Country-Specific ETFs: An Efficient Approach to Global Asset

EDHEC RISK AND ASSET
MANAGEMENT RESEARCH CENTRE
393-400 promenade des Anglais
06202 Nice Cedex 3
Tel.: +33 (0)4 93 18 32 53
E-mail: [email protected]
Web: www.edhec-risk.com
Country-Specific ETFs: An Efficient
Approach to Global Asset Allocation
May 2006
Joëlle Miffre
Professor of Finance, EDHEC Business School
Abstract
The article shows that country-specific exchange-traded funds (ETFs) enhance global asset
allocation strategies. Because ETFs can be sold short even on a downtick, global strategies that
diversify risk across country-specific ETFs generate efficiency gains that cannot be achieved by
simply investing in a global index open- or closed-end fund. Besides, the benefits of international
diversification can be achieved with country-specific ETFs at low cost, with low tracking error
and in a tax-efficient way. For all these reasons, country-specific ETFs may be considered serious
competitors to traditional country open and closed-end funds.
Keywords: Country-Specific ETFs, Global Asset Allocation, Short-Selling
EDHEC is one of the top five business schools in France. Its reputation is built on the high quality of
its faculty (110 professors and researchers from France and abroad) and the privileged relationship
with professionals that the school has cultivated since its establishment in 1906. EDHEC Business
School has decided to draw on its extensive knowledge of the professional environment and has
therefore focused its research on themes that satisfy the needs of professionals.
2
EDHEC pursues an active research policy in the field of finance. The EDHEC Risk and Asset
Management Research Centre carries out numerous research programmes in the areas of asset
allocation and risk management in both the traditional and alternative investment universes.
Copyright © 2009 EDHEC
1. Introduction
Exchange-traded funds (ETFs) are shares that closely track the performance of an index.
They offer, in one trade, the benefits of diversification and index tracking1 at low cost.
The first ETF, the SPDR, was launched in 1993 and was designed to passively mimic the S&P500
index. Since then the variety of indices that ETFs track has kept expanding, ranging from equity
and bond indices worldwide to style and sector portfolios. As an investment vehicle, ETFs have
become so popular that Fuhr (2001) estimated that “most days, two or three ETFs are on the list
of the top five most actively traded stocks in the AMEX”. ETF assets under management grew at
a 132% average annual rate from 1995 to 2000, with a combined value in the US that stands at
$240 billion today (Investment Company Institute, 2005).
ETFs are effective tools for investment, risk and tax management. They are frequently used for
passive investment, short-term speculative trading, tactical asset allocation, hedging and arbitrage.
Alternatively, asset managers use them to carry out market-neutral strategies, to time the market,
implement sector and style rotation strategies or diversify a domestic portfolio across countries
and global industries. In a way similar to futures, they can also be used to simulate full investment
of small mandates into stocks. At times, the most liquid ETFs are even substitutes for the cash
necessary to meet redemption. In the end, the many advantages of ETFs lead Gastineau (2001) to
the conclusion that “in the year ahead… we would expect nearly all equity index funds to have
an ETF share class”. Developments in this field have already occurred. For example, Vanguard, an
established leader in the passive fund management industry, announced in January 2004 that it
was expanding its own class of ETFs to include US style and sector MSCI benchmarks.
This article focuses on 16 iShares MSCI country funds (Barclays Global Investors’ country-specific
ETFs). Like country open- and closed-end index funds, country-specific iShares increase meanvariance efficiency. However, unlike country index funds, country-specific ETFs can be sold short
even on a downtick. This paper shows that the ability to short ETFs even on a downtick enables
US investors to increase the mean-variance efficiency of their portfolio; namely, achieve higher
average returns for a given level of risk or, alternatively, lower risk for a given average return.
These efficiency gains are not available to investors in global open- or closed-end index funds,
as the latter cannot be sold short. Relative to their country open- and closed-end fund siblings,
country-specific iShares offer the additional advantages of being tax-efficient, of having low
tracking error and low costs. All these reasons explain why they have become an integral part of
an asset manager’s global asset allocation strategy and serious competitors to the traditional open
and closed-end country funds.
The remainder of the paper is organized as follows. The paper first presents the advantages of
country-specific iShares relative to conventional country funds. I then introduce the methodology,
the dataset and the empirical results, highlighting the efficient gains that can only be achieved
with country-specific ETFs.
2. Country-Specific ETFs Versus Open and Closed-End Country Funds
Like conventional country index funds, iShares MSCI country funds present in one trade the benefits
of international diversification and index tracking at a low cost. However country-specific ETFs
offer investors four additional benefits that cannot be obtained by merely trading open or closedend country funds.
First, unlike open-end funds, which trade only at 4:00pm NAV (net asset value), iShares trade
on a continuous basis and are exempt from the uptick rule. They can therefore be sold short at
any time during the trading day. These two features, continuous trading and short-selling, make
them particularly attractive in down markets, as investors then very much value the possibility
1 - Academic research suggests that it is nearly impossible to consistently beat the market (Jensen, 1968; Malkiel, 1995; Gruber, 1996; Carhart, 1997; Frino and Gallagher, 2001).
3
of either liquidating their iShares position or taking a short position with immediate effect.
Like country-specific ETFs, closed-end country funds trade continuously throughout the day.
However they can be sold short only on an uptick and this, as we will see, leads to efficiency losses
relative to ETFs.
Second, iShares MSCI country funds (and ETFs in general) are typically acknowledged as a taxefficient alternative to investing in mutual funds. They are created through a process called
“creation in kind”, whereby exchange specialists, market makers and arbitrageurs buy ETF shares
by depositing with the fund the index constituents. Similarly, ETFs can be liquidated through a
process called “redemption in kind”, whereby large market participants receive the portfolio of
underlying shares, rather than cash, in exchange for ETF shares. The ETF typically redeems in kind
the stocks with high unrealized capital gains and keeps the stocks with low unrealized capital
gains. Because of this, it is unlikely that the ETF will have to sell high capital gain stocks in the
future and, thus it is unlikely that its investors will be forced to pay capital gain taxes (Gastineau
(2001, 2002); Poterba and Shoven (2002)). In contrast, to raise cash for redemptions, traditional
mutual funds sell shares that may have appreciated relative to their initial cost. As a result,
redeeming clients realize taxable capital gains.
Third, country-specific ETFs are also well-known for their low tracking error relative to conventional
closed-end funds. Because they can create new shares or redeem existing ones at NAV on a daily
basis, arbitrageurs ensure close tracking and transparency with regards to the fund constituents.
Harper, Madura, and Schnusenberg (2003) confirm this by showing that country-specific ETFs
have returns that are not statistically different from the underlying MSCI returns and offer higher
Sharpe ratios than competing closed-end funds. Chang and Swales (2003) also show that because
of creation in kind, redemption in kind, and their open-end structure, country-specific ETFs trade
at lower discounts to NAV than country closed-end funds. This ultimately reduces the risk that
investors sell their ETFs shares at too low a price relative to NAV. On the other hand, it increases
the probability that investors buy ETFs shares at a more expensive price than otherwise similar
closed-end funds.
Finally, country-specific ETFs are cheaper than closed-end country funds. Chang and Swales
(2003) report that the average expense ratio on iShares MSCI country funds is only 0.87%, while
the average expense ratio on country closed-end funds is 1.59%. None of the closed-end funds
studied had expense ratios cheaper than their ETF counterparts. Looking at the universe of ETFs
and not just at iShares, we find that expense ratios can be as low as 0.09% a year for the most
liquid ones. This compares favourably to the fact that a typical index fund charges up to 0.5% a
year, while the expense ratio of an active fund can be as high as 2% (Poterba and Shoven (2002),
Kostovetsky (2003)).
To summarize, iShares MSCI country funds, unlike their open and closed-end country funds
siblings, can be sold short even on a downtick, are tax-efficient, have low tracking error and
low cost. This paper also argues that ETFs offer efficiency gains in as much as they increase
average returns for a given level of risk or, alternatively, decrease portfolio risk for a given return.
These incremental risk-adjusted return are not available to investors in global closed and
open-end index funds, as the latter cannot be sold short. This does not mean that we should
all use them as our preferred investment vehicle. Small investors, who cannot afford to
create country-specific ETFs in kind and are used to investing small amounts periodically,
typically find the brokers commissions and bid and ask spreads of ETFs too expensive.
They would rather invest in no-load country open-end index funds. Because they carry no
commissions, mutual funds are cheaper and more flexible to small investors, who very much value
the possibility of defining a transaction size that perfectly matches their resources.
4
3. Methodology
The methodology aims at constructing portfolios with long positions in US stocks and long or
short positions in iShares MSCI country funds. Following Sharpe (1964), I calculate the portfolio’s
weights ω that maximize the portfolio’s expected Sharpe ratio. The optimisation problem is
subject to two constraints. The first constraint ensures that the sum of the weights equals one
(It is needed to guarantee that the mandate is fully invested). The second constraint forbids or
allows short-selling up to a certain level K. In mathematical terms, the optimization problem reads
as follows
⎡ E (RP ) − Rf ⎤
⎡ ω' R − Rf ⎤
max ⎢
= max ⎢
⎥
⎥
ω
ω
σP
⎣ ω' V ω ⎦
⎣
⎦
N
subject to ∑ i =1 ω i =1 and ωi ≥ K for i =1,...,N . K is positive for iShares MSCI country funds if
short-selling of ETFs is not allowed or is negative for iShares MSCI country funds if short-selling
of ETFs is allowed. ω is a N-vector of portfolio weights, R is a N-vector of mean returns, σP is
the portfolio standard deviation and V is a covariance matrix. The resulting portfolio is called the
optimal risky portfolio.
The optimization problem is first solved for a portfolio containing the S&P500 and long positions
in iShares. Then, asset managers can combine risk-free borrowing and lending with the optimal
risky portfolio and define the asset allocation that matches their client’s preferences for risk and
expected return. Unlike traditional asset classes, ETFs can be sold short. For this reason, I relax, in
a second step, the assumption of no ETF short-sale and define the characteristics of the optimal
portfolio when ω ETF ≥ −0.05 , ω ETF ≥ −0.20 and ω ETF ≥ −1 . The choice of a 20% short interest was
dictated by the fact that short ETF positions typically range from 10 to 40% of the ETF market
capitalisation (Gastineau (2002)). To test the sensitivity of the results to the short-sale restriction,
this paper looks at the case of i) an investor who limits his short ETF interest to 5% and ii) a hedge
fund manager who considers up to 100% short positions in any ETF.
4. Data
The data include end-of-month total returns on the S&P500 index over the period April 1996 to
January 2004. The purpose of the analysis is to test how sensitive the efficient frontier of an USbased investment is to the inclusion of 16 iShares MSCI country funds. These include four iShares
from Asian-Pacific countries (Australia, Hong Kong, Malaysia and Singapore), ten iShares from
European countries (Austria, Belgium, France, Germany, Italy, the Netherlands, Spain, Sweden,
Switzerland and the UK) and two iShares from North-American countries (Canada and Mexico).
The US one-month Treasury-bill is used as a proxy for the risk-free asset.
Since means, standard deviations and correlations change over time, I calculate summary
statistics for the assets over increasing samples ranging from April 1996-April 1999 to the whole
dataset. It is indeed commonly acknowledged that in periods of recession, mean returns fall, while
standard deviations and correlations rise (Solnik, Boucrelle, and Le Fur (1996), Longin and Solnik
(2001), Dahlquist and Harvey (2001)), inducing changes in optimal asset allocations.
Table 1 presents summary statistics (annualized mean, standard deviation and expected Sharpe
ratio) for the different asset classes, where the statistics are averaged over these increasing
samples. Figure 1 plots the annualized standard deviations against the annualized average
returns. It is easy to see from this graph and table 1 that the S&P500 performs better than any
iShare on a risk-adjusted basis. The South-East Asian ETFs are among the poorest performers.
This is expected since the sample covers the Asian and Russian crises of 1997 and 1998.
5
Table 1 - Summary statistics
The means, standard deviations and correlations are annualized and measured over increasing
samples ranging from April 1996-April 1999 to April 1996-January 2004 in increments of one
observation at a time. The table presents averages of the mean returns, averages of the standard
deviations and averages of the correlations over these increasing samples.
Figure 1 - Risk and average return: averages across increasing samples
Table 1 also presents of the return correlations between the S&P500 index and the countryspecific iShares. As with the means and standard-deviations above, the correlations are measured
over increasing samples and the average correlations are reported in table 1. As expected, the
correlations are higher with the North American and UK ETFs (average correlation of 0.75) and
lower with the Pacific Asian ETFs (average correlation of 0.58). These correlations are less than
one, suggesting, as in Solnik (1974), that the partial segmentation of capital markets leaves ample
room for risk reduction through international diversification. The high standard deviations in
the correlations between the S&P500 and some iShares (in particular, those of Sweden, Austria,
Germany and Italy) confirm that correlations did change over time and therefore that optimal
asset allocations are also time-dependent.
6
Altogether table 1 and figure 1 illustrate that, while country-specific iShares show poor performance
as stand alone investments, their low correlations with the S&P500 make them good candidates
for inclusion in a well diversified portfolio.
5. Results
Optimal Asset Allocations
Since the parameters of the optimization problem (means, standard-deviations and correlations)
change over time, I calculate the optimal asset allocation over increasing periods, ranging from
April 1996-April 1999 to April 1996-January 2004, in increment of one observation at a time.
The optimal asset allocations are calculated at the end of each month from April 1999 until
January 2004. Figures 2 to 5 present the average asset allocations of the optimal portfolios and
the standard deviations of the optimal asset allocations (in parenthesis). Figure 2 looks at the case
where the asset manager only takes long ETF positions, while figures 3 to 5 allow for up to 5, 20
and 100% short ETF interests respectively.
Figure 2 - Average optimal asset allocation with long ETF positions (standard deviations in parentheses)
Figure 3 - Average optimal asset allocation with up to 5% short ETF positions (standard deviations in parentheses)
7
Figure 4-Average optimal asset allocation with up to 20% short ETF positions (standard deviations in parentheses)
Figure 5-Average optimal asset allocation with up to 100% short ETF positions (standard deviations in parentheses)
Figure 2 shows that an investor maximizes his expected Sharpe ratio by investing half of his
wealth in the S&P500. This result is intuitive as the United States represents roughly 50% of
world market capitalisation.2 The remaining half is invested in the iShares of Spain, Italy, the
UK, Sweden, Canada, Mexico and France (in order of decreasing importance). The other nine
ETFs have a zero weight. A glance at figure 1 can explain the relative importance of each asset.
As expected, the assets that lie in the North-West of the graph have higher weights than the
dominated assets in the South-East of figure 1. Note also that the relatively high standard
deviations of the optimal asset allocations in parentheses (especially for the S&P500 and the
iShare of Spain) indicate that the optimal asset allocations are quite volatile and time-dependent.
This is in line with the idea that the world economy underwent a recession over part of the period
covered. As a result, average returns decreased, standard deviations and correlations increased,
leading to optimal asset allocations that were also changing.
Figure 3 allows for short iShares positions of up to 5%. Per dollar invested, the optimizer
recommends short-selling $0.44 and investing the proceeds in the S&P500 and the iShares of Spain,
8
2 - International Federation of Stock Exchanges www.fibv.com
Italy, the UK, France, Canada and Sweden (in order of decreasing importance). The iShares that are
sold short are those of Australia, Austria, Belgium, Germany, Hong Kong, Malaysia, Mexico, the
Netherlands, Singapore and Switzerland. As expected, these iShares were dominated in figure 1
and got zero or near zero weights in figure 2. As such, they are good candidates as risk minimisers
through short positions.
Typically, the average short ETF position equals 20% of the ETF market capitalisation. With this in
mind, figure 4 presents the average optimal asset allocation when short ETF interests are restricted
to 20%. Now, per dollar invested, the investor increases his short position threefold relative to the
5% short case. The optimizer indeed recommends short-selling $1.38 and investing the proceeds
in the asset classes that offer the best risk-adjusted mean returns in figure 1.
Figure 5 considers the case of a hedge fund with short iShare positions of up to 100%.
The average optimal asset allocation changes dramatically once again. Per dollar invested, an
investor sells $6.84 short and invests the proceeds in the assets that offer relatively higher
returns and/or lower risk (such as the S&P 500 and the iShares of Spain, France and Mexico).
As in figures 3 and 4, the iShares that attract the shortest positions in figure 5 are those
of Singapore, Germany, Switzerland, Austria, Australia, Belgium, Hong Kong and the Netherlands.
These ETFs have relatively lower risk-adjusted returns in table 1. Therefore, selling them short
and investing the proceeds in assets with relatively lower risk and higher return enhances meanvariance efficiency.
A word of caution is warranted at this stage. This paper assumes that asset managers take the
asset allocations the optimizer defines as face value. In reality, asset managers only consider the
output of the optimizer as indication of portfolio weights. They typically recognize the limitations
of optimizers3 and combine the optimizer’s quantitative approach with their own forecast of the
relative performance of the different asset classes. This paper, on the other hand, reports only the
solutions of the optimization problem. Because of this limitation, out-of-sample return forecasts
are meaningless in this setting.
6. Mean-Variance Efficient Frontiers
Figure 6 plots the mean-variance efficient frontiers with and without risk-free lending and
borrowing of an US investor who considers i) the S&P500 only, ii) the S&P500 along with long and/
or short iShares positions over the whole sample. It is obvious from the graph that the inclusion of
country-specific ETFs shifts the frontier of an US investor to the North-West. The results are even
more dramatic when ETFs are sold short. Because long and short strategies with ETFs consistently
dominate the US mean-variance efficient frontier, investing in country-specific ETFs enhances the
asset allocation of an US investor.
Figure 6 - Mean-variance efficient frontiers under different restrictions
3 - Mean-variance optimizers are very sensitive to small changes in the initial estimates of the return distribution. Besides, the estimates with the biggest measurement errors are likely to have the most
prevalent influence on the optimal asset allocation. For this reason, mean-variance optimizers have been criticized as “estimation error maximisers” (Michaud (1989)). Attempts have been made to correct
for these limitations, using, for example, Bayes-Stein estimators (Chopra, Hensel, and Turner (1993)) or by constraining the outputs (Sheedy, Trevor, and Wood (1999)). In the end, mean-variance optimizers
cannot replace the informed judgement of an asset manager. The optimizer is simply a computational convenience that helps him translate his forecasts of assets’ returns, risks and correlations in an efficient
portfolio.
9
In an attempt to quantify the gains from investing in country-specific ETFs, I consider two
representative clients with different degrees of risk aversion. The first one is relatively more risk
averse and requires a standard deviation of only 10%. The second one is willing, if necessary, to
borrow at the risk-free rate to achieve an average return of 25%.
Table 2 reports the average return of client 1’s portfolio and the standard deviation of client
2’s portfolio under different restrictions. The table shows that we can systematically increase
the average return and decrease the standard deviation, of a US-only portfolio by investing in
country-specific ETFs. For example, for a standard deviation of 10%, allowing long positions in
country-specific ETFs substantially increases average returns by 17%. This illustrates the benefits
of portfolio diversification as previously reported by Levy and Sarnat (1970), Solnik (1974), or
Odier and Solnik (1993). By diversifying internationally using iShares MSCI country funds, we
eliminate country-specific risk and consequently enhance mean-variance efficiency.
Table 2-Efficiency Test
Because it is expensive, and even often impossible, to sell conventional country funds short, global
investors could not in the past fully benefit from international portfolio diversification. However,
the introduction of country-specific iShares and the possibility of selling them short have opened
up new horizons. Figure 6 and table 2 clearly indicate that short selling ETFs enhances the benefits
of international portfolio diversification even further. Compared to a US investment, short-selling
up to 20% iShares increases the average return of a 10% risk US portfolio by 73%.
Similarly, for an average return of 25%, an investor can reduce his risk by more than half if
he takes short positions of up to 20% in ETFs. A more conservative investor, unwilling to short
sell ETFs by more than 5%, can still enjoy lower risk than an investor who is only long ETFs. The
efficiency gains are even more substantial for hedge funds; namely, when iShares are allowed
to be sold short by up to 100%. Then, relative to as US-only investment, the average return of a
portfolio with a 10% risk is increased by 130%, while the risk of a portfolio with a 25% average
return is decreased by 69%.
7. Conclusions
It is well known that country-specific ETFs offer the benefits of international portfolio diversification
at a lower cost, with a lower tracking error and in a more tax-efficient way than passive open or
closed-end country funds. This article focuses on another important feature of country-specific
ETFs, short-selling. It argues that this feature enhances the benefits of diversification even further
and makes ETFs a more efficient approach to global asset allocation. The results indicate that US
investors, who take long and short positions in iShares MSCI country funds, can achieve higher
average returns for a given level of risk or, alternatively, lower risk for a given average return. On
the other hand, global index funds cannot be sold short on a downtick. As a result, they are poorer
portfolio diversifiers: their inclusion in a portfolio leads to asset allocation that is inefficient
relative to country-specific ETFs.
Since stock index futures can also be sold short at any time during the trading day, similar efficiency
gains could be achieved with stock index futures. Unlike country-specific ETFs however, stock
10
index futures are not tax friendly, they need to be rolled over before maturity, they are marked to
market and they are subject to margin calls. These features could make them less profitable and
more burdensome to long-term investors. ETF futures4 were introduced in November 2002, making
it possible now for investors to benefit from the tax advantages of ETFs and the leverage of futures
all at once. In the not too distant future we can reasonably expect more financial innovation in
this area. The introduction of futures on country-specific ETFs would then make it possible to
replicate the efficiency gains of short-selling mentioned in this paper, without sacrificing the tax
efficiency of country-specific ETFs.
4 - US examples include the DIAMONDs futures and the ETF futures that track the Nasdaq 100, Russell 1000, Russell 2000 and Russell 3000 indices. Similar financial innovation took place in Europe with the
Euro Stoxx 50, DAX and SMI ETF futures.
11
References
• Carhart, M., (1997), ‘On Persistence in Mutual Fund Performance’, Journal of Finance, 52,
57-83
• Chang, E., and G. S. Swales, (2003), ‘Do Country-Specific Exchange-Traded Funds Outperform
Closed-End Country Funds?’, Working paper
• Chopra, V. K., C. R. Hensel, and A. Turner, (1993), ‘Massaging Mean-Variance Inputs: Returns from
Alternative Global Investment Strategies in the 1980s.’ Management Science, 39, 845-856
• Dahlquist, M., and C. Harvey, (2001), ‘Global Tactical Asset Allocation.’ Emerging Markets Quarterly,
6-14.
• Fuhr, D., (2001), ‘Exchange Traded Funds: A Primer.’ Journal of Asset Management, 2, 260-273
• Gastineau, G., (2001), ‘Exchange-Traded Funds: An Introduction.’ Journal of Portfolio Management,
27, 88-96
• —, (2002), The Exchange-Traded Funds Manual, John Wiley & Sons
• Frino, A., and D. R. Gallagher, (2001), Tracking S&P500 Index Funds, Journal of Portfolio
Management, Fall, 44-54
• Gruber, M., (1996), Another Puzzle: The Growth in Actively Managed Mutual Funds, Journal of
Finance, 55, 783-810
• Harper, J., J. Madura, and O. Schnusenberg, (2003), ‘Performance Comparison Between Exchange
Traded Funds and Closed-End Country Funds’, Working paper
• Jensen, M., (1968), The Performance of Mutual Funds in the Period 1945-1964, Journal of
Finance, 23, 2, 389-416
• Kostovetsky, L., (2003), ‘Index Mutual Funds and Exchange-Traded Funds.’ Journal of Portfolio
Management, 29, 80-92
• Levy, H., and M. Sarnat, (1970), ‘International Diversification of Investment Portfolios.’ American
Economic Review, 60, 668-675
• Longin, F., and B. Solnik, (2001), ‘Extreme Correlation in International Equity Markets.’ Journal
of Finance, 56, 649-676
• Malkiel, B. G., (1995), Returns for Investing in Equity Mutual Funds: 1971 to 1991, Journal of
Finance, 50, 549-72
• Michaud, R., (1989), ‘The Markowitz Optimization Enigma: Is ‘Optimized’ Optimal?’ Financial
Analysts Journal, (January-February), 31-42
• Odier, P., and B. Solnik, (1993), ‘Lessons for International Asset Allocation.’ Financial Analyst
Journal, 49 63-77
• Poterba, J., and J. Shoven, (2002), ‘Exchange Traded Funds: A New Investment Option for Taxable
Investors.’ American Economic Review, 92 422-427
• Sharpe, W., (1964), ‘Capital Asset Prices: A Theory of Market Equilibrium under Conditions of
Risk.’ Journal of Finance, 19 425-442
• Sheedy, E., R. Trevor, and J. Wood, (1999), ‘Asset Allocation Decisions when Risk is Changing.’
Journal of Financial Research, 3, 22 301-315
• Solnik, B., (1974), ‘Why Not Diversify Internationally Rather Than Domestically?’ Financial
Analysts Journal, 30 48-54
12
• Solnik, B., C. Boucrelle, and Y. Le Fur, (1996), ‘International Market Correlation and Volatility.’
Financial Analysts Journal, 52, 17-34