THE INTERNATIONAL ROLE OF THE EURO
ECON 101

Part I: THE EURO AS AN INTERNATIONAL CURRENCY

	1. The international use of major currencies since 1974
	2. The euro as an international currency

Part I: THE EURO AS AN INTERNATIONAL CURRENCY

An international currency is one that is used by the residents of countries that are not the country of issue. International currencies as well as national currencies fulfill three functions: as means of payments, as unit of account and as store of value. According to Krugman (1991), these functions can be subdivided into six, when taking into account the use of the currency by private and public sectors (Table 2).

Table 2: The functions of the international currency

Source: Krugman (1991).

As a means of payments, an international currency is used by non-residents for trade and capital flows. Private non-residents use the international currency as a vehicle, i.e. as an intermediate value in transactions between two smaller currencies. Typically, transactions between Portugal and Thailand are split between escudo/dollar and dollar/baht transactions. The monetary authorities also use international currencies as means of payments, when they intervene in the foreign exchange market.

Figure 1: The main synergies between the international monetary functions

As a private unit of account, an international currency is used to invoice, i.e. to set the price of goods and of assets, as well as when issuing bonds or defining a bank loan. This function is different from the means of payments function, since prices may be set in one currency, and payments in another. National authorities also can use the international currency as a unit of account, when they peg their currency to it.

As a store of value, an international currency is used both by the private sector and by the public sector to maintain the value of savings. The motivation of private investors is an optimal trade-off between return and risk diversification. The motivation of the public sector differs across to the exchange rate regimes. It can resemble private holders optimisation, or be devoted to exchange rate management.

The private and public functions of a specific currency can experience unequal developments: one currency can be chosen as a vehicle for transactions and be seldom used for public interventions in the foreign exchange market. A large part of the trade of a country can be invoiced in a specific currency, while the latter is not chosen as a monetary anchor. Finally, official reserves will be denominated in a specific international currency only if monetary authorities want to stabilise the exchange rate against this currency, whatever the composition of private portfolios.

However, there are synergies between the various functions of the international currency. These synergies use various channels (Figure 1):

• Transaction costs (bid-ask spreads): because the market for the international vehicle is large and deep, transaction costs are low, so the monetary authorities will likely use the same currency for their interventions, and private investors will likely hold assets denominated in that currency.
• Security issues: the denomination function determines the supply of securities; the availability of securities in a specific currency is necessary for the currency to expand both as a store of value and as a means of payments.
• Policy incentives: the monetary authorities will have an incentive to peg their currency to an international currency if a large share of trade and capital flows are denominated in that currency (because this will protect the domestic economy from harmful exchange rate fluctuations).
• Policy instrument: when an international currency is used as an anchor, it is necessary to hold official reserves and to intervene in the currency in order to defend the peg.
• Risk: the incentives to denominate trade and capital flows in a specific international currency, as well as to hold assets in this currency either for the store of value purpose or for the vehicle purpose, increase if the currency is used as a monetary anchor. This is because the risk of exchange rate variations is lower, which makes hedging either useless or less costly ⁽¹⁾.

We first outline the present status of international currencies for the various functions mentioned (Section 1). Then, the conditions for an internationalisation of the euro are discussed (Section 2). Although it is a crucial determinant of the risk borne when using the international currency for all private functions, the anchor function is often neglected in the existing literature. Special focus is given to this function in this report, based on original estimations presented in Appendix.

1. The international use of major currencies since 1974

Since the breakdown of the Bretton Woods regime, the US dollar has no longer been the institutional key currency of the IMS. However, neither European integration nor the affirmation of Japan as a major economic and financial power have entailed the end of the domination of the dollar as an international currency. Both the DM and the yen still play a modest international role compared to their economic and political weights, although their roles vary across the monetary functions..

1.1. Official reserves and interventions: no strong diversification out of the dollar

Over the last twenty years,theshare of the dollar in official reserves has been falling. From 76.1% of official foreign exchange reserves held by central banks around the world in 1973, the share of the dollar fell to 58.9% at end-1996 (IMF Annual Report, 1997 p. 159). During the same period, the share of the main European countries grew from less than 15% to almost 25% ⁽²⁾. However, the dollar's share in official reserves was twice as large as that of European currencies in 1996.

Although it has sharply declined, from 76.1% in 1973 and 93.3% in 1976 to 55,5% in 1996 (IMF Annual Report, 1997), the share of the dollar in official reserves of industrial countries is still higher than the share of any other single currency, and higher than it was ten years ago (54.8% in 1986). The decline in the dollar has benefited all other currencies, but mostly the DM whose share rose from 7.1% to 16.4% between 1973 and 1997. The share of the yen has risen from almost zero in 1973 to 5.9% of the total of identified official reserves in industrial countries, although it has been declining since 1991 (IMF Annual Report, 1997, p159).

The decline in the share of the dollar may partly be explained by the dollar's trend to depreciate. Composition effects are also important. Specifically, US official reserves increased from 0% of world reserves in 1973 to 3.7% in 1994. In addition, the share of other European countries in the official holdings of all the industrial countries rose as an effect of the creation of the European Monetary System. From 52% at the end of 1978, the share of European central banks, excluding the Bundesbank, abruptly increased to 62% at the end of 1979. Thus, the apparent decline of the dollar's share may not reveal a fundamental movement towards more diversified reserves in industrial countries.

As a whole, developing countries did not diversify their reserves out of the dollar. The yen's share increased at the expense of that of European currencies. This phenomenon can be explained by the fact that Asian reserves, which include more yen than other LDCs reserves, increased more rapidly than total world reserves in the eighties. However, the increase in the yen's share was limited by the persistent use of the dollar as a nominal anchor. In the mid-1990s, some Asian countries started to diversify their official reserves. Indonesia increased the share of the yen in its reserves from 27% to 35% in 1994 and reduced the share of the dollar from 52% to 49%. China announced its willingness to divide its reserves into equal parts between the dollar, the mark and the yen, and recently in September 1997 announced its intention to use the euro. Taiwan reduced the dollar's share. However, the fall in Asian reserves during the 1997 currency crisis, and the subsequent relinquishing of dollar pegs may have lasting effects on the composition of LDCs reserves.

1.2. The dollar still prominent as a vehicle

In the foreign exchange markets, the dollar, in 1995, was used in more than 80% of two-way transactions, the DM in 37%, the French franc in 8%, other EMS currencies in 13%, the pound sterling in 10% and the yen in 24% (Table 3).

Table 3: Foreign exchange turnover in April 1995
Daily averages for spot, outright forward and foreign exchange swap transactions
(Total: USD 1571.8 billion)

Source: BIS, Central Bank Survey of Foreign Exchange and Derivatives Market Activity, 1996, p.8.

It is interesting to note that if, as expected, the share of the dollar (90%) was higher in 1989, so were the respective shares of the yen (27%) and of the sterling. The DM has gained 10% (27% in 1989), the French franc 6% (2% in 1989).

The overwhelming use of the USD in the foreign exchange market proves that the USD is used not only for transactions between US residents and non-residents, but also as an intermediate currency in transactions between third currencies. This is the strict definition of a vehicle, in the foreign exchange market.

Still, the Deutsche mark is used as a vehicle, but on a regional basis. For instance, a Danish kroner/French franc transaction usually goes through the Deutsche mark (Danmarks Nationalbank, 1992, quoted by Hartmann, 1997b).

1.3. Slow diversification of denominations

The dollar declined as a trade invoicing currency from 56% of total world trade in 1980 to 48% in 1992 (Table 4). This decline was partly due to composition effects (for instance, the OPEC countries' share of world exports fell from 16% in 1980 to 5% in 1992). An increasing part of world exports is invoiced in the importing country's currency. Only Japan is invoicing a larger share of its exports in its own currency (40% in 1992 compared with 29% in 1980), which can be interpreted as a standardisation of its behaviour. Yet the dollar remains the only currency used as a vehicle, i.e. as an invoicing currency for trade between countries other than the issuing country. Even for intra-EU trade, the DM is hardly used as a vehicle (Ecu Institute, 1995).

Table 4: Denomination of international trade
Shares of the major currencies in denominating international trade

* share of the currency in world exports/share of the issuing country in world exports.
Source: Ecu Institute (1995).

As far as the denomination of international bonds and notes is concerned, the dollar is still dominant in floating rate issues (70.3% in 1996), while its share is less than 40% for straight fixed rate issues. (Table 5). Yen and DM issues each only represent 7-8% for floating rate issues and 14-18% for fixed rate issues. Although the dollar's share was smaller for 1995 issues, announced issues displayed stabilisation in 1997.

Source : BIS (1997)

1.4. The dollar remains the main de facto anchor outside Europe

The use of an international anchor is frequent in LDCs and in transition countries. In these countries, pegging the currency to a foreign one helps achieving disinflation despite the lack of reputation of the monetary authorities. It also reduces uncertainty for foreign investors. Finally, when some flexibility is introduced through a crawling peg, such a policy contributes to maintaining a stable real exchange rate, which is favourable both for promoting exports and for attracting foreign direct investment.

Exchange rate regimes are often classified according to the commitment of the monetary authorities, from no commitment (free float) to a complete commitment (currency board). Intermediate regimes include crawling pegs (the peg moves according to a pre-announced schedule), pegs with fluctuation bands (the exchange rate can fluctuate within some pre-defined margins) and a managed float (the exchange rate is stabilised, but without any specific commitment). Exchange rate regimes are also classified according to the anchor currency(ies) (single currency or basket).

The structure of exchange rate regimes has changed since 1978 (Table 6). Fewer currencies are pegged to the dollar and conversely more and more follow crawling pegs, or managed and independent floats. In 1978, almost one third of the countries had their currencies pegged to the US dollar, and only one fourth opted for independent or managed floats. By 1997 only one tenth of the currencies were pegged to the dollar and more than a half were officially floating.

Table 6 : Exchange rates regimes (number of currencies)

Source: IMF, Exchange Rate Restrictions and Exchange Rate Arrangements, various issues.

However, the distinction between a fixed peg and a managed float is not easy when the peg is frequently adjusted. At the same time, a fixed exchange rate is always adjustable, except under a currency board or in a monetary union. It is also possible to have a fixed, pre-announced central rate with discretionary fluctuation bands, as in France, where there is a discretionary, narrow band, inside the wide +/- 15% official fluctuation band.

Indeed, official exchange rate regimes do not always fit the practice of exchange rate management. In Asia, for instance, before the 1997 currency crises, most exchange rate regimes were managed floats, whereas currencies were de facto pegged to the USD (Bénassy-Quéré, 1996). More generally, the dollar has remained the main anchor currency outside Europe, which could explain why the its weight in LDCs official reserves has not declined.

1.5. Private portfolios: a substantial decline of the dollar

International portfolios of the private, non-banking sector include securities and eurocurrency deposits. Banks also hold bonds and stocks in foreign currencies, together with international loans. The whole private portfolio has experienced a significant diversification out of the dollar since the early 1980s: aggregated assessments made by the Ecu Institute show that, the share of European currencies in the world private portfolio increased from 13.2% to 36.9% between 1981 and 1995, while the share of the yen rose from 2.2% to 11.5%, and the share of the dollar fell from 67.3% to 39.8%. The DM alone contributed 15.6% in 1995 (Table 7).

Table 7 : Share of world private portfolio (%)

Source: Ecu Institute (1995).

However, the decline of the dollar's share varies across the types of assets.

a. International bonds and notes

In September 1996, the outstanding amount of international debt securities denominated in EU14 currencies or ECUs was $1,056.3 billion (BIS). The corresponding figures for dollar and yen securities were $1,139 billion and $520.8 billion respectively.

The total outstanding amount of international debt securities was $349.1 billion for the United States, $1406.1 billion for the EU countries and $360.4 billion for Japan in September 1996. The ratio of currency shares over country shares (internationalisation coefficient) thus was 0.75 for the European Union, 1.45 for Japan and 3.26 for the US. Thus, European currencies remain under-represented compared to the very high share of European issuers in international bonds and notes markets.

However, the data suggest a substantial decline in the role of the dollar from 62% of the stock of bonds outstanding in 1985 to 39.6% at the end of 1996, with a sharp rise in the share of yen denominated bonds to 16.9% (BIS 1997, International Banking and Financial Market Developments, p.41).

b. Bank loans

The share of the dollar in the international cross-border positions of the banks in foreign currencies was still around 50% in March 1997, whether in assets or liabilities, in total positions or in positions vis-à-vis non-banks ( BIS Monthly Report, August 1997, Table 4). This share is 10 points higher than the weight of the dollar in international bonds. Nevertheless, the share of the dollar was 75% in 1977 and 65% in 1985. Thus, there has been a significant decline of the role of the dollar as a store of value for banks (euro-loans) and for non-banks (euro-deposits).

Altogether, the currencies of the European Exchange Rate Mechanism (ERM) altogether represent 23-28% of the total cross border positions in foreign currencies and 29-32% of cross border positions vis-à-vis non-banks. This is much more than the shares of the yen (4-6%): the development of yen-denominated international assets seems to be limited to bonds and notes.

The rise of the yen in international bank loans (from almost 0% in 1977 to 6.1% in March 1997) can be explained by the decline, since the eighties, of the share of the dollar in developing country debt vis-à-vis the industrial countries' banks (Bénassy 1996). This movement has mainly benefited the yen in Latin America and over all in Asia where the yen's share was already 28% at the end of 1988. Despite this rise, it should be pointed out that the Japanese currency remains under-represented compared to the weight of Japanese banks in total cross-boarding asset positions: the yen represents only 9.4% of total asset positions in all currencies (international and domestic), while banks located in Japan totalise 13.2% of the total (BIS Monthly Report, tables 4A and 2A).

1.6. Summary

The decline in the dollar as an international currency is already taking place, leaving some room for the euro. However, this decline is partly due the ERM, which enlarges the regional role of the DM as a monetary anchor and as a reserve currency. In addition, the decline of the dollar is more pronounced for the private store of value function than for the public store of value, vehicle and unit of account functions (Table 8). This discrepancy can be explained by inertia which is less determinant for the store of value function than for other functions. EMU will constitute a huge shock in the sense that it will suddenly create a large zone with deep financial markets, no intra-zone exchange rate risk and an independent central bank. This shock may override inertia and favour the emergence of the euro as an international currency. However, the outcome of the balance between inertia and size effects is uncertain, at least in the short run. The factors of internationalisation are discussed in Section 2.

Table 8: Summary statistics on the present internationalisation of the main currencies

* Exact composition depending on the topic, but always including £, FF, NGL, Ecu.

2. The euro as an international currency

In the past, the internationalisation of a currency usually started with its use as a means of payment for trade, before it was used as a store of value and finally as a unit of account (Bourguinat, 1992). It is believed that the first international currency was Alexander the Great's currency which was widely used in Asia Minor in the 3 ^rd century b.c. Ever since, the vehicle function for trade has been the key determinant of the internationalisation process. In more recent years, the Bretton Woods system of fixed exchange rates against the US dollar was coupled with the Marshall plan which boosted the US as the major goods supplier of Europe.

Today, capital flows are forty times larger than trade flows ⁽³⁾. Thus, the use of the international currency for capital flows seems determinant. The emergence of the euro will be slow, and not automatic. It will depend on various factors among which policies aiming at making European financial markets more efficient, and exchange rate policies of non-EMU European countries vis-à-vis the euro.

2.1. How strong is inertia ?

The pound sterling remained an international currency for half a century once the United Kingdom had lost its leading position as an economic superpower after World War I. As pointed out by Bourguinat (1992) and Kenen (1993), hysteresis characterises the internationalisation process. Clearly, the currency that is already used as an international currency benefits from a strong inertial bias. All agents ­ currency traders, lenders and borrowers, exporters and importers, private and public sectors - are more likely to use the currency that everyone else is using. This is because the international currency benefits from economies of scale and from network externalities.

2.1.1. Economies of scale

Economies of scale occur mainly because transaction costs are lower for larger volumes.

On foreign exchange markets, transaction costs can be measured as bid-ask spreads ⁽⁴⁾. These costs are small for interbank transactions: for a $10,000 transaction, the typical quoted cost is $5, which means a 0.05% cost  ⁽⁵⁾: the equivalent of a big hamburger compared to the value of a regular car. However, bid-ask spreads reflect not only the liquidity of each market, but also the volatility of the exchange rates (they increase with volatility). As exchange rate volatility increases with the turnover, the impact of a larger turnover on transaction costs is ambiguous a priori. Nevertheless, Hartmann (1997a) shows that only surprises in the daily turnover are related to increased volatility. Because size effects due to EMU will not be surprising, it can be assessed that EMU should reduce bid-ask spreads.

Alogoskoufis et al. (1997) use Hartmann's estimates to compute bid-ask spreads with zero volatility (Table 9). The differences in transaction costs due to differences in liquidity are very small: the largest difference reported in Table 9 is only 55 cents on every $10,000 transaction. However, for only the spot market, these transaction costs amount to at least $160 million a day.

Table 9 : Spot foreign exchange transaction costs
($ for a $10,000 transaction, assuming zero volatility)

Source: Alogoskoufis, Portes and Rey (1997)

According to Hartmann (1997), a 1% increase in the turnover reduces transaction costs by 0.03% approximately, for a given exchange rate volatility. From Section 1, it can be stated that rebalancing the role of the DM and of the USD in the forex market would entail a rise in the DM/yen turnover by 100% and a fall in the yen/USD turnover by 50% approximately. With such assumptions, the transaction cost for a $10,000 transaction between the DM and the yen would decline from $4.37 to $4.24, whereas the transaction cost for a $10,000 transaction between the yen and the USD would rise from $4.16 to $4.22. Hence, it would become about as cheap, for a Japanese investor, to use the DM or the USD as vehicles in the foreign exchange, i.e. as intermediate currencies when they buy third currencies (See Box 1). However, there is no reason why the DM/yen turnover should suddenly increase and the yen/USD turnover be reduced. Since transaction costs are lower for the transactions involving the dollar, there is a strong incentive to use the dollar for new transactions, which reinforces the cost advantage of the USD. Thus, economies of scale maintain the existing status of currencies as vehicles on the foreign exchange markets.

To a lesser extent, economies of scale on transaction costs also affect the store of value function: when two assets only differ in their associated transaction costs, a liquidity premium has to be paid to the investor to equalize the yield of the two investments. The liquidity premium is low if yields are certain, because the transaction cost is low compared to the expected return. However, as exchange rates are volatile, there is much uncertainty about the yields. Thus, international portfolio reallocations are frequent, and transaction costs add up. Alogoskoufis et al. (1997) estimate that the USD benefits from a liquidity discount of 25-50 basis points. It means that, other things equal, US interest rates are lower by 25-50 basis points, due to the fact that the dollar market is more liquid  ⁽⁶⁾. This liquidity discount is based on low transaction costs on the foreign exchange market and also on the domestic, financial market. The latter are important since the store of value function is concerned: when international investors buy and sell US financial assets against US money, they have to pay a transaction cost which can be measured by the bid-ask spread. Due to greater liquidity, this cost is lower for the US market than for any other market.

Table 10 : Bid-ask spreads on 10-year government bonds
($ for a $10,000 transaction)

Source: Alogoskoufis et alii (1997)

To sum up, persistent differences in transaction costs will likely slow down the internationalisation of the euro as a vehicle and as a store of value, which could in turn limit the internationalisation for other functions in the short run (see Section 2.2.3).

Box 1: transaction costs

Take a Japanese investor who intends to buy assets denominated in European currencies (later in euros) or in dollars. For the sake of simplicity, we shall use the terms euros for both pre-EMU European currencies and the forthcoming single currency.

To buy euro-denominated assets, he can sell yens directly against euros or prefer indirect exchange through the dollar. In the first case, he will pay transaction costs on the exchange of yen against euros plus domestic euro transaction costs, on the stock exchange for instance. In the second case, he will pay exchange transaction costs when he buys dollars with yens, exchange transaction costs when he converts his dollars into euros, and domestic euro transaction costs. Let us call Ed the cost of direct transaction and Ei the cost of indirect transaction through the dollar. If Ed>Ei, the dollar is used as a vehicle as in the present situation (except for Deutsche-mark assets).

To buy dollar denominated assets, he can choose between direct and indirect exchange. In the first instance, the direct transaction, he will pay transaction costs on the exchange of yen against dollars, and the American domestic transaction costs (total cost: Dd). For the indirect transaction, he will pay transaction costs on the exchange of yen against euros, on the exchange of euros against dollars, and the American domestic transaction costs (total cost: Di). If Di>Dd, the euro is not used as a vehicle to buy dollars, and if Ed>Dd, the dollar is prefered as a store of value. Both features apply currently.

EMU may have the following implications:
Ed<Ei: the dollar is no longer a vehicle to euros,
Ed=Dd: the euro is as good as the dollar as a store of value,
Dd<Di: then the euro is not a vehicle for buying dollars.
Finally, suppose the Japanese investor intends to buy forint-denominated assets.

He can choose direct exchange, and pay the yen/forint transaction costs plus the Hungarian domestic transaction costs (total cost: Fd). Or he can choose indirect exchange through euro (cost: FEi) or dollar (FDi).
If Fd>FEi>FDi, the dollar is used as a vehicle as it is now often the case.
Conversely, EMU may lead to Fd>FDi>Fei: the euro would then be used as a vehicle.

2.1.2. Network externalities

Money is a network good: the more people are using it, the larger the incentive to use it. Network externalities are different from economies of scale in the sense that they do not work through prices. On the contrary, externalities appear when a market (and thus, a price) is missing. The text-book example for network externalities is that of telephone: the connection price does not rise when more people are connected to the network, although the welfare of being connected increases. This raises the incentive to use the network that is already used extensively, even if it is not the best one.

The case of the international currency is similar: when using the international currency for its various functions, agents do not have to pay an additional price for its international status. Still, the utility is higher when using the international currency, because it reduces uncertainty:

• the financial market of the international currency offers a large range of instruments that better meet the needs for hedging (store of value function).
• the profit uncertainty is lower if the same currency is used for invoicing exports and imports, and external competitiveness is more stable if prices are set in the same currency as those of the competitors (private unit of account function);

• treasury management is easier if only one foreign currency is used. This argument applies to small firms which seldom hedge exchange rate risks (private means of payment function).

2.1.3. Synergies between the various functions

The synergies between the various functions of the international currency have been detailed in the introductory section. The use of the international currency for a specific function (i) reduces the cost and (ii) produces positive externalities to the other functions.

(i) Reduced cost: we have already underlined the synergy between the vehicle function and the store of value function. This is because both functions benefit from low transaction costs, and transaction costs are lower the deeper the market. There is also a synergy with the unit of account function, especially with the anchor function, because defending a peg requires official reserves and interventions with transaction costs. In addition, the use of a currency as a unit of account reduces the information costs when using it both as a means of payment and as a store of value, since it is no longer necessary to forecast exchange rate fluctuations.

(ii) Cross function externalities: the use of a currency as a unit of account reduces uncertainty when using it both as a means of payment and as a store of value, especially if the international currency is used as a monetary anchor. Conversely, monetary authorities have an incentive to peg their currency to the international anchor if the net foreign asset position and security issues are denominated in that currency, and if trade is largely invoiced in that anchor. Finally, holding official reserves in the international currency is more useful when the domestic currency is pegged to it.

Hence, a low internationalisation in one function may impede the further internationalisation in other functions. This may have happened to the DM, whose internationalisation could have been slowed down by the weakness of the vehicle function. It may also happen to the euro if size effects are less powerful for some functions than for others.

2.2. Although an advantage, size will not automatically trigger internationalisation

Historically, size and currency status have been correlated. First the pound sterling and then the dollar became dominant during the periods when the United Kingdom and the United States were the world's main economies and traders. Nowadays, the only significant international currencies are those of the world's three largest economies and traders: the United States, Germany, and Japan. However, the above analysis shows that the size of the financial markets will be determinant.

Table 11 : Various measures of size

⁽¹⁾ Austria, Benelux, France, Germany and Ireland.
⁽²⁾ The EU15 figure excludes intra-EU trade.
⁽³⁾ Amounts outstanding
Sources: Funke and Kennedy (1997) and Hartmann (1996).

2.2.1 EMU GDP will compare with US GDP

Back-of-envelope estimates of portfolio reallocations made by Bergsten (1997) are based on the size of GDP. The GDP of the monetary union will range between 62 and 112% of US GDP, according to the number of countries involved in EMU, while German GDP is only 31% of that of the US.

GDP is important since the absolute amounts of private investment and of public deficit grow with GDP, leading to a parallel development of financial markets. In addition, a zone with a large GDP is an important trade partner for foreign countries, who have an incentive to use the currency of that zone for various functions, especially as a unit of account and as a store of value.

2.2.2 Large EMU exports will not necessarily entail an internationalisation of the euro

Excluding intra-European flows, the European Union 15 will remain the largest exporting zone in the world (Table 11). According to Grassman's law (Box 2), this should boost the euro as an international invoicing currency and means of payments.

However, the issue is not that simple. Friberg (1997) shows that Grassman's law no longer held for Sweden in 1993, and that invoicing in a third currency was not an uncommon practice.

Box 2: Grassman's law

S. Grassman studied the currency denomination of Swedish exports and imports in 1968, and found that exporters tended to invoice in their own currency: 2/3 of Swedish exports were invoiced in kronas but only ¼ of Swedish imports, 12% of the exports were invoiced in dollars although only 8% were sold to the United States. Other empirical studies in the 1970s found the same pattern.

In fact, exporters have the choice between (i) pricing in their own currency, which entails certain profit rates and uncertain demand volumes, (ii) pricing in the importer's currency, which leads to certain demand volumes and uncertain profit rates, or (iii) pricing in a third currency, with uncertain profit rates and uncertain demand volumes. The outcome of this choice mainly depends on market structures.

When price competition is tough, i.e. when demand is reactive to small price variations (as in the energy and raw materials markets), exporters have little incentive to price in their own currency. This is because the world price is given. Thus, exchange rate variations must be fully compensated by domestic price variations. In such markets, pricing in a reduced number of currencies helps reducing information costs for both exporters and importers. Thus, pricing in a third currency is common practice. As forex transaction costs must be supported both by the exporters and by the importer, there is a strong incentive to use the dollar which benefits from low transaction costs. The potential use of the euro in such markets will depend on the transaction costs on the euro.

When the products are differentiated, small price variations (in the importer's currency) do not have dramatic effects on demand volumes. However, exchange rate variations are all but small. Goldberg and Knetter found that in the United States, since the 1970s, dollar prices of foreign products have not responded fully to exchange rates. A price response equal to one half the exchange rate change would be a good estimation. This muted price response (in the importer's currency) can be explained by price discrimination  ⁽⁷⁾ in a world of global firms. A second explanation is that menu costs lead to lagged price reactions to shocks ⁽⁸⁾. A third explanation is that transfering the forex hedging cost from the exporter to the importer needs price compensation from the exporter, which leaves little incentive for the exporter to invoice in his own currency. This could explain why exports tended to be invoiced in the exporter's currency when the variability of exchange rates was low (before 1973), but no longer when it increased: exporters did not want to suffer from demand variability for a small benefit in terms of hedging costs.

The latter argument shows that the variability of exchange rates matters: those exporters from countries maintaining relatively stable exchange rates against the euro (ERM II countries, Central and Eastern European Countries and African countries) will surely have an incentive to invoice in euro, because this will remove all hedging costs. Friberg (1997) goes further, concluding that pricing in a third currency should be preferred if the variability of the third currency vis-à-vis that of the consumers is sufficiently low, relative to the variability of the exporter's currency vis-à-vis the consumers' currency. Hence, pricing in euro as a third currency will be chosen if the euro is more stable against the importer's currency than is the exporter's own currency against the that of the importer.

To sum up, the potential role of the euro for trade invoicing will be related to the level of transaction costs in euro for for goods that are little differentiated (like energy or raw materials), and to its variability against key currencies for differentiated goods (manufactured items). In any case, seeing that capital flows are forty times larger than trade flows, it is obvious that the key factor of the internationalisation of the euro will be the store-of-value and of vehicle functions. Portfolio allocation and domestic financial market's efficiency are pushed in the forefront.

2.2.3 A rush of international investors into the euro should not be taken for granted

In 1999, intra-European debts will become domestic, and international assets will have to be measured net of domestic holdings. Although the Group of Ten members of the EU produce about one third of the world GDP and have, net of their intra-EU trade, a similar share of international trade, the proportion of international assets denominated in European currencies (net of intra-European holdings) can be evaluated only at one-eighth of world GDP (MacCauley & White, 1997). In order to bring the European share of international assets to the level of the EU contribution to world GDP and trade, holdings in euro would have to rise by more than $700 billion, an amount roughly equivalent to 12% of the outstanding domestic debt of the EU 7 and to 15% of GDP. Most evaluations estimate the portfolio reallocations in a range of $500 billion to $1 trillion (Bergsten, 1997), twice to three times the total amount of exchange reserves of the European Union countries at the time of the Maastricht Treaty. These huge figures assess the magnitude of the potential portfolio reallocation. But they do not guarantee that such a reallocation will take place.

a. There will be some diversification out of the euro

Firstly, there will be some diversification out of the euro, for two reasons.

(i) The present diversification of institutional portfolios is low in Europe. Although the EU matching rule (liabilities in a foreign currency must be 80% matched by assets in the same currency) will apply when diversifying outside the euro, this rule will not be binding at least in the short run: EU institutional investors will be able to reach Japanese or British diversification ratios without transgressing the matching rule.

Table 12 : Diversification rates of institutional investors portfolios

Source: Artus (1996)

The question is whether institutional investors will need to diversify outside the euro. It can be argued that they will have large opportunities to diversify within the euro, through holdings of private and public securities of various EMU countries, and that the EU matching rule will never be binding within the euro. However, currency diversification will no longer be possible within the EMU zone.

(ii) Currency diversification: when the intra-European exchange rates are definitely fixed, institutional investors of all countries will no longer be able to reduce the exchange rate risk of their portfolios by holding assets in various European currencies. Indeed, the standard portfolio choice model shows that, unless expected returns are higher for the euro than for average European currencies, private holdings in euros should be lower than private holdings in European currencies: risk diversification will entail a higher share of non-euros holdings. However, this argument already applies to the ERM core whose currencies are already very stable relative to each other. Thus, the diversification out of the euro should be limited (Bénassy, Italianer and Pisani-Ferry, 1994; Arias, 1997). It may be larger if monetary unification and the uniformisation of financial markets lead to higher correlations between asset prices and returns.

b. The behaviour of the ECB will be crucial in the short run

The amount of the reallocation will depend on the monetary policy of the European Central Bank, as well as on the confidence investors will have in future policy. The independence of the ECB will favour confidence; but uncertainties about future members or about fiscal cooperation may reduce this confidence. In addition, in the absence of price stability reputation, investors may take the nominal exchange rate of the euro as an indicator of monetary policy. Large fluctuations in the euro/dollar exchange rate may lead to scepticism vis-à-vis the euro. At present, this is not the market opinion: a survey reported by Artus ⁽⁹⁾ showed that 490 out of 500 Asian investors want to diversify their holdings out of the dollar into the euro. However, the market opinions are volatile, and one can never exclude weak demand for euros.

c. Transaction costs will not automatically be reduced

A large portfolio shift will not necessarily imply the parallel development of turnover which depends more heavily on transaction costs, the latter relying on the size of the market (see Section 2.1.1).

Finally, currency traders, lenders and borrowers will not change their behaviour if European financial markets are not at least as efficient as American ones. This crucial point is studied below.

2.3. The efficiency of the financial market will be crucial for the emergence of the euro

The achievement of the full potential financial market benefits of EMU cannot be taken for granted, and in any case it will take time. The introduction of the euro is an opportunity to dismantle the barriers between the segmented European financial markets (bank deposits and loans market, securities, financial services, etc). The euro may enhance the impact of EU financial directives, increase transparency in credit evaluation, accelerate the integration of the European financial market and expand Europe's institutional investor base.

After many previous steps, the introduction of the euro is the most significant move towards monetary integration. In the area of financial and monetary integration, the EU Second Banking Directive, the Capital Adequacy, the Investment Services, and other financial directives have been implemented. But European financial markets still remain segmented. Full implementation of the EU Investment Services Directive (which creates a single passport for securities firms, portfolio managers and investment advisories) will enhance the impact of these structural changes.

2.3.1. Banking competition and securitisation of the financing is not yet completed

Restructuring through consolidation has largely occurred at the wholesale level in the European banking system, but at the retail level, Europe is overbanked. Cross-border competition will be enhanced. Operational efficiency as well. This will provide additional pressures for consolidation. On the other hand, national barriers still exist and inefficient retail banking systems might call for public support and delay adjustments.

Traditionally, European firms finance their activities mostly through indirect finance, bank loans amounting 54% of all outstanding financial assets in Europe, at the end of 1995. By contrast, in the United States, bank loans amount only 22% of total assets outstanding, because US firms rely more heavily on bond and equity financing. This divergence tends to be more noticeable as the quality of the credit diminishes: small- and medium-size enterprises almost never issue junk bonds and equities in Europe. However, the European market for corporate bonds is roughly 3/4 the size of the U.S. market, but the bulk of these bonds were issued by European financial institutions. Bonds account for a small share of the total liabilities of non-financial firms (less than 1% in Germany ⁽¹⁰⁾). Again, in short term financing, European companies rely heavily on banks. By contrast, U.S. firms rely on short-term financing because they have access to a very liquid commercial paper market, which accounts for more than half of the world's outsanding of banking law of 1984, but European corporate securities markets have remained relatively underdevelopped.

Driven by financial deregulation and disintermediation, European securities markets have become more liquid and integrated. Large sovereign debt issues have developed more efficient secondary bond markets. Concomitantly, economic stability and the convergence of macroeconomic policies have permitted greater capital mobility. As the links between national securities markets tighten, and bond spreads are reduced.

The introduction of the euro will likely foster further securitisation of European finance. Greater uniformity in market practice, more transparency in pricing will increase market integration. The elimination of currency risk, greater uniformity in market practices, the convergence of credit spreads should increase the depth and the liquidity of European securities markets. The design and implementation of monetary-policy operating procedures will be especially important, since EMU financial and monetary policies can be used to encourage or discourage the development of deep and liquid EMU-wide euro securities markets. The proposals of the Giovannini Report (DGII, 1997) aim at favouring financial liquidity through various harmonisation techniques.

Technological progress will enhance the impact of the introduction of the euro, because the location of trading, clearing and settlement will become less and less relevant. Whether technological progress leads to centralize activities in one or two locations, or whether it leads to the development of more locations (national markets), the securities and derivative markets will soon be fully integrated.

2.3.2. Small intra-EMU credit spreads may remain

Integration of the private security market in euros could proceed rapidly, at least for major firms. Some analysts say that institutional investors may switch from a country to a sectoral approach, privileging the largest European firms at the expense of the smallest, whatever the country  ⁽¹¹⁾. However, it is often argued that the government bond market in euros may remain fragmented. There will no longer be any sovereign issuer in EMU, since member countries will not control money creation and the Treaty prohibits the central bank bailing out governments. Ratings will deteriorate for the states showing the highest debt to GDP ratios, compared to the least indebted. Punishment for fiscal laxity may come from the financial markets through significant and persistent credit spreads.

On the other hand, the no-bail-out rule is often said not to be credible since it would be difficult to let an EMU member state go bankrupt. This alternative view is reinforced both by the management of the recent Asian crisis, and by the convergence of European long term interest rates (which already include the EMU regime), at least for the sovereign debt of Spain and Belgium denominated in DM (McCauley and White, 1997). In addition, credit risk will be reduced in Europe by the Growth and Stability Pact.

2.3.3. The international money market will likely favour US Treasury bills in the short run

As far as liquidity is concerned, US Treasury bills will likely be prefered, at least in the short run. The US market operates 24 hours a day, anonymously, and any quantity can be traded. It is deep and liquid. In fact, US Treasury bills are very nearly interest-bearing money. The closest market is that of British pound Treasury bills. Paris bons du Trésor market is small, and Germany virtually does not use Treasury bills. After EMU is completed, all members' treasury bills will be denominated in euros, but will still be in national securities. They will bear the same currency risk, and the risk due to interest rate variations will be very similar due to the unified monetary policy. However, the various Treasury bills will not be perfect substitutes, because of fiscal disparities and because few countries will be able to provide a large range of financial products (in terms of maturities, indexation, etc).

On the other hand, there will be a harsh competition between member states' Treasuries to provide sophisticated bills to the markets. Such a competition is already underway. It could lead to a very rapid enlargement of European financial markets. In addition, if the risk of interest rate variation is perceived as equal across European countries, asset holders will be able to use the Treasury bills of any country to hedge a specific risk.

To sum up, the efficiency of the European financial market will rise because of (i) increased competition among financial intermediaries, (ii) the rising role of institutional investors, (iii) the currency uniformisation of European securities, and (iv) competition among euro asset issuers. However, EMU by itself will not provide a market for Treasury bills as liquid as that of US Treasury bills. This may slow down the emergence of the euro as a vehicle currency.

2.4 The euro will be attractive as an anchor currency

The public unit of account function is often neglected when dealing with the competition for international money status. However, this function has a strong impact on the use of the international currency in other functions, for two main reasons:

(i) Defending an international peg requires official reserves (stock) and official interventions (flow) in the foreign exchange market concerned. Thus, the anchor function boosts the public store of value and means of payments functions. Due to increased turnover, transaction costs decline, which gives more incentive to all agents to use the same currency for the two functions.

(ii) The risk and cost of using a specific foreign currency for the various functions is lower when it is used as a monetary anchor, i.e. when the domestic currency is pegged to the foreign one. This is because a stable exchange rate makes hedging either unnecessary or less costly, and because pricing in this international currency (instead of another one) leads to more stable demand and to more stable profits.

However, Asian or Latin American countries will unlikely peg their currencies to the euro alone, at least in the early phases of EMU. This is because the European Union is not their main partner, and because these countries have been used to US dollar peg in the past.

A natural expansion area for the euro would be Central and Eastern European countries (CEECs). In the past, however, these countries have generally not pegged their currencies to core-EMS currencies.

The choice of an international monetary anchor is generally explained by the theory of optimum currency areas (Mundell, 1961; McKinnon, 1963). According to this theory, two countries A and B have an interest in fixing their bilateral exchange rate if they face mostly symmetric shocks (i.e., shocks of the same type at the same time in the same direction), if their bilateral trade is important, and if output factors are mobile between the two countries. This theory can be used to assess whether the anchoring strategies of various countries are optimal. On the basis of cross-section estimations for 36 countries, we show that, the CEECs should have attributed a higher weight to the DM in their implicit basket peg over 1992-1995 (Appendix).

The theory of optimum currency areas assumes that monetary authorities aim at stabilising the output growth rate. However, the development strategies of emerging countries are based on exports and direct investment inflows. Their final target in terms of output growth is dominated by intermediate targets in terms of price competitiveness and endebtedness capacities. This can modify the optimum currency area diagnostic. For instance, the implicit dollar peg of most Asian countries before 1997 crisis cannot be explained by the optimum currency area theory. However, this strategy was a typical non-cooperative strategy for them, given the distribution of their foreign trade (about 25% with the US and another 25% with Asian partners other than Japan)  ⁽¹²⁾. Taking into account the external constraints does not modify the conclusion for CEECs currencies however, because their external trade is mostly carried out with the European union.

Therefore, it can be concluded that there will be a strong incentive for the CEECs to choose the euro as a monetary anchor.

NOTES

1. Hartmann (1997a) shows that transaction costs are lower the smaller the exchange rate volatility. The role of the risk of exchange rate variations is enhanced by the fact that it cannot be fully hedged. For instance, an exporter cannot hedge the risk borne before a contract is signed. In the same way, there is no hedging for direct investment.

2. If ECUs issued against dollars are not assumed to be dollars.

3. At the time of the Treaty of Maastricht, in 1992, the world exports amounted to $10 billion a day and total daily spot trading to $394 billion.

4. The bid-ask spread is the discrepancy between the purchasing price and the selling price.

5. Hartmann P. (1997). Effective transaction costs are lower than quoted bid-ask spreads. Their typical value is $3 for a $10,000 transaction.

6. 10 basis points correspond to 0.1 percentage point.

7. Price discrimination means that the same product is sold at a different price in the various countries.

8. Menu costs are the costs which are supported when changing the price of the products. They cover the time spent when calculating the new price and the cost of publicising it (printing commercial documents, etc). Although small, these costs can explain why prices are not adjusted to optimal prices at any time (Akerlof and Yeelen, 1989).

9. Panel discussion, XIVèmes Journées Internationales d'Economie Monétaire et Bancaire, Orléans, June 1997.

10. Prati and Schinasi (1997).

11. See Financial Times, January 2nd, 1998.

12. See Bénassy-Quéré (1997).