange rate co-ordination at the regional level, in the form of exchange rate targets complemented by ad hoc co-operation.

34Hence, several issues which are essential to the smooth functioning of international co-operation may be raised in relation to the efficiency of such co-operation. Central bankers and other authorities already invest large amounts of resources in such co-operation. Moreover, such co-operation covers not only macroeconomic issues, but also market developments, as well as statistical and supervision issues. The number of available fora has increased rapidly. As a result, the discussion on the international financial architecture, which has tended to focus on how best to organise crisis management and the functioning during such crises of institutions such as the IMF or the World Bank, might usefully consider the following questions also :

• Are fora for co-operation, including, inter alia, the G7, G10, G20, the OECD, the IMF and the BIS, as well designed as possible with regard to the efficient exchange of information ? In particular, do they bring together the right participants ? For instance, the way in which the European Union and the European Central Bank participate in the various international fora is an important issue.
• Moreover, are all fields in which co-operation might be necessary adequately covered ? Monetary co-operation is, for instance, unlikely to be very efficient unless sufficient fiscal co-operation is available. In the context of the EU, the Growth and Stability Pact has been drawn up to ensure that a clear framework for fiscal co-operation between EU countries is available. Although it is unlikely that such a constraining framework could be developed at international level, the development of adequate co-operation in this field could also be envisaged. In this respect, the IMF has already put forward a code of good practices.
• Furthermore, are these fora as well equipped as they can be to promote awareness of relevant issues and shared views ? This is linked to their ability to identify new issues in a timely manner and to ensure that the benefits from co-operation are distributed evenly. [1] It also raises the question of whether such fora identify best practices and exert peer pressure efficiently and rapidly. This, in turn, is closely related to the ability of their support staff or secretariats to formulate issues in ways which are relevant from the point of view of both economic analysis and the institutional responsibilities and mandates of participants in ad hoc co-operation.
• Finally, what degree of overlap and of co-ordination between the various fora is optimal ? Are there issues which would best be overseen by a more restricted group of countries, leading to regional co-operation ?

35Such issues are certainly relevant to the effectiveness of the international financial architecture. In view of the large amount of time and resources currently being invested by central bankers in the functioning of such fora and their success in making progress in various fields of central banking, including European Economic and Monetary Union, their study is becoming a key area of the global debate on the international financial architecture.