Applied Financial Economics, 2005, 15, 1299–1304
International indexing as a means of portfolio diversification Hakan Saritasa,* and Hakan Aygorenb a
Pamukkale University, I.I.B.F. Kinikli Kampusu, Kinikli 20040, Denizli, Turkey b Department of Business Administration, The Faculty of Economics and Administrative Sciences, Pamukkale University, Turkey
Global investing offers investors a larger pool of investment opportunities and tremendous diversification. However, despite the increased integration of world economies, there are still important variations among overseas capital markets. Complexities of overseas investing can often ensnare even the best active asset managers. International indexing is an option to overcome the difficulties of global investing. This study considers international indexing as a means of portfolio diversification. Performances of 15 international indexes are evaluated using monthly return data from 1998 through 2002. Returns are measured against ISE-100 Index (Istanbul Stock Exchange-100 Index) returns. The results of the study suggest that international indexing does not offer superior returns compared to the ISE-100 index.
I. Introduction International investment is growing rapidly among private and institutional investors, fuelled by the deregulation of national markets and the relaxation of capital controls. The trend towards international investments reflects the growing recognition that international diversification can yield a superior risk–reward trade-off compared to domestic investments. The parallel and increasing use of indexed portfolios reflects the perceived advantages of passive versus active management. Recently these two trends have converged in indexed international investments. Advocates of passive international equity portfolios argue that transaction costs are significantly higher in international markets (Jorion and Roisenberg, 1993). A superior rate of return is one key benefit of international investing. A global focus also offers a larger pool of investment opportunities. However, while the profit potential and diversification opportunities of global investing are huge, such pitfalls
as administrative barriers, currency fluctuations, and higher costs do exist in international investments. One option to overcome the difficulties of global investing is international indexing. International indexing is the effective strategy to hedge against these potential international investment pitfalls, and lowers the costs associated with international investing. The variety of indexes and index funds ensures investors have ample opportunity to customize an indexing strategy that will more than fulfil their financial goals with minimal risk.
II. Passive Investment Strategy Passive investment strategy involves a minimal amount of oversight and very few transactions once the portfolio has been selected (Maness and Zietlow, 1993). Proponents of the efficient market hypothesis believe that active management is largely wasted
*Corresponding author. E-mail:
[email protected] Applied Financial Economics ISSN 0960–3107 print/ISSN 1466–4305 online ß 2005 Taylor & Francis http://www.tandf.co.uk/journals DOI: 10.1080/09603100500187844
1299
H. Saritas and H. Aygoren
1300 effort and unlikely to justify the expenses incurred. Therefore, they advocate a passive investment strategy that makes no attempt to outsmart the market. A passive strategy aims only at establishing a welldiversified portfolio of securities without attempting to find under- or overvalued stocks. Because the efficient market theory indicates that stock prices are at fair levels, given all available information, it makes no sense to buy and sell securities frequently, which generates large amount of brokerage fees without increasing expected performance (Bodie et al., 1993). Investors who believe in efficient market theory often prefer indexing strategy. The strategy is based on the assumption that it is almost impossible to get a return over indexes in the long run and that indexes fully reflect whole market (Saritas, 2001).
III. International Investing and Indexing International investing provides investors tremendous diversification. A superior rate of return is one key benefit of international investing. A global focus also offers a larger pool of investment opportunities. Fouse (1992) suggests that price formation and the dynamics of security price changes are the same in all markets. You don’t have to learn a new language of finance to invest abroad. This should be reassuring to investors, given that the low correlation among the world’s equity markets provides the opportunity to enhance risk-adjusted expected return. The clear value of moving assets to international equities is diversification. Plans can reduce their volatility by investing in international securities. Aiello and Chieffe (1999) suggest that international diversification can reduce investment risk to about 56% of the level that can be achieved with national diversification. While the profit potential and diversification opportunities of global investing are huge, the fact that some obstacles as administrative barriers, currency fluctuations, and higher costs exist in international investments has made international indexing an option to overcome the difficulties associated with international investment. The logical way to achieve the international diversification is with an index fund. During the 1980s when returns averaged over 17% annually, it was difficult to beat the market, making indexing popular (Elgin, 1994). Indexing is the structuring of a passively managed portfolio of stocks or bonds that seeks to replicate the
returns of market indexes like the S&P 500 stock index or the Shearson/Lehman government bond index (Miller and Meckel, 1999). Indexing can take place in two principal forms. First, it can be accomplished through the physical replication of securities in an index, such as owning the exact number of shares of Microsoft, in order to match Microsoft’s percentage of the S&P 500. This can be done in the form of exact matching, or in simpler close approximations such as ‘stratified sampling’. Second, indexing can be accomplished through the use of derivatives contracts that seek to replicate the returns, not the holdings, of an index. The S&P 500 futures contract is by far the most recognizable equity-based derivatives contract, and is the common tool used in indexing through derivatives. Index funds are relatively liquid and do not significantly alter the market when their securities are bought and sold. Indexing provides impressive returns – the EAFE Index consistently ranks above median on an annual basis – at a risk equal to that of the market (Peters, 1988). If political-risk analysis is one way to minimize risk in emerging markets, passive indexing is another. Global indexing makes it possible for those who believe in passive investing to ease the problem of political risk by spreading it broadly (Hirsh, 1993). Index managers say international indexing has experienced tremendous growth mainly because it is a way to reduce risk and is less expensive than traditional international management (Price, 1989). International indexing has become popular among institutional investors for two reasons: Pension funds are increasing in general their long-term policy allocation to international asset classes. In addition, the performance of active managers outside the United States has been very disappointing (Durgin and Williams, 1990).
IV. Current Study Aiello and Chieffe (1999) found that investment in international mutual funds that are based on international indexes might offer significant diversification benefits. Authors also found that the performance of the average international index does not outperform the S&P 500 benchmark. In this study, we tried to find out if investors can benefit from investing in international index funds based on international indexes. We find similar results with Aiello and Chieffe (1999) in that international investing provides great diversification potential and international indexes cannot outperform the benchmark index.
International indexing as a means of portfolio diversification V. Research Design To find out that international index funds provide domestic individuals a good opportunity to invest in international securities, we have performed tests to determine whether international indexes outperform domestic indexes. We also tried to determine if there are diversification benefits to investing in international index funds. We have tested the monthly return data of 15 international indexes to determine whether they outperform the ISE-100 benchmark, using Sharpe, Treynor and Jensen measures. The Sharpe measure is the ratio of average risk premium to the total risk of the portfolio during the evaluation period and expressed as follows: ari arf Si ¼ i
ð1Þ
where Si is the Sharpe measure for index i; ari is the average return on index i; arf is the average risk-free rate; I is the standard deviation of returns on the index. The Treynor measure is the ratio of average risk premium to systematic risk of the portfolio and expressed as follows: Ti ¼
ari arf i
ð2Þ
where Ti is the Treynor measure for index i; I is the Beta or systematic risk for index i. The Jensen measure is the intercept in regression of excess portfolio returns against excess market returns and expressed as follows: i ¼ ðari arf Þ i ðarM arf Þ ð3Þ where i is the Jensen measure; arM is the average return on the market. The Jensen measure compares portfolio performance against market. In order to compare performances of different portfolios with each other, the modified Jensen measure should be used: i ð4Þ Ji ¼ i We then examined the correlation of the international indexes with the ISE-100 Index to find out if there are diversification benefits to investing in international index funds. We attempt to answer the following questions: (1) Does a linear relationship exist between the individual international index returns and the ISE-100 returns? (2) Does a linear relationship exist between
1301
the returns on the entire portfolio of international indexes and the returns on the ISE-100? In this study, we used monthly return data in USD bases on 15 international indexes that are formed, designed and updated by MSCI. The sample period is January 1998 through December 2002. We chose the indexes on such basis that different regions of the world are covered. The ISE-100 is the benchmark Turkey Index. The three-month Turkish Treasury bill yield is used as the risk-free rate. We compute excess returns of the indexes in USD bases from this security.
VI. Analysis of Data and Results Table 1 shows the risk-return data for the 15 international indexes and the ISE-100. During the evaluation period, the average monthly return on international indexes is 0.099%. The average monthly return on the ISE-100 is 0.401%. All of the 15 international indexes return less than the ISE-100. The EM Europe & Middle East Index has the largest monthly returns at 0.206% and the Pacific Index has the smallest at 0.328%. The lowest standard deviation is that of the EAFE Index at 0.0491. The EM Europe & Middle East Index has the greatest standard deviation at 0.0886. The standard deviation of the ISE-100 is 0.2061. The overall average standard deviation for international indexes is 0.0587. The benchmark index ISE-100 seems to be too risky compared to international indexes. Table 2 exhibits the portfolio performance measures for the indexes. According to the Sharpe measure, all of the indexes underperformed the ISE-100 Index. Indeed, Sharpe measures for all the indexes including the ISE-100 are negative. The average Sharpe measure for international indexes (0.18507) is less than that of the Sharpe measure for ISE-100 (0.02581). Therefore, we cannot reject the null hypothesis that the average performance of international indexes is equal to or less than that of the ISE-100 Index in terms of the Sharpe measure. According to the Treynor measure, none of the international indexes outperformed the ISE-100 Index. The average index has a 0.07598 Treynor measure, lower than the ISE-100 Treynor measure of 0.00532. Therefore, we cannot reject the null hypothesis that the average performance of international indexes is equal to or less than that of the ISE-100 Index in terms of the Treynor measure. According to the Jensen measure, also none of the international indexes outperformed the ISE-100 Index. Therefore, we cannot reject the null hypothesis
H. Saritas and H. Aygoren
1302 Table 1.
Returns for international indexes and ISE-100
Index Developed markets International indexes The world index North America EAFE Europe Pacific International free indexes World ex USA USA Emerging markets International indexes EMF (Emerging Markets Free) EM Europe & Middle East Combined markets All country indexes AC Europe EU All country free indexes AC world index free ex USA EAFE þ EMF AC Asia Pacific free ex Japan AC Americas free Average international ISE-100 (IMKB-100)
Table 2.
Average monthly return
Standard deviation
0.126% 0.024% 0.260% 0.179% 0.328%
0.0504 0.0555 0.0491 0.0530 0.0586
0.236% 0.013%
0.0495 0.0553
0.017% 0.206%
0.0813 0.0886
0.172% 0.168%
0.0534 0.0544
0.227% 0.247% 0.184% 0.012% 0.099% 0.401%
0.0509 0.0505 0.0740 0.0559 0.0587 0.2061
Performance measures for international indexes and ISE-100
Index
Sharpe
Rank
Treynor
Rank
Jensen
The world index North America EAFE Europe Pacific World ex USA USA EMF (Emerging Markets Free) EM Europe & Middle East AC Europe EU AC world index free ex USA EAFE þ EMF AC Asia Pacific free ex Japan AC Americas free ISE-100 (IMKB-100) Average international
0.20997 0.16373 0.24305 0.20982 0.21532 0.23627 0.16626 0.11258 0.0821 0.20686 0.20246 0.2281 0.2337 0.10124 0.16465 0.02581 0.18507
11 5 16 10 12 15 7 4 2 9 8 13 14 3 6 1
0.08162 0.06846 0.09613 0.08304 0.12617 0.0931 0.06985 0.04278 0.02181 0.0799 0.07772 0.08671 0.08888 0.05579 0.06774 0.00532 0.07598
10 6 15 11 16 14 7 3 2 9 8 12 13 4 5 1
0.0763 0.06314 0.09081 0.07772 0.12085 0.08778 0.06454 0.03746 0.01649 0.07459 0.0724 0.0814 0.08356 0.05047 0.06242 0 0.07066
that the average performance of international indexes is equal to or less than that of the ISE-100 Index in terms of the Jensen measure. The fact that all the indexes including the ISE-100 posted negative results in terms of the three measures most likely resulted from the fact that real interest rates are very high in Turkey.
Rank 10 6 15 11 16 14 7 3 2 9 8 12 13 4 5 1
Table 3 shows the results of the regression of the international indexes on the ISE-100 Index: EISE-100 ¼ a þ EACAF þ EEMEM þ EACAPFXJP þ EPAC þ EEUR þ EEMF þ EEU þ EUSA þ EACWXUS þ ETWI þ EEAFE þ "
ð5Þ
International indexing as a means of portfolio diversification Table 3.
1303
Regression of the international indexes on ISE-100
Index
Coefficients
t
Sig.
(Constant) The world index EAFE Europe Pacific USA EMF (Emerging Markets Free) EM Europe & Middle East Eu AC world index free ex USA AC Asia Pacific free ex Japan AC Americas free
0.006521357 5.510550353 34.89417215 25.5804684 5.98121148 10.459236 1.400667225 2.087686246 9.792294325 18.3012904 0.53 8.010240211
0.33915815 0.24087526 1.4286715 1.9458282 1.1692171 0.6859494 0.86835198 4.51308055 2 1.0446225 0.677613 0.55087502
0.73597 0.8106777 0.1595741 (*)0.0575434 0.2480898 0.496045 0.3895227 (*)4.138E-05 (*)0.0673721 0.3014299 0.5012704 0.5842744
Excluded variables Collinearity statistics Index
Beta In
Sig.
t
Partial correlation
Tolerance
North America 12.8366789 0.865591 0.3911115 0.1252649 3.83321E-05 World ex USA 8.60166432 0.5746754 0.5682516 0.0835321 3.79621E-05 AC Europe 3.903268059 0.41268564 0.6817135 0.06008762 9.53943E-05 EAFE þ EMF 2.480571113 0.14457248 0.8856666 0.02108338 2.90794E-05 Predictors in the model: (constant). ACAF. EMEM. ACAPFXJP. PAC. EUR. EMF. EU. USA. ACWXUS. TWI. EAFE Dependent variable: ISE-100 Note: (*) statistically significant at the 95% level.
Table 4.
Individual regressions of international indexes on ISE-100
Index
Constant
Coefficient
t
Sig.
R2
The world index North America EAFE Europe Pacific World ex USA USA EMF (Emerging Markets Free) EM Europe & Middle East AC Europe EU AC world index free ex USA EAFE þ EMF AC Asia Pacific free ex Japan AC Americas free
0.003 0.001 0.004 0.003 0.0004 0.0035 0.0014 0.0028 0.0051 0.0036 0.0037 0.0040 0.0041 0.0014 0.0017
0.786 0.744 0.741 0.779 0.457 0.751 0.736 0.882 1.432 0.810 0.821 0.807 0.801 0.520 0.764
2.289 2.220 2.196 2.337 1.545 2.220 2.201 3.366 6.791 2.441 2.504 2.396 2.381 1.784 2.293
(*)0.026 (*)0.030 (*)0.032 (*)0.023 0.128 (*)0.030 (*)0.032 (*)0.001 (*)0.000 (*)0.018 (*)0.015 (*)0.020 (*)0.021 (*)0.080 (*)0.025
0.083 0.078 0.077 0.086 0.040 0.078 0.077 0.163 0.443 0.093 0.098 0.090 0.089 0.052 0.083
Note: (*) statistically significant at the 99% level.
where E is the excess returns (monthly index return monthly T-Bill return). The variation in the Europe Index, the EM Europe & Middle East Index and the EU Index significantly explains some variation in the ISE-100 Index. While the Europe Index is negatively correlated to the ISE-100 Index, the EM Europe & Middle East Index and the EU Index are positively correlated to the ISE-100 Index.
Since collinearity tolerance statistics are low, four of the indexes (North America, World ex USA, AC Europe, EAFE þ EMF) are excluded from the model by the SPSS program. Table 4 exhibits the individual regression results of the 15 international indexes on the ISE-100 Index: EISE-100 ¼ þ E1 þ " where I is one of the 15 international indexes.
ð6Þ
H. Saritas and H. Aygoren
1304 According to the individual regression results, while the EM Europe & Middle East Index has a moderate degree of correlation, the other indexes explain only a small portion of the variation in the ISE-100 Index. Looking at the R2 statistics of the international indexes, the variation in the 15 international indexes explains between 4 and 44.3% of the variation in the ISE-100. That is to say that the international indexes have little correlation with the Turkish market. This result shows that international investing provides investors with substantial diversification potential. With investing in a fund based on an index that has little correlation with the domestic Turkish market, investors can take advantage of the existing diversification potential.
VII. Conclusion The results of this study suggest that investing in international index funds that are based on international indexes may provide investors with significant diversification opportunity. That is because international indexes have little correlation to the Turkish market. However, the results show that investors may not receive a return from international index funds over the benchmark index of ISE-100. This is because the performance of the average international index underperforms the ISE-100 benchmark index. In terms of performance, the results suggest that investors may be disappointed with risk-adjusted return by investing in the international indexes, including the ISE-100 Index. The negative results posted by the indexes including the ISE-100 in terms of Sharpe, Treynor and Jensen measures are most
likely a result for the fact that real interest rates are very high in Turkey. In sum, this study suggests that investors can gain significant diversification benefits from international indexing. However, investors may be disappointed with returns received.
References Aiello, S. and Chieffe, N. (1999) International index funds and the investment portfolio, Financial Services Review, 8, 27–35. Bodie, Z., Kane, A. and Marcus, A. J. (1993) Investments, R. D. Irwin Inc., IL. Durgin, H. and Williams, T. (1990) Indexing surges on foreign input, Pensions & Investments, 18, 1–2. Elgin, P. R. (1994) Sundae school of investing: vanilla index basis for active managers’ fancy flavours, Corporate Cashflow Magazine, 15, 38–9. Fouse, W. L. (1992) Allocating assets across country markets, Journal of Portfolio Management, 18, 20–7. Hirsh, M. (1993) Have political-risk analysts got Asia’s number?, Institutional Investor, 27, 65–7. Jorion, P. and Roisenberg, L. (1993) Synthetic international diversification, The Journal of Portfolio Management, 19, 65–74. Maness, T. S. and Zietlow, J. T. (1993) Short-Term Financial Management, West Publishing Company, St. Paul, MN. Miller, T. and Meckel, T. S. (1999) Beating index funds with derivatives, Journal of Portfolio Management, May, 75–88. Peters, E. E. (1988) The case for international index funds, Pension World, April, 26–8. Price, M. (1989) The world of indexes is getting bigger, Pensions & Investment Age, 17, 3–4. Saritas, H. (2001) Mutual fund performance evaluation: index funds as an alternative investment tool, PhD dissertation, Dokuz Eylul University, Izmir.