The Effect of the Euro on Country Versus Industry portfolio Diversification.

The effect of the Euro on country versus industry portfolio diversification Thomas J. Flavin* National University of Ireland, Maynooth Abstract We examine the relative benefits of industrial versus geographical diversification in the Euro zone before and after the introduction of the common currency. A priori, one may expect that increased stock market correlation would precipitate a move from geographical towards industrial diversification. We employ the empirical model of Heston and Rouwenhorst but show that adopting a panel data approach is a more efficient estimation method. We find evidence of a shift in factor importance; from country to industry. However, this is not exclusive to the Euro zone but is also present for non-EMU European countries. Therefore, fund managers should pursue industrial rather than geographical diversification strategies. JEL Classification: F36, G11, G15. Keywords: Portfolio diversification, industry and country effects, Euro. * Address for correspondence: Dept. of Economics, NUI Maynooth, Maynooth, Co. Kildare, Ireland. Tel: + 353 1 7083369, Fax: + 353 1 7083934, Email: [email protected] 1. Introduction A fundamental principle of financial theory, dating back to Markowitz (1952), is that portfolio diversification allows an investor to earn higher returns for each unit of risk and hence leads to greater portfolio performance. Grubel (1968) and Levy and Sarnat (1970) were among the first to show that diversification across international assets increased these benefits due to their relatively low correlation compared to those of domestic stocks. Many empirical papers find that these benefits are still present despite increasing integration across financial markets in both stock markets (Grauer and Hakansson, 1987; De Santis and Gerard, 1997) and bond markets (Levy and Lerman, 1988) and in the face of time-varying correlations (Ang and Bekaert, 2002). Many authors have posed the question whether or not equivalent benefits can be obtained from diversifying portfolios across industries rather than across national borders. The early literature provided overwhelming evidence that international diversification is better than industrial diversification. Grubel and Fadnar (1971) report that industries within a country are more highly correlated than industries across countries. However, Heston and Rouwenhorst (1994,1995) were among the first papers to rigorously address this issue. They focus on European markets - 12 in total - and assign each stock to one of seven industrial sectors. Their main finding was that the majority of diversification benefits stem from international rather than industrial diversification. They report that on average less than 4% of the variation in country indexes is attributable to their industrial composition. Griffin and Karolyi (1998) include developed non-European markets as well as some emerging markets, while allowing for “more finely partitioned industrial classifications” but find no greater importance for industry effects in portfolio selection. Rouwenhorst (1999) again focuses on European countries over the post Maastricht Treaty time period up to August 1998 and finds that the relative strengths of country effects is unaffected by time and increased economic integration. More recent studies have been less supportive of the view that country effects dominate industrial sectoral effects. Brooks and Catao (2000) estimate the impact of ‘new-economy’ versus ‘old-economy’ stocks in portfolio diversification and find that the introduction of ‘new-economy’ stocks finds an increased role for diversification across industrial sectors. Baca et al. (2000) also report an increased role for sectoral effects in determining asset returns and conclude that country effects have declined in importance. Their focus is on the seven 1 largest world stock markets so therefore you may expect relatively high levels of integration. Moreover, Cavaglia et al. (2000), using an extended sample of countries, agree with this finding and state that for the purposes of portfolio risk reduction, industrial factors are more important than country effects. Recent studies on emerging markets have identified a similar pattern with Wang et al. (2003) finding that industrial effects have been significantly more important than country effects in Asian markets since at least 1999. Until now, studies focussing on Europe have found that industrial composition plays a relatively minor role in determining country correlations and that low correlations are primarily due to country-specific sources of return variation. This paper focuses on the determination of cross-country correlation and hence on the optimal portfolio diversification strategy from the perspective of an investor from a Euro zone country. In particular, we assess the relative importance of countryand industry-specific shocks to the variability of stock returns. There are a number of legitimate reasons for a re-examination of this issue. Firstly, from a financial markets perspective, there has been sufficient change in the investment landscape to warrant further investigation. In the aftermath of the introduction of irrevocably fixed exchange rates between member countries on January 1, 1999, a typical investor who wants to hold a portfolio without foreign exchange risk has had their investment opportunity set altered significantly. The portfolio set has been expanded enormously as all Euro zone investors may diversify across international borders between participating states without worrying about currency fluctuations. Bodart and Reding (1999) found that exchange rate risk reduces market integration. Of course, the magnitude of the benefits of increasing the investment set will be dictated by the correlations between stocks in these countries. There are likely to be large risk-return benefits to be reaped if the previously observed low country correlations are maintained. However some of the textbook explanations for low correlation no longer apply to the Euro zone countries, such as differences in fiscal and monetary policies. All states have now transferred responsibility for monetary policy from domestic central banks to the European Central Bank (ECB), while the degree of fiscal autonomy among member countries has also been dramatically reduced. This policy co-ordination has led to a substantial narrowing of interest rates across the Euro zone countries. We should expect that increased 2 economic integration would reduce the asymmetry of responses to shocks to fundamental variables. Furthermore, few institutional or legal impediments remain. Consequently, one might expect that cross-country correlations would be mainly driven by differences in the industrial structure of domestic markets. Therefore, it is reasonable to expect that diversification across industries may be more important in this new era, especially since Carrieri et al. (2004) find that increased country level integration does not rule out industry-level segmentation. However, there are a number of factors that could work in the opposite direction. Goetzmann et al. (2002) find that episodes of integration are not only characterised by increased cross-country correlation but also by an expansion of the investment opportunity set. The latter effect may offer improved investment diversification possibilities. Another competing view comes from Francis et al. (2002) who show that above average levels of currency volatility leads to increased stock market correlation, so it is possible that the effect of eliminating exchange rate variability could result in lower correlation between markets. The adoption of the Euro provides as near to a natural experiment as you are likely to find in financial economics and allows us to assess the potential explanations of low cross-country correlations mentioned above. Secondly, from an econometric viewpoint, we also apply more efficient estimation techniques to the model than those usually employed. In particular, we form a panel data set and show that pooling the data and estimating a cross-section of time series regression leads to more precise estimation. This allows us to attach statistical as well as economic significance to our results and has important implications for fund managers in making their decision whether to pursue active geographical or industrial diversification. We find that there has been a shift in importance from country to industrial effects. In the early years of our sample, our results are consistent with the other literature focussing on European stock markets; country effects outweigh industrial effects. However this result is reversed following the introduction of the Euro. Therefore Euro zone investment strategies would be better off concentrating on industrial rather than geographical diversification. This is consistent with increased integration between Euro zone markets after the adoption of the single currency, which has been documented by Fratzscher (2001). However, we use a group of nonEMU European countries to show that this result is not just confined to the Euro zone. 3 The robustness of our results to the inclusion of these additional markets suggests that the decline in importance of country effects may be due to factors other than the introduction of the Euro. In particular, when taken with the other literature, increased country correlations appear to be a global phenomenon. The remainder of the paper is organised as follows. Section 2 reviews the literature on the sources of low crosscountry correlation and analyses whether these are likely to apply within the common currency area. Section 3 describes the data while section 4 outlines the model and discusses its estimation. Our results are presented in section 5, while conclusions are contained in section 6. 2. Sources of low cross-country correlation Given its importance in portfolio selection models, the sources of low crosscountry correlation of financial asset returns have generated a great deal of literature. A number of common themes have emerged. Firstly, a potential explanation of low correlation may be due to low levels of market integration. In segmented or partially segmented markets local factors may be more important than global factors. Without full integration, it is possible to observe pricing differences or different speeds of price adjustment. There is empirical evidence to show that stock market correlation is positively linked to levels of both economic and financial integration. Ferson and Harvey (1991) find a positive relationship between the degree of real and financial integration. Bekaert and Harvey (1995) show that market integration has a strong influence on the co-movement of emerging market returns with a global market factor. Furthermore, there is evidence of market integration increasing over time (De Santis and Gerard, 1997 and Hardouvelis et al., 1999). Reductions in transaction costs, institutional and legal impediments are generally credited with increasing integration among developed markets. Following the substantial political, economic and financial co-ordination within the Euro zone, stock market co-movements are unlikely to be low for lack of market integration. However, stock market integration may still be restricted by the home bias in equity portfolios displayed by many investors (see Lewis, 1999 for a review of this topic). One possible explanation of this phenomenon is that investors are better informed about domestic (or regional) market conditions or they are more optimistic 4 about the future performance of domestic markets (investor sentiment). Flavin et al. (2003) show that geographical variables, which may be a proxy for these psychological barriers, have significant explanatory power for determining the level of stock market correlation. Secondly, following Roll (1992), differences in the industrial composition of national indices have been put forward as an important determinant of cross-country correlation. However, more recent empirical evidence does not support this view. Heston and Rouwenhorst (1994,1995), Griffin and Karolyi (1998) and Flavin et al. (2003) all show that industrial composition explains little of stock market comovements. Thirdly, economic fundamentals and economic shocks may also play a role in determining stock market correlation. Campbell and Hamao (1992) show that economic fundamentals, such as interest rates and dividend yields, help to explain US and Japanese market co-movement. Conversely, Karolyi and Stulz (1996) find little evidence that macroeconomic announcements or shocks to exchange rates or interest rates influence U.S. and Japanese stock return correlations. Ammer and Mei (1996) find that equity risk premia rather than fundamental variables account for most comovements across national indices. Obviously, country-specific shocks will impact on domestic market returns and hence reduce co-movements with other markets, but also global shocks to which markets have different sensitivities may also result in low cross-country correlation. With the co-ordination of monetary variables within the Euro zone, the main focus of our paper is to examine the role of economic shocks, the final explanation outlined above. In particular we seek to assess the relative importance of countryand industry-specific shocks. The degree to which such shocks have differential crosscountry and cross-industry effects may help to identify the optimal diversification strategy available to a portfolio manager. 3. Data We use monthly total returns and market capitalisations on 1193 companies across the eleven original members of the ‘Euro zone’. Greece is omitted from the analysis, as it did not join the EMU on January 1999. A control group is created using similar data for the UK, Switzerland, Denmark and Sweden. All returns are expressed in a 5 common currency, the Euro. Pre-Euro returns for all markets and post-Euro returns for the non-EMU countries are computed by converting from the domestic currency to the Euro via the ECU end-of-month exchange rate. Our sample stretches from January 1995 to December 2002. The starting point was chosen to give an equal span before and after the introduction of the Euro. In this respect, we hope to capture changes in optimal diversification strategies that may have been induced by the adoption of the common currency. All data are collected from Datastream and each company is assigned to an industrial sector and a country according to the Datastream classification. These are consistent with the FTSE industry sectors. In this application, we use ten broad industrial classifications. Griffin and Karolyi (1998) have already shown that using very fine industry definitions does not significantly change the findings. Given that we are trying to assess the impact of the introduction of the common currency, we have decided to work with a balanced panel of companies. For the Euro zone, this leaves us with 740 companies. The industrial and geographical breakdown of these companies is reported in Table 1. It is clear that there is a nonuniform distribution of companies across industrial sectors and especially across geographical boundaries, e.g. Luxembourg has relatively few stocks and these tend to be concentrated in the financial sector, whereas Germany accounts for almost 20% of the companies in our sample but 75% of these operate outside of the financial sector. Table 2 presents information on the average market capitalisation of the firms in our sample. In particular, we report the average proportion of the Euro zone market that is attributable to each country and each sector over the whole sample. Again we see important differences across countries and industrial sectors. Information technology stocks accounted for about 7% of the Euro zone value-weighted index but almost 40% of these were located in Finland (mainly Nokia). These stocks represented over 60% of the Finnish market. The highest value weights for German stocks are in Financials and Industrial firms, while France has a higher concentration of Service providing companies. Tables 3 and 4 present sample correlations for the Euro zone countries and industries respectively. These correlations are computed using monthly returns for the preand post-Euro period. Initially, focusing on the country correlations, we can see that there is significant variation between the samples. The average pair-wise correlation falls from 0.679 in the pre-Euro sample to 0.586 in the post-Euro sample. 6 This is counter-intuitive given that we would have expected the correlation to rise in light of the increased integration within this economic zone. However, this finding is consistent with Adjaouté and Danthine (2004) who argue that this could be due to the cyclical nature of country correlation. We would require an extended post-Euro sample to verify this. Another potential explanation is that EMU has created even greater return dispersion through an expansion of the investment opportunity set as suggested by Goetzmann et al. (2002), thereby offsetting the integration effect with new diversification possibilities. Alternatively, it may be that the elimination of currency volatility has lowered equity market correlation, in line with Francis et al. (2002). The falling average correlation does mask the fact that over 38% (21 out of 55) of the correlations did increase. Table 4 contains the corresponding matrix for the Euro zone industries. Here we see a relatively large decrease in the average pair-wise correlation, from a pre-Euro level of 0.747 to 0.568 in the post-Euro sample. A decrease in correlation was recorded in 38 of the 45 (nearly 85%) correlation coefficients. It is noteworthy that the average correlation is higher for industries than countries in the former time period and slightly lower in the later period. This suggests that there may have been a relative shift or, at least, convergence in country and industry effects over the sample. Equivalent correlations for the non-EMU sample are both characterised by increasing correlation. Average correlations have increased from 0.668 to 0.704 for countries and 0.428 to 0.453 for industries. It is noticeable that all pair-wise crosscountry correlations are much higher than those for industries. These correlations are presented in Tables 5 and 6. 4. Methodology Heston and Rouwenhorst (1994) propose a model of stock return that is capable of disentangling country and industry effects. Solnik and de Freitas (1988) allow for exchange rate effects, but this is clearly redundant in our specification. The return for any stock i that belongs to industry j and country k is given by: . itkjit R        (1) 7 In this formulation, α represents a common component of all stocks, βj captures the industry effect and γk the country effect. The error term, εi, is asset specific and is assumed to be zero mean with a finite variance. This specification rules out any interaction between industry and country effects. Using our data, we have companies located in one of eleven countries (k = 1 to 11), with each belonging to one of ten industries (j = 1 to 10). We define industry dummies, Iij, to have a value of one if stock i belongs to industry j and zero otherwise. Likewise, country dummies, Cik, take a value of one if stock i belongs to country k and zero otherwise. Thus we can rewrite equation (1) for each time period as ...... 111111101011 itiiiiit CCIIR            (2) Of course, equation (2) cannot be estimated in its current form as both the industry and country dummies sum to unity, resulting in perfect multicollinearity between the regressors. We could proceed by dropping an arbitrary industry and country and measuring everything else relative to these. However for portfolio managers, it would be more desirable to measure country and industry effects relative to some more easily identifiable and accepted benchmark such as an equally (or value)-weighted index of stocks. Heston and Rouwenhorst (1994) follow Suits (1984) and Kennedy (1986) by estimating a constrained dummy variable regression. In essence, this amounts to constraining the weighted industry and country effects to sum to zero. Imposing such restrictions is equivalent to measuring each industry relative to the average firm or in this case a weighted portfolio of Euro zone stocks. If we apportion the weights simply as the number of stocks in each country and industry, then our benchmark is an equally weighted index.       11 1 10 1 0 ,0 kkk j jj m n   (3) where nj and mk represent the number of firms in industry j and country k respectively. 8 advocated as a benefit of EMU, it could be argued that this may account for some of the increase in importance of industry effects within the single currency zone. Brooks and Catao (2000) argue that the increasing importance of the industry factor in their study could be due to the Information Technology sector. It is generally accepted that there was a bubble in this sector during the late-90’s, being fuelled by internet companies in the main. However, in our analysis, we are dealing with a balanced panel and therefore only include those stocks for which a full history from 1995-2002 is available. Therefore the influence of short-lived, mis-priced companies is greatly reduced, if not totally eliminated. An alternative explanation for the increased importance of industry factors stems from the cyclical behaviour of country effects (Adjaouté and Danthine, 2004). The decline in country factors may be temporary and if so fund managers should be careful about the absolute adoption of industrial strategies. Presently such strategies seem to offer better diversification possibilities but in so far as country factors are cyclical, this could be reversed again. 6. Conclusion The goal of our paper is to assess the relative importance of country and industry effects in European portfolio diversification and the impact of the Euro on this. Many earlier studies have addressed this issue and generally, for European markets, concluded that country effects were greater and consequently diversification along geographical lines was more important for fund managers. Our motivation for undertaking this analysis is two-fold. Firstly, this is the first study to focus exclusively on the Euro zone markets in the post-EMU period. The elimination of foreign exchange risk lifted barriers for investors who are averse to this risk source and as such provided a much-expanded ‘domestic’ market. The adoption of a common monetary policy and the greater alignment of fiscal policy across member states, together with few legal or institutional barriers to investment served to reduce many of the usual explanations for low cross-country correlation. One remaining plausible explanation is that low stock market co-movement stems from the differing industrial composition of the indexes. Therefore, a priori, one might expect that within this region industrial effects may play a more important role in portfolio choice in the aftermath of the Euro being adopted. Secondly, we apply panel data estimation 15 techniques that improve the efficiency of our results. Compared to the more traditional estimation approach, we are able to attach statistical as well as economic significance to our results. Using data on the Euro zone markets from 1995-2002, our findings suggest that in the purely post-Euro sample industry effects outweigh country effects and hence industrial diversification is more likely to confer greater portfolio performance on the investor. On average, correlations between national stock markets in this area have decreased but by less than cross-industry correlations. It is also noteworthy that correlations vis-à-vis the larger markets actually increased. Now, industrial portfolios appear to be less correlated than country portfolios. Industry-specific shocks create more return dispersion than country-specific shocks and hence offer greater portfolio diversification benefits. However, further analysis reveals that this change in the relative importance of country and industry effects is not exclusive to the Euro zone. In fact, it is also to be found in a sample of non-EMU European countries whether analysed separately or in a larger pan-European sample. Therefore we conclude that this reversal in the relative fortunes of country and industry diversification is not due to the introduction of the common currency but is part of a global phenomenon that has also been documented for other regions. Consequently, portfolio managers would be well advised to adopt diversification strategies based on industry portfolios rather than country portfolios. A number of explanations for the reported increase in the importance of industry effects are suggested. Firstly, global market conditions since 1999 have been turbulent with a number of financial crises and the crash following September 11, 2001. Such events generally tend to increase stock market co-movements. Secondly, industry-specific shocks could become a more important source of return variation as some industries become more exposed to international trade. Thirdly, Adjaouté and Danthine (2004) suggest that the decline in country effects may be due to their cyclical behaviour. A longer post-Euro sample will be needed to verify this. 16 References Adjaouté, K., Danthine, J.P., 2004. Portfolio diversification: alive and well in Euroland. Applied Financial Economics, forthcoming. Ammer, J., Mei, J., 1996. Measuring international economic linkages with stock market data. Journal of Finance 51 (5), 1743-1763. Ang, A., Bekaert, G., 2002. International asset allocation with regime shifts. Review of Financial Studies 15 (4), 1137-1187. Bekaert, G., Harvey, C.R., 1995. Time-varying world market integration. Journal of Finance 50 (2), 403-444. Baca, S.P., Garbe, B.L., Weiss, R.A., 2000. The rise of sector effects in major equity markets. Financial Analysts Journal 56 (5), 34-40. Brooks, R., Catao, L., 2000. The new economy and global stock returns. IMF Working Paper no. 00/216. Washington, DC. Campbell, J.Y., Hamao, Y., 1992. Predictable stock return in the United States and Japan: a study of long-term capital market integration. Journal of Finance 47 (1), 4369. Carrieri, F., Errunza, V., Sarkissian, S., 2004. Industry risk and market integration. Management Science 50 (2), 207-221. Cavaglia, S., Brightman, C., Aked, M., 2000. The increasing importance of industry factors. Financial Analysts Journal 56 (5), 41-54. De Santis, G., Gerard, B., 1997. International asset pricing and portfolio diversification with time-varying risk. Journal of Finance 52 (5), 1881-1912. Ferson, W., Harvey, C.R., 1991. The variation of economic risk premiums. Journal of Political Economy 99 (2), 385-415. Flavin, T.J., Hurley, M.J., Rousseau, F., 2003. Explaining stock market correlation: a gravity model approach. The Manchester School 70, 87-106. Francis, B.B., Hasan, I., Hunter, D.M., 2002. Return-volatility linkages in international equity and currency markets. Bank of Finland Discussion Papers no. 9/2002. Helsinki. Fratzscher, M., 2001. Financial market integration in Europe: on the effects of EMU on stock markets. European Central Bank Working Paper no. 48. Frankfurt. Goetzmann, W., Li, L., Rouwenhorst, K.G., 2002. Long-term global market correlations. Working Paper no. 8612, National Bureau of Economic Research, Cambridge, MA. 17 Grauer, R.R., Hakansson, N.H., 1987. Gains from international diversification: 196885 returns on portfolios of stocks and bonds. Journal of Finance 42 (3), 721-739. Griffin, J.M., Karolyi, G.A., 1998. Another look at the role of the industrial structure of markets for international diversification strategies. Journal of Financial Economics 50, 351-373. Griffin, J.M., Stulz, R.M., 2001. International competition and exchange rate shocks: a cross-country industry analysis of stock returns. Review of Financial Studies 14, 215-241. Grubel, H., 1968. Internationally diversified portfolios: welfare gains and capital flows. American Economic Review 58, 1299-1314. Grubel, H., Fadnar, K., 1971. The interdependence of international equity markets. Journal of Finance 26 (1), 89-94. Hardouvelis, G., Malliaropulos, D., Priestley, R., 1999. EMU and European stock market integration. CEPR Discussion Paper no. 2124. London. Heston, S.L., Rouwenhorst, K.G., 1994. Does industrial structure explain the benefits of international diversification? Journal of Financial Economics 36, 3-27. Heston, S.L., Rouwenhorst, K.G., 1995. Industry and country effects in international stock returns. Journal of Portfolio Management, 53-58. Karolyi, G.A., Stulz, R.M., 1996. Why do markets move together? An investigation of U.S. – Japan stock return comovements. Journal of Finance 51 (3), 951-986. Kennedy, P., 1986. Interpreting dummy variables. Review of Economics and Statistics 68, 174-175. King, M.A., Wadhwani, S., 1990. Transmission of volatility between stock markets. Review of Financial Studies 3, 5-33. Levy, H., Lerman, Z., 1988. The benefits of international diversification in bonds. Financial Analysts Journal 44, 56-64. Levy, H., Sarnat, M., 1970. International diversification of investment portfolios. American Economic Review 60, 668-675. Lewis, K.K., 1999. Trying to explain home bias in equities and consumption. Journal of Economic Literature 37, 571-608. Markowitz, H.M., 1952. Portfolio Selection. Journal of Finance 7 (1), 77-91. Roll, R., 1992. Industrial structure and the comparative behavior of international stock market indices. Journal of Finance 47 (1), 3-41. 18 Rouwenhorst, K.G., 1999. European equity markets and the EMU. Financial Analysts Journal 48, 57-64. Schwert, G.W., 1989. Why does stock market volatility change over time? Journal of Finance 44 (5), 1115-1153. Solnik, B., De Freitas, A., 1988. International factors of stock price behavior. In: Khoury, S., Ghosh, A. (Eds.), Recent Developments in International Finance and Banking. Lexington MA. Suits, D.B., 1984. Dummy variables: mechanics v. interpretation. Review of Economics and Statistics 66, 177-180. Wang, C., Lee, C., Huang, B., 2003. An analysis of industry and country effects in global stock returns: evidence from Asian countries and the U.S. The Quarterly Review of Economics and Finance 43, 560-577. 19 Panel A TOTLF NCYSRCYSER NCYCGCYCGDGENIN BASIC ITECH RESOR UTILS TOTAL OE 10 0 2 5 0 5 7 0 1 2 32 BG 18 5 3 1 0 9 8 2 0 1 47 LX 13 0 2 2 0 0 1 0 0 3 21 FN 3 1 7 5 1 9 9 2 0 1 38 FR 26 6 36 20 17 12 13 9 6 0 145 BD 35 3 8 21 18 24 19 1 0 7 136 IR 6 0 6 9 1 1 8 0 6 0 37 IT 35 4 5 2 10 13 14 1 2 0 86 NL 22 5 18 10 8 11 9 2 7 0 92 ES 23 7 6 7 4 7 15 0 2 5 76 PT 6 4 7 3 1 3 6 0 0 0 30 197 35 100 85 60 94 109 17 24 19 740 Panel B UK 122 9 121 32 8 31 41 14 13 9 400 SW 33 2 13 12 6 18 9 2 0 6 101 DK 13 1 8 10 0 3 1 0 0 0 36 SK 11 0 3 4 2 11 9 3 0 0 43 179 12 145 58 16 63 60 19 13 15 580 Table 1: Panel A (B): Number of Companies in Balanced Euro zone (Non-EMU) Panel by Industrial Sector and Country. Key: OE = Austria, BG = Belgium, LX = Luxembourg, FN = Finland, FR = France, BD = Germany, IR = Ireland, IT = Italy, NL = Netherlands, ES = Spain, PT = Portugal, UK = United Kingdom, SW = Switzerland, DK = Denmark, SK = Sweden, TOTLF = Financials, NCYSR = Non-cyclical Services, CYSER = Cyclical Services, NCYCG = Non-cyclical Consumer goods, CYCGD = Cyclical Consumer goods, GENIN = General Industrials, BASIC = Basic Industrials, ITECH = Information Technology, RESOR = Resources, UTILS = Utilities. 20 TOTLF NCYSRCYSER NCYCGCYCGDGENIN BASIC ITECH RESOR UTILS TOTAL OE 0.0029 0.0002 0.0004 0.0005 0.0002 0.0007 0.0015 0.0000 0.0008 0.0009 0.0081 BG 0.0177 0.0019 0.0016 0.0031 0.0002 0.0025 0.0037 0.0002 0.0000 0.0066 0.0374 LX 0.0028 0.0000 0.0013 0.0000 0.0000 0.0000 0.0002 0.0000 0.0000 0.0001 0.0045 FN 0.0019 0.0035 0.0010 0.0017 0.0001 0.0019 0.0054 0.0263 0.0007 0.0001 0.0427 FR 0.0394 0.0331 0.0293 0.0409 0.0177 0.0311 0.0185 0.0218 0.0244 0.0013 0.2575 BD 0.0722 0.0295 0.0125 0.0139 0.0259 0.0447 0.0259 0.0167 0.0003 0.0066 0.2482 IR 0.0063 0.0000 0.0014 0.0042 0.0002 0.0000 0.0027 0.0000 0.0002 0.0000 0.0151 IT 0.0520 0.0312 0.0103 0.0016 0.0086 0.0053 0.0023 0.0006 0.0141 0.0093 0.1354 NL 0.0489 0.0126 0.0163 0.0188 0.0023 0.0111 0.0059 0.0050 0.0345 0.0000 0.1553 ES 0.0278 0.0181 0.0042 0.0022 0.0009 0.0017 0.0055 0.0000 0.0057 0.0149 0.0811 PT 0.0052 0.0046 0.0009 0.0001 0.0001 0.0001 0.0016 0.0000 0.0000 0.0021 0.0148 0.2773 0.1346 0.0792 0.0871 0.0562 0.0992 0.0733 0.0706 0.0805 0.0421 1.0000 Table 2: Percentage of Value-weighted index by Industry and Country. Key: OE = Austria, BG = Belgium, LX = Luxembourg, FN = Finland, FR = France, BD = Germany, IR = Ireland, IT = Italy, NL = Netherlands, ES = Spain, PT = Portugal, TOTLF = Financials, NCYSR = Non-cyclical Services, CYSER = Cyclical Services, NCYCG = Non-cyclical Consumer goods, CYCGD = Cyclical Consumer goods, GENIN = General Industrials, BASIC = Basic Industrials, ITECH = Information Technology, RESOR = Resources, UTILS = Utilities. 21 PRE-EURO OE BG FN FR BD IR IT LX NL PT ES OE 1.000 BG 0.738 1.000 FN 0.648 0.666 1.000 FR 0.760 0.828 0.562 1.000 BD 0.807 0.807 0.675 0.797 1.000 IR 0.674 0.642 0.711 0.627 0.734 1.000 IT 0.644 0.768 0.547 0.777 0.638 0.579 1.000 LX 0.628 0.621 0.457 0.564 0.541 0.550 0.604 1.000 NL 0.821 0.819 0.693 0.856 0.878 0.731 0.704 0.558 1.000 PT 0.759 0.625 0.508 0.776 0.628 0.573 0.661 0.465 0.686 1.000 ES 0.705 0.702 0.627 0.757 0.727 0.702 0.725 0.611 0.748 0.686 1.000 POST-EURO OE BG FN FR BD IR IT LX NL PT ES OE 1.000 BG 0.568 1.000 FN 0.021 0.122 1.000 FR 0.336 0.621 0.716 1.000 BD 0.438 0.631 0.611 0.946 1.000 IR 0.361 0.608 0.346 0.645 0.669 1.000 IT 0.267 0.498 0.593 0.862 0.834 0.526 1.000 LX 0.384 0.411 0.317 0.679 0.732 0.534 0.678 1.000 NL 0.497 0.735 0.560 0.905 0.899 0.723 0.832 0.679 1.000 PT 0.079 0.410 0.508 0.728 0.721 0.433 0.719 0.561 0.618 1.000 ES 0.375 0.576 0.499 0.832 0.830 0.669 0.767 0.605 0.797 0.735 1.000 Table 3. Correlation Matrix for Euro zone countries preand post-Euro. Key: OE = Austria, BG = Belgium, LX = Luxembourg, FN = Finland, FR = France, BD = Germany, IR = Ireland, IT = Italy, NL = Netherlands, ES = Spain, PT = Portugal 22 PRE-EURO TOTLF N CYS R C YSE R N CYCG C YCGD GENIN B ASIC I TECH R ESOR R ESOR TOTLF 1.000 N CYSR 0.784 1.000 CYSER 0.789 0.851 1.000 N CYC G 0.818 0.846 0.875 1.000 CYCGD 0.833 0.859 0.855 0.874 1.000 GENIN 0.896 0.867 0.887 0.917 0.930 1.000 BASIC 0.810 0.795 0.878 0.870 0.914 0.932 1.000 ITECH 0.805 0.756 0.743 0.692 0.802 0.825 0.773 1.000 RESOR 0.591 0.575 0.649 0.632 0.742 0.692 0.700 0.500 1.000 UTILS 0.534 0.663 0.606 0.564 0.580 0.584 0.557 0.423 0.573 POST-EURO TOTLF N CYS R C YSE R N CYCG C YCGD GENIN B ASIC I TECH R ESOR R ESOR TOTLF 1.000 N CYSR 0.517 1.000 CYSER 0.761 0.773 1.000 N CYC G 0.630 0.070 0.301 1.000 CYCGD 0.786 0.459 0.715 0.432 1.000 GENIN 0.850 0.698 0.917 0.389 0.795 1.000 BASIC 0.800 0.429 0.710 0.454 0.834 0.786 1.000 ITECH 0.707 0.832 0.817 0.272 0.614 0.841 0.606 1.000 Table 4. Correlation Matrix for Euro zone Industrial Sectors preand post-Euro. Key: TOTLF = Financials, NCYSR = Non-cyclical Services, CYSER = Cyclical Services, NCYCG = Non-cyclical Consumer goods, CYCGD = Cyclical Consumer goods, GENIN = General Industrials, BASIC = Basic Industrials, ITECH = Information Technology, RESOR = Resources, UTILS = Utilities. 23 1995-1998 UK SW DK SK UK 1 SW 0.639548 1 DK 0.728706 0.657548 1 SK 0.677344 0.690473 0.616662 1 1999-2002 UK SW DK SK UK 1 SW 0.790238 1 DK 0.685414 0.677749 1 SK 0.731875 0.632709 0.708063 1 Table 5. Correlation Matrix for non-EMU countries preand post-Euro. Key: UK = United Kingdom, SW = Switzerland, DK = Denmark, SK = Sweden. 24 End \ Start Jan 95 Jan 96 Jan 97 Jan 98 Jan 99 Jan 00 Jan 01 Jan 02 Dec 95 0.5720 Dec 96 0.5258 0.5202 Dec 97 0.5263 0.7019 0.6522 Dec 98 1.6243 1.2697 1.0905 1.5573 Dec 99 1.4045 1.4199 1.3717 1.3483 0.9530 Dec 00 0.8722 0.8773 0.7136 0.6711 0.6612 1.7161 Dec 01 0.6223 0.3946 0.4395 0.4170 0.6440 2.4733 1.2010 Dec 02 0.9368 0.6274 0.8083 1.0230 1.1898 1.8569 1.2757 1.0803 Table 11: Ratio of industry to country effects based on balanced panel of 1320 companies from Europe. 31 32 End \ Start Jan 95 Jan 96 Jan 97 Jan 98 Jan 99 Jan 00 Jan 01 Jan 02 Dec 95 0.9634 Dec 96 0.8009 0.6040 Dec 97 0.5914 0.7677 0.6771 Dec 98 1.5275 1.3004 0.9431 1.3616 Dec 99 1.8263 1.3265 1.1920 1.3552 0.9742 Dec 00 0.9459 0.6075 0.4563 0.5112 0.5422 1.9693 Dec 01 0.7637 0.3163 0.2858 0.4332 0.6872 1.9958 0.7769 Dec 02 0.7212 0.3911 0.5626 0.7751 1.0943 1.7869 1.0350 0.9891 Table 12: Ratio of industry to country effects based on balanced panel of 543 companies from EMU excluding financials.