It is a pleasure for me to address the third conference of the International Research Forum on Monetary Policy, sponsored jointly by the European Central Bank, the Board of Governors of the Federal Reserve System, the Center for German and European Studies at Georgetown University and the Center for Financial Studies at Goethe University.
Let me first welcome all participants and in particular Governor Kohn, an esteemed member of the Board of Governors of the Federal Reserve System. I am pleased to see that this high level conference series has become a major occasion for a transatlantic exchange between central bankers and academics on international monetary policy issues. I would also like to congratulate the organizers for gathering together papers that provide significant contributions to the theory and practice of monetary policy. I was particularly happy to see that the papers address issues that often appear in the agenda of meetings among central bankers, such as the conduct of monetary policy in the presence of uncertainty, the implementation of optimal policy, the communication of operational procedures to the public, the co-ordination of central banks in a global environment, and the interaction between monetary and fiscal policy.
I would like to address the issue of a one-size-fits-all monetary policy for a heterogeneous currency union – and specifically for the euro area. This is by no means a “new” topic. But the choice is not – or at least not entirely – due to lack of originality. There are two reasons why I think it is an important topic to address tonight.
The first reason lies in the renewed interest on this question in the recent public discussion. The desirability of a single monetary policy for the euro area was highly debated at the start of Economic and Monetary Union (EMU). Indeed, the common judgment on the ECB’s policy in the early stages was that it was too tight for low growth countries such as Germany or Italy, and too loose for high-growth countries such as Ireland and Spain. Complaints declined over time, due to the gradual reduction in the diversity of inflation rates and the stable dispersion of GDP growth rates among euro area members. They were renewed recently when media alarmed about increased divergence of growth rates in the euro area in the last quarter of 2004. Concerns were particularly strong for Germany, whose permanently weak growth was taken as an indicator of the costs imposed by EMU and by the single monetary policy.
The second reason for choosing this topic is that it fits the twofold scope of the International Research Forum. The IRF aims at fostering a dialogue on monetary policy issues that are relevant on both sides of the Atlantic, and more generally from a global perspective. It also aims at providing a forum where policy-related questions search for research-based answers and where unanswered questions provide stimulus for further research. The conduct of a single monetary policy in the presence of heterogeneity is relevant for a currency union among independent states such as the euro area, for a large federation of states such as the US, but also for countries where monetary policy operates in the presence of regional or sectoral divergences. The topic is also one where policy-makers are in a quest of research-based advice: What role should disaggregated information have in the monetary policy strategy of a central bank? Should a central bank react differently to sectoral, regional or national information even though its mandate is to achieve price stability for the currency union as a whole?
The euro area is a currency union among 12 countries. Prior to the establishment of EMU, the idea of a single monetary policy for those countries encountered scepticism among the economic profession supported by the arguments of the theory of optimum currency areas, which recommended definitely fixing the exchange rate only among economies with flexible markets, free mobility of factors and limited incidence of asymmetric shocks. Clearly, this was not the case for the euro area, which is characterized by unique institutional and economic features such as significant rigidities in labour and product markets, limited labour mobility, differing national industrial structures and rates of productivity growth, lack of a significant centralised fiscal transfer mechanism, and decentralised responsibility for fiscal and other economic policies.
Under these circumstances, asymmetric and even common macroeconomic shocks are likely to induce divergent price developments. It is often argued that a single monetary policy implying a common interest rate in the currency union, combined with inflation differentials, leads to different real interest rates across countries. This may destabilize the currency union by contributing to strengthen inflation differentials further and by creating divergence in output growth. At first sight, the mechanism is straightforward. For a given official interest rate, member countries with higher than average inflation rate experience lower real interest rates, which in turn fuel domestic demand and national inflation. Conversely, countries with lower than average inflation experience higher real interest rates, which lead to further downward pressure on domestic demand and inflation.
However, such a simple analysis neglects several important factors. First of all, the argument is generally illustrated by using the dispersion of ex post measures of the real interest rate, computed by deflating the nominal interest rate by the observed national inflation rate. On the contrary, to capture the effect of inflation differentials on investment and consumption, one needs to look at ex ante measures, computed by deflating market interest rates by the relevant expectations of inflation over the appropriate horizon. This is justified because the real interest rate affects economic activity by changing the price of consumption and investment today relative to tomorrow. Due to the credible commitment of monetary policy to the achievement of price stability, the dispersion of national inflation expectations in the euro area is much lower than that of realized inflation. As a consequence, the dispersion of the relevant measure of real interest rates is limited.
Second, it is uncertain whether the relevant inflation expectations for constructing real interest rates are country-specific or area wide. The appropriate concept differs across categories of agents (consumers or investors) and activities (investment in housing or in traded durables). It also depends on the degree of integration in goods and capital markets across the euro area. Limited integration may justify the use of expectations of national inflation to construct the relevant real interest rate. However, increasing integration may reduce the dispersion of real interest rates, as nominal rates would eventually have to be deflated by euro area expected inflation.
Finally, it should be observed that in a monetary union there are channels that work in a stabilising manner, thus counteracting the effect of potentially diverse real interest rates. The first is a competitiveness channel. Consider a country with lower than average inflation due to weak demand and with higher than average real interest rates. Lower inflation relative to other trading partners would increase competitiveness in that country and thereby the demand of its goods, hence counteracting the initial effect of higher than average real interest rates. The second is the risk sharing channel described by Robert Mundell in his 1973 work. In a currency area, economic agents can better mitigate country-specific shocks through portfolio diversification. Under flexible exchange rates, a country hit by an adverse shock would experience a devaluation of the currency and the country’s domestically-denominated assets would buy less from the same trading partners. Hence, the higher the degree of asymmetry and the divergence in inflation and real interest rates, the larger the benefits from portfolio diversification and the shock absorber role of the currency union.
Let me now turn to list some facts for the euro area, which might help to quantify the relevance of the arguments I have just described. Over the period 1990-1998, the 12 countries now comprising the euro area experienced a strong downward trend in the degree of inflation dispersion. This latter reached its lowest level in 1999, remaining remarkably stable thereafter. By way of comparison, throughout Stage III of EMU, inflation dispersion within the euro area has been on average close to the level observed in the 14 US metropolitan statistical areas, although it has remained somewhat higher than the one recorded in the four US census regions. At the same time, the persistence of inflation differentials in the euro area has been higher than the one experienced by the US metropolitan statistical areas. Since the introduction of the euro, seven of the twelve member states have systematically maintained either a positive or a negative inflation gap against the euro area average.
Remarkably, the process of nominal convergence in the euro area was not accompanied by greater dispersion of real GDP growth rates, despite the irrevocable fixing of the exchange rates and the adoption of the euro. Since 1999, dispersion in GDP growth rates has remained close to its historical average of around 2 percent. The figures recently referred to by the media, suggesting increasing divergence, are constructed using the volatile quarter-on-quarter real GDP and only relate to the four largest countries in the euro area. On the contrary, in the course of 2004 dispersion in annual real GDP growth rates among the 12 member states declined even further, while dispersion in quarterly rates remained broadly unchanged.
Concerning real interest rates, the dispersion of ex ante measures has decreased to about one third of that prevailing in the period 1990-1998, while the dispersion of ex post measures has remained broadly unchanged. Moreover, in the case of long-term real interest rates – those most relevant for investment decisions – the dispersion of ex ante measures has been approximately half that of ex post measures since 1999. These comparisons refer to ex-post measures computed by deflating market interest rates by the national inflation forecasts, as obtained from Consensus Economics. The dispersion is likely to be even lower for measures based on euro area inflation forecasts.
On the underlying causes of the inflation dispersion experienced since 1999, recent ECB research has shown that in nine of the twelve member countries, internal factors were the most important contributors. External factors (as import costs) played a major role only in a limited number of cases, such as for Belgium, France and Luxembourg. Among internal factors, the main contribution came from the unit labour costs and gross operating surplus components, while net indirect taxes contributed to a lower extent. Concerning unit labour cost developments, the compensation per employee component was generally more important than labour productivity in contributing to differentials.
The diversity of inflation rates among euro area countries also has an important sectoral dimension. A remarkable feature is the higher, although decreasing throughout the 1990s, dispersion in service price inflation relative to the one observed for the HICP index as a whole. By contrast, non-energy industrial goods prices converged significantly throughout the 1990s, stabilizing at a low level of dispersion from 1999 onwards. Finally, the evolution of energy prices varies substantially from country to country, as a consequence of the considerable heterogeneity of the euro area countries’ exposure to external oil shocks. Given the large weight of the service sector in the HICP basket, service price dynamics provides the largest sectoral contributor to overall HICP inflation dispersion.
One factor that contributes to explain the persistence of the inflation differentials observed in the euro area is the presence of rigidities affecting the price and wage formation mechanism, as they delay the necessary adjustment of relative prices to economic shocks. Recent preliminary research carried out in the Eurosystem indicates, on average, greater rigidity in price-setting in the euro area than in the US. As regards sectoral differences, energy and unprocessed food prices seem to change most frequently in the euro area, while service prices appear to be modified less frequently. Given the large weight of the non-tradable sector in the economy, the longer adjustment process in the price of services may partially explain the persistence in the overall inflation divergence. The fact that a large share of output in the service sector is accounted for by employment compensation also suggests that a substantial part of the lasting inflation divergence may stem from differences in wage developments and wage-setting mechanisms across euro area countries.
Overall, several elements can be singled out as being important in accounting for the persistence of euro area inflation differentials, such as sluggish price adjustments, wage dynamics in excess of productivity developments, the role of the service sector, the degree of openness of national economies to international trade and their exposition to oil shocks.
Some limited evidence is available on the relevance of the competitiveness channel and of the risk sharing mechanism as forces that counteract the dispersion in inflation and in real interest rates. National competitiveness indicators show that the competitiveness channel is active. All member countries having experienced lower inflation than the euro area average over the period 1999-2004 (i.e. Germany, Finland, France, Belgium and Austria) have also registered improvements in price competitiveness relative to euro area trading partners. The adjustment in national competitiveness seems to develop slowly but to finally offset the real interest rate channel over time.
Concerning the relevance of risk sharing as a counteracting force, the creation of EMU has induced cross-border diversification for investors in the euro area. For instance, the share of mutual funds’ investment made in other EU countries has increased from 10% to 30% over the period 1997-2002. This process has been accompanied by growing cross-border banking activity, mainly in wholesale banking, which has offered diversification opportunities both to credit institutions and borrowers.
To a large extent, inflation and output differentials in a currency union reflect long-term equilibrium phenomena and the necessary adjustments in relative prices following economic shocks, which policies should promote and facilitate. However, the persistence of these differentials in the euro area is also the product, at least to some extent, of misaligned national policies, wage dynamics not linked to productivity developments, and structural inefficiencies such as nominal and real rigidities in product and factor markets. These features are typically impediments in the adjustment to common and asymmetric economic shocks, and need to be addressed by economic policies. The question is: should national, regional or sectoral diversity appear among the objectives of monetary policy? Should it appear among the objectives of national economic policies?
Monetary policy is conducted by the Governing Council of the ECB with the primary objective of maintaining price stability in the euro area as a whole. By pursuing this objective over the medium-term, the ECB’s monetary policy anchors inflation expectations and increases market transparency, thereby facilitating the necessary adjustment of relative prices across different countries, regions and sectors. A recent branch of the academic literature has suggested a more direct reaction of the monetary policy to disaggregated information, particularly when economic divergence results from nominal rigidities. In such circumstances, monetary policy should assign larger weights to units where price developments are more sluggish. However, there are substantial problems related to any implementation of this suggestion. First, there are problems in the measurement of the degree of nominal rigidity and it is unclear at which level such rigidity should be measured (national, regional or sectoral). Second, by assigning higher importance to more rigid units, monetary policy would accommodate structural inefficiencies, ultimately creating perverse incentives. These difficulties would introduce some arbitrariness in the conduct of monetary policy and negatively affect the transparency and accountability of the central bank.
The focus of the ECB on the achievement of a goal defined for the euro area as a whole, however, should not be interpreted as neglect of disaggregated information. While inflation differentials cannot be an objective for its monetary policy, the ECB carefully monitors disaggregated information to assess the underlying causes of such differentials and to formulate the most appropriate monetary policy response. This also facilitates the identification of structural barriers in the euro area that delay macroeconomic adjustment and thus helps to identify areas in which structural reforms are particularly necessary.
National economic policies are better instruments to enhance the ability of individual countries to respond to economic shocks and to national, regional or sectoral divergences. Structural policies conducted at the national level can contribute to ensure a smooth adjustment to shocks or changing economic conditions. Although further integration in financial markets is certainly needed, the creation of EMU has fostered the mobility of capital within the euro area by reducing market segmentation and by increasing cross-border flows. On the contrary, labour mobility remains low between countries and regions, as well as between sectors and professions. Labour reacts too slowly to developments in wages and demand conditions. It is important for national labour market policies to enhance flexibility at the national and regional level, particularly given the existence of differences in languages and cultures, which inhibit mobility across countries. Similarly, structural policies should aim at improving the efficiency of the price setting mechanism in the goods and service markets to reduce the persistence of inflation divergence.
Decentralized fiscal policies also provide important instruments. They can be set in reaction to shocks in such a way as to counteract the presence of structural rigidities and the consequent emergence of differentials. However, sound government finances are necessary if automatic stabilisers are to work fully without the risk of excessively high deficits. In this respect, the Stability and Growth Pact provides governments with sufficient leeway over the long term, provided that they strive to achieve surpluses or balanced budgets in periods of favourable economic activity.
Let me take the occasion to briefly comment on an idea that seems to gain more and more support. It has been recently argued that the ECB should use its collateral policy as a sanction to exert fiscal discipline on those euro area member states that breach the 3 % limit. One possibility would be for the ECB to impose haircuts on the bonds issued by those governments that fail to comply with the Pact, thereby making those bonds less attractive for counterparties to hold and use as collateral in the ECB’s regular operations.
Although superficially appealing, this suggestion would be misguided. First of all, such a measure would exceed the mandate of ECB’s collateral policy, which is to manage risk in monetary policy operations. Assigning additional roles to collateral policy would deflect it from its primary and crucial purpose.
Second, such a proposal ignores the differentiation already applied by the ECB in valuing collateral. All financial assets offered as collateral, including government bonds, are valued daily at market prices. In its collateral policy, the ECB therefore relies on the judgement of the market to distinguish among government bonds and, implicitly, the fiscal behaviour of member states. Moreover, the ECB sets credit standards for the eligibility of assets as collateral and is bound by the Treaty not to distinguish between government and private issuers in the implementation of these standards.
Third, and most importantly, it is clear that the design of the Stability and Growth Pact and its implementation are governmental responsibilities, to be controlled by parliaments. Of course, in giving its assessment of macroeconomic developments, the ECB must always assess all elements that influence the outlook for price stability in the medium term, including changes to the fiscal framework and fiscal measures adopted by national governments. However, it is not and cannot be the ECB’s role to enforce fiscal discipline and to correct shortcomings in the implementation of the Stability and Growth Pact. Attempting to do so would politicise the ECB’s operations and ultimately threaten its independence, on which the credibility and effectiveness of monetary policy crucially rely. The ECB therefore focuses on its own mandate, with the primary objective to maintain price stability, leaving others to meet their own responsibilities.
Let me conclude with a citation. On the eve of the changeover, I wrote a commentary on diversity and monetary policy in the euro area. To the question whether a single one-size monetary policy could fit all parties involved – be they national entities, social partners or economic actors – my answer was: “One size must fit all”. The political decision on the creation of EMU had resolved all discussions on whether monetary union should precede or follow political unity and the fulfilment of the criteria for an optimum currency area. Today, in light of the evidence gathered so far in the euro area, I am more confident in saying: “One size does fit all!”
Reproduction is permitted provided that the source is acknowledged.