Properties of the euro–Europe’s new currency
Allen, Andrew T
This paper analyzes the portfolio properties of the euro as it has been initially composed with 11 currencies and also in its anticipated future extended version comprising all 15 EC members. The period under investigation is January 1992 to October 1998. Using a U.S. investor’s perspective, investing in the euro is a low risk-low return alternative compared to investing in the U.S. dollar. Results of the Gibbons, Ross and Shanken multivariate test show that the composition of the European common currency is mean-variance inefficient.
PROPERTIES OF THE EURO – EUROPE’S NEW CURRENCY
The launch of Europe’s single currency, the euro, on January l, 1999, was without any doubt one of the most significant events in the economic and political development of Europe. In this major experiment in monetary arrangements, 11 European countries did unite in the European Monetary Union (EMU) with the European Central Bank (ECB) setting the short-term interest rate for the entire euro zone. The 11 countries that have adopted the euro are: Austria, Belgium, Finland, France, Germany, Ireland, Italy, Luxembourg, the Netherlands, Portugal and Spain. To participate in the single currency zone, countries were required to meet several criteria prescribed in the Maastricht Treaty of 1991, relating to inflation, interest rates, government debt, and exchange rate volatility. The four national currencies not taking part in the euro – the British pound, Danish kroner, Greek drachma, and the Swedish. kronor – will set fluctuation bands against the euro. These will be 15 per cent bands, as with the old Exchange Rate Mechanism, although some may want to keep their currencies in narrower bands. On January 1, 1999, the conversion rates between the euro and currencies from the 11 nations participating in the shared currency were set (see the appendix for the official exchange rates). All ECUs were automatically converted to euros on a one-to-one basis.
The introduction of the euro creates two important long-term factors which may lead to an increased demand for the euro at the expense of the U.S. dollar: ( 1 ) the use of the euro as a reserve asset and (2) as a means of payment. Jager and de Jong (1987) examined the attractiveness of the ECU as a reserve asset and find that its reserve function will be substantially undermined by the exchange rate stability that is pursued in the EMU. Following the euro’s introduction, Europe’s foreign exchange reserves will become somewhat `excessive’ as there will be little need for them to hold reserves as intervention in the foreign exchange market will be carried out by the ECB. Temperton ( 1997) estimates that the excess holdings of foreign exchange reserves could be as much as $300 billion. Attempts to reduce these reserves holdings could put a severe downward pressure on the U.S. dollar. On the other hand, EC authorities may not want to see the value of the euro strengthen too much against the dollar out of fear of losing international competitiveness. Undoubtedly, the euro will gain importance as a vehicle currency for international trade and finance. Although about half of world trade is denominated in U.S. dollars, the U.S. accounts for only 13% of global exports. Since the 11 euro-zone nations represent a much higher percentage of world exports, it is to be expected that a larger share of world trade will be denominated in euro, at the expense of the U.S. dollar. The introduction of the euro will, therefore, have far-reaching consequences for financial and commercial enterprises worldwide. According to Robinson (1998) as a result of the euro, corporate treasurers will see greater degrees of consistency and conformity because of money market and interest rate convergence, a concentration of banking relationships, and the rise of an efficient secondary financial market that resembles the U.S. secondary markets.
The euro’s arrival will also force Europe’s investors to formulate new strategies and expand the range of asset classes. As the euro spells the end of currency risk, the emphasis shifts from country-based investments to sectoral analysis and a widening of the universe in terms of including new markets not invested in before.
The portfolio properties of Europe’s newest currency has important implications for corporate management as well as investors. The euro becomes attractive to corporate treasurers to the extent that it offers a better reward/risk trade-off than its component currencies. Furthermore, how does the new currency compare in terms of return and risk with a minimum-variance currency cocktail of the 11 (or 15) EC members? Also for international investors, the return/risk characteristics of the euro are a major consideration in their asset allocation decisions across the international currency spectrum. More in particular, is the euro an attractive substitute for the U.S. dollar in terms of yield and risk diversification? Should the euro be held in lieu of a combination of the euro-zone countries currencies or, in other words, is the euro a mean-variance efficient currency portfolio?
The use of currency-cocktails, such as the Special Drawing Right (SDR) and the ECU, to mitigate exchange rate risk has been extensively researched, for instance, by Mathur and Ogden (1991) and Jorion (1991). This paper attempts to analyze the portfolio properties of the euro as it has been initially composed with 11 currencies and also in its anticipated future extended version comprising all 15 EC members. In particular, we investigate the attractiveness of the euro based on its performance as an investment currency for U.S. investors. The potential attractiveness of the euro arises from the way its value and interest rate – are determined, viz. weighted averages of the currency values and interest rates, respectively of the member countries of the EMU. A meanvariance framework is used to assess the return/risk properties of the euro. The efficiency of the euro is tested using the Gibbons, Ross and Shanken (1989) F-statistic. The F-statistic efficiency test is appealing because it is valid for small samples, while the chi-square efficiency test in Jorion (1991) is only valid asymptotically.
MEAN-VARIANCE PROPERTIES OF THE EURO
The period under investigation is January 1992 to October 1998. The starting date was chosen because of the Maastricht Treaty (December 1991 ) in which the conditions and a timetable for monetary union and creation of the single currency (euro) were set. The data were all taken from the Datastream database and consist of monthly observations for spot exchange rates (Reuters) and 1-month inter-bank interest rates or their equivalent. To study the properties of the euro, a reference currency has to be chosen that is not a member country. The U.S. dollar was chosen as the reference currency.
Numerous studies have found that the behavior of speculative prices could best be described by a non-Gaussian distribution. McFarland, Petit and Sung ( 1982), for instance, have found that bilateral exchange rates fit the Symmetric Stable Paretian (SSP) better than the normal distribution. Since the ECU (euro) is composed of bilateral exchange rates, it seems reasonable that excess returns of the ECU and its component currencies is distributed other than normally. The Jarque-Bera statistic, a test for normality based on the skewness and kurtosis measures combined, is reported in Table 1. The probability for the Jarque-Bera statistic is zero for Ireland and Italy, and very low for all European currencies including the ECU.
Another point to note from Table 1 is that the correlation matrix for the excess returns in the EMS contains a lot of very high correlations, reflecting the pegging of currencies within the EMS. The average correlation between the currencies initially adopting the euro is 0.81, while it is 0.76 between all EMS countries. Four countries, Austria, Belgium, Germany and the Netherlands, show correlation between their excess returns and that for the ECU of 0.98 or higher. It is well known that Belgium, Germany and the Netherlands had an agreement to keep their bilateral exchange rates within a 1% margin, while Austria has close links to Germany in terms of trade and monetary policy.
To examine the attractiveness of the euro for U.S. investors, we perform a mean-variance analysis. A set of efficient investment portfolios composed first of a selection of Euro-zone countries and next including the other four EMS countries. Using a mean-variance approach, the optimization problem involves finding the optimal weights that minimize the portfolio variance for a given level of expected return. The optimal (minimum variance) weights and summary statistics for different levels of expected returns are reported in Table 2 where no short-selling is allowed and Table 3 where short-selling is allowed.
As shown in Table 2, the (ex post) minimum-variance composition of these cocktails is very different from the actual currency weights in the ECU. At lower levels of risk, investment in Austria is high, while at higher levels of risk investments in Italy and Portugal dominate as both countries display high mean excess returns (Table 1).
Table 3 shows the minimum-variance composition when short-selling is allowed and when the ECU is included as one of the portfolio choices. Table 3 shows that a large long position in the ECU is required in all of the minimum-variance portfolios.
The efficient portfolio reported in Tables 2a and 2b are plotted in an excess return/variance graph in Figure 1. The ECU has also been entered in terms of mean return and volatility. Figure 1 shows that extending the asset selection to all 15 EMS countries results in a dominating efficient frontier. This means that the 15currency euro offers a higher excess return for the same level of risk and a lower amount of risk for the same level of excess return than the current eleven currency euro. The majority of the minimum-variance portfolios are largely composed of only two investments, U.K. and Greece. This is caused by the high excess return realized in Greece, the low volatility for British investments and the relative low correlation (0.49) between both markets (see Table 1 ).
As is clearly illustrated in Figure 1, investing in the euro is, ex post, a low risk – low return alternative. There are several efficient investment portfolios formed by EMSinvestments which may dominate investing in euro, i.e. much higher return for the same or even lower levels of risk. The euro will definitely bring stability, i.e. elimination of currency volatility, but offers a correspondingly low level of excess return abovethat for the U.S. dollar. The next section determines whether the euro by itself is a minimum-variance portfolio, and can be held in lieu of a combination of the 11 euro-zone countries.
MEAN-VARIANCE EFFICIENCY OF THE EURO
The ex ante efficiency of the euro currency portfolio is tested using the Gibbons et. al. (1989) multivariate test statistic, W, which follows an F distribution. Their test statistic is an improvement over both formal inference approaches that rely on asymptotic approximations and over informal univariate test procedures that had been used in other studies of portfolio efficiency.
Two assumptions that may affect the application of the efficiency test to the ECU are: ( 1 ) the normality of excess currency returns and (2) the requirement that the ECU not be linearly related to the N currencies. The previous section showed that for some currencies, returns are not normally distributed. Departures from normality may affect the test but MacKinlay (1985) shows that the F test is robust to this type of misspecification. The second problem is that the ECU, as discussed previously, is a weighted average of the component currencies and so is linearly related to them. However, because these weights change over the sample period, the assumption that the ECU is not linearly related to its component currencies is not violated. Table 4 displays the Ordinary Least Squares regressions of the excess.returns from individual ECU countries on the excess ECU return. All but two of the univariate t-tests have intercepts different from zero, suggesting that the ECU portfolio is inefficient.
In addition to the compelling statistical properties, the multivariate test also has a useful geometric interpretation. Before Gibbons et. al. the efficient frontier was taken as given and various measures were proposed to determine how far the efficiency of a given portfolio was from the (ex post) optimal portfolio. For example, one such test compares the difference in the slope of the rays from the origin through the optimal portfolio (given by the tangent to minimum-variance frontier of risky assets) and the portfolio under investigation (i.e. it Figure 2 displays the (ex post) minimumvariance frontier for the inaugural euro. This looks like the top part of the efficient frontier represented in Figure 1 but allows for shortselling; the bottom part is dominated as portfolios can be constructed with equal variance and higher excess returns. We use a multivariate F statistic to test whether the slope of the tangency line to the efficient frontier is different from the slope of the line through the ECU. The slope of the tangency line is 2.43 and the slope of the line through the ECU portfolio compares the Sharpe performance ratios, that is the ratio of a portfolio’s mean return to. its standard deviation). In contrast the Gibbons et. al. statistic is a test of the ex ante efficiency of a given portfolio. They show that the multivariate test statistic they propose is based on the squares of the slopes for the optimal portfolio and the portfolio under investigation. They show that their test statistic compares the distance along the ex post frontier tangency ray up to a given level of risk to the distance along the ray to the portfolio under investigation for the same level of risk. When the ratio of these two distances are not statistically different, the ex ante efficiency of the given portfolio cannot be rejected. is 0.34. These slopes are significantly different, the resulting F statistic equals 35.5. For an F with 10 degrees of freedom in the numerator and 71 degrees of freedom in the denominator, the critical value at the 1% level of significance is 2.55. The sample data rejects the null hypothesis that the euro is mean-variance efficient. The euro should not be held by in lieu of a combination of the euro-zone country currencies but in addition to them. CONCLUSION Europe’s great experiment to create an economic and monetary union has taken a significant step forward with the launch of the single currency, the euro, joined by 11 member countries. This paper investigated whether the newly created euro is an attractive investment vehicle for U.S. investors. In particular, using a mean-variance framework we examine the efficiency of the euro as a currency portfolio. Due to the weighting of the 11 member currencies, the euro offers the lowest volatility of all 15 EMS currencies, only with the exception of the UK pound sterling. However, investing in the euro, during the January 1992 to October 1998 period, was a low risk and low return alternative compared to investing in the US dollar. Using the mean-variance approach, the actual currency weights in the ECU (euro) are very different from the optimal composition of the currency cocktails composed of the 11 (and 15) European currencies. Results of the Gibbons et.al.(1989) multivariate test show that the composition of the European common currency is inefficient. The euro will definitely bring stability (i.e. elimination of currency risk) but offers a correspondingly low level of excess return above that of the U.S. dollar. Over time, the euro could become a rival for the US dollar as a major reserve and trading currency to the extent the ECB succeeds in pursuing price stability across Europe.
Gibbons, M., S. Ross and J. Shanken, “A Test of the Efficiency of a Given Portfolio,” Econometrica, September 1989, 1121-1152.
Jager, H. and E. de Jong, ” The Exchange-Rate Mechanism of the EMS and the ECU as a Reserve Asset,” European Economic Review, 31, July 1987, 1071-1091.
Johnson, R. and L. Soenen, “The European Perspective on the Composition and Variability of the ECU,” Journal of Business Finance & Accounting, 21, 7, October 1994, 963-974.
Jorion, P., “Properties of the ECU as a Currency Basket,” Journal of Multinational Financial Management, 1, 2, 1991, l-24.
MacKinlay, A., “An Analysis of Multivariate Financial Tests, ” Ph.D. Dissertation, Graduate School of Business, University of Chicago, Chicago, Il., 1985.
Mathur, I. and W. Ogden Jr., “Using Ex Ante Currency Cocktails in Managing Foreign Exchange Risk,” Journal of Multinational Financial Management, 1, 1, 1991, 23-34.
McFarland, J, R. Petit and S. Sung, “The Distribution of Foreign Exchange Price Changes: Trading Day effects and Risk Measurement,” Journal of Finance, 37, June 1982,693-715.
Robinson, C., “The Euro as a Catalyst for Achieving Strategic Business Objectives,” TMA Journal, September/October 1998, 1418.
Temperton, P., “The Euro, the Yen and the Dollar, chapter 13 in P. Temperton, ed., The Euro, John Wiley & Sons Ltd., Chichester (NY), 1997, 161-170.
Andrew T. Allen, Robert Johnson University of San Diego
California Polytechnic State University
Copyright College of Business Administration. University of Detroit Mercy Fall 2000
Provided by ProQuest Information and Learning Company. All rights Reserved