EAFE is used by the institutional investment community for the following purposes:

• A performance benchmark for the non-U.S. stock market.
• A proxy for foreign stocks as an asset class in asset allocation (AA) models.
• The eligible universe of non-U.S. stocks from which portfolio holdings must be selected.
• The basis for foreign industry and sector groups.
• The basis for derivatives and structured equity products.
• A vehicle for index-driven quantitative management techniques, principally indexation.

Five major developed index "families" are currently available for the developed foreign markets. The first and most prevalent family of indices, of course, belongs to MSCI, a subsidiary of Morgan Stanley & Company which maintains separate operations and a "Chinese wall" from its parent investment bank. This family of indices was first introduced by Capital International, an investment company, in 1968. In addition to EAFE, the MSCI family includes (as of December 31, 1996): 49 country indexes, 51 regional indices, 2,000 industry indices, 350 custom indices, company-level and index-level valuation ratios, and historical total returns and constituent data that date back to January 1, 1970. According to MSCI, "the overall goal of the MSCI Indices is to represent the opportunity set of the global investor, by retaining the spirit of an all-share index with a key difference. MSCI indices are designed to be investable and replicable."²

In this manner, each index in the MSCI family is intended to represent an accurate normal portfolio. An ancillary objective of the indices is to provide one consistent standard across all MSCI indices. In 1986, Morgan Stanley & Company, Incorporated acquired the indices and data from Capital International Perspective, S.A. (CIPSA). Again, according to the MSCI literature³, CIPSA is solely responsible for decisions regarding constituent additions and deletions as well as any other methodological changes to the indices, the contributions of Morgan Stanley include its expertise, technology, and marketing resources.

The MSCI indices were created at a time when the only significant barometers of foreign market returns were local market indices. Local market indices differ drastically from market to market in terms of the ways in which they are calculated and compiled. Arriving early to the party has been a boon to the popularity of MSCI. According to screens provided by the Frank Russell Company's Russell Performance System⁴, more than 90% of pension plans that were benchmarked internationally were benchmarked to EAFE as of December 31, 1995. In addition to MSCI, four other sets of indices available to international investors. The competition is intriguing because each one has its own set of strengths and weaknesses relative to the others, and the differences can be substantial. These index families include:

• The Financial Times/Standard & Poor's Actuaries World Indices (FT/S&P).
• The Salomon Brothers Broad Market Indices (BMI).
• The Dow Jones Global Indices (DJGI).
• The local market indices on which futures are based.

Financial Times/Standard & Poor's-Actuaries World Indices are jointly owned by FT-SE International, Standard & Poor's Corporation, Goldman, Sachs & Company - although the last no longer has calculation or computation responsibilities. FT-SE (usually pronounced "footsie") International is a limited partnership between the Financial Times Limited and the London Stock Exchange. McGraw-Hill-owned Standard & Poor's joined the consortium in 1995, replacing founding member, NatWest Securities. The Financial Times, in turn, is owned by United Kingdom conglomerate Pearsons Limited.

The FT/S&P were introduced in1986 as the Financial Times Actuaries World Indices. To serve the needs of the investment community, a complete set of reconstructed monthly total returns and is constituent data has been created which dates back to January 1981. There are no official fundamentals available for these indices, but Goldman, Sachs & Company publishes a monthly booklet that provides unofficial ratios and statistics for both the indices and their constituents, according to Mark Makepeace⁵ president of FT-SE International. It is widely used by European fund managers for global benchmarking, but places distant second in pervasiveness to the MSCI in the Americas, Asia, and Australia.

The objective of the FT/S&P, according to Makepeace is, "to create and maintain a series of high-quality indices of the international equity market for use as a benchmark by the global investment community." In pursuit of this objective, the indices aspire to be comprehensive, consistent, flexible, accurate, investable, representative, and user-driven.

SALOMON BROTHERS BMI

The Broad Market Indices were introduced jointly by Salomon Brothers and Frank Russell Company on June 30, 1989. It is the only one of the five global families that is 100%-heuristically driven, and  only family that attempts to represent fully every investable company worth more than $100 million in free float market capitalization. The family includes (as of January 31, 1996) 21 country indices, 18 regional indices, and 39 industry indices that are updated regularly in Salomon Brothers publications. However, the nature of the index construction methodology lends itself to customization into any user defined sub-groupings desired. into nine sectors. Thomas S. Nadbielny, director of index research at Salomon Brothers and co-creator of the indices, maintains a profound conviction that the BMI are the best of the international benchmarks.⁶ "When we looked at the existing set of global indexes," he recalls, "we noticed that the high ground was left open and we took it. By that I mean that the existing set of indexes presented like a beauty contest with a lot of perfume and makeup, but none of the other indexes actually measured the global equity marketplace as it pertains to the institutional investor. "Not only were the indexes to be weighted by market capital" None of the other indexes were either comprehensive or float-weighted, two attributes essential to actually measuring the marketplace from the portfolio's shareholders perspective. We present the market as it is, not as we wish it to be. If you had all the money the world, the Salomon Brothers World Equity Index is what you could buy." Compilation and calculation responsibilities are now handled by Salomon Brothers. (Research is separated by Chinese wall from the sales area.). The primary institutional uses are for customized bench-marking and to serve as a basis for structured over-the-counter derivative products. Customized benchmarks include, but are not necessarily limited to style indices and industries and sectors at the country, regional, and global levels.

Nadbielny asserts that the Salomon Brothers' indices are in the best position to meet the specialized needs of pension funds and retirement systems: "Most pension funds come to us because coverage of small-cap issues. Overall we have three to four times number of companies as our main competitors. Our competition is weakest with their small-cap product. In addition, our index can be tracked and hedged with baskets of local market indexes [e.g., the
DAX, CAC 40, TOPIX, etc.] closer than the competitor products. This means that a fund can get a double benefit by using our index: small-cap coverage and lower hedging and synthetic tracking costs."

DJGI

The Dow Jones Global Indexes, the relative newcomer to the group, were introduced in 1993 and today encompass 29 countries, into three main regions and four sub-regions. More than 2,800 stocks are classified into 120 industry groups, which are in turn gathered into nine sectors. It is noteworthy that the DJGI is the only one of the major index families with no investment banking relationships.

John A. Prestbo, editor of the Dow Jones Global Indexes, states that these indexes were created from a fresh perspective:"⁷The initial impetus for Dow Jones to create its own global index series was the need for a quantitative means of telling the story of daily stock trading around the globe in the company's news products, including The Wall Street Journal. From the start, however, the indexes were developed with the investment community in mind.

"Not only were the indexes to be weighted by market capitalization, but also the geographic groupings of countries and regions would be complemented by a more detailed breakdown of economic sectors and industry groups than was available from other index providers," Prestbo explained. "Moreover, the indexes were to be made up of investable stocks that investment managers could easily buy and sell; three liquidity screens were developed to weed out slow-trading securities.

"In 1995, Dow Jones authorized licensing its global indexes for investment products, the first time it had done so in its history. In 1996, it began calculating the indexes in real time, every 15 seconds, and distributing them over Dow Jones Markets [formerly known as Telerate].

"The ambitions of Dow Jones for its global indices are formidable, indeed," Prestbo adds. "For years, Dow Jones has been associated with the best-known and most widely quoted stock index in the world, which is the Dow Jones Industrial Average. It is our intent to make the Dow Jones Global Indexes the standard for the 21st century, in terms of accuracy and scope as well as clarity and overall use fulness to investors."

LOCAL MARKET INDICES

Finally, the local market indices are the global indices best known by investors, especially by retail investors and the popular media. Each one is owned, compiled, and managed by a different entity - usually, but not always, the exchange in question. Most foreign index futures contracts are based upon the local market indices. In coupling a general lack of breadth with a leadership in media citations, the local market indices can be thought of as generalized pulse-takers - most similar to the Dow Jones Industrial Average. With a few notable exceptions, such as Australia, most of these indices are narrower and less diverse than the more institutionally oriented families discussed earlier in this chapter.Construction and calculation methodologies vary widely among the local market indices.

COMPARATIVE DIFFERENCES

Among the institutional indices, the greatest distinctions between families are generally in the construction, compilation, calculation, and dissemination methodologies. A comprehensive discussion on these methodologies would be far too lengthy and tangential for the scope of this chapter. Putting the local market indices aside for the moment, the key differences, in my opinion, could be summarized as follows:

Varying Proportions

Each index tries to represent a different proportion of the investable capitalization in each market: the MSCI, 60%; the DJGI, 80%; the FTIS&P, 85%, and the BMI, 100%.

Rules

Each index publishes construction and maintenance rules and standards. The BMI is the only family for which the rules are applied 100% heuristically without need of a committee. The DJGI and FT/S&P have very precise rules which, in practice, govern about 99% of index maintenance issues even though both have Policy committees that are consulted in highly unusual situations. The MSCI rules, first made available in 1996, are closer to broad brush strokes than precise prescriptions. Although these rules Provide general guidelines for how constituents are selected, MSCI is the only one of these four families in which it is not generally possible for investors to anticipate index changes based upon market constituency changes and corporate actions. The committee generally is needed to choose among a number of available stocks - all of which satisfy the stated criteria. A West Coast-based investment manager who declined to be quoted for attribution muttered that the MSCI index committee can be depended upon to add securities "Out of left field."

Market Capitalization

The BMI, DJGI, and FT/S&P all include only the capitalization of those issues that are available to international investors. The BMI and DJGI adjust all holdings' market capitalization for free float and cross-ownership.

The FT/S&P adjusts for the percentage of capitalization held by the government, but not for cross-ownership. MSCI addresses the cross-ownership issue in a different manner: It simply does not include a stock at all if the committee determines that its cross-ownership with another index stock is so great that its inclusion would have the effect of double-counting. But any stock that is selected goes in at 100% of its full market capitalization.

Index Constituents

The different treatments of these issues can lead to wide variations in country index constituents. For example, when Deutsche Telekom was privatized in 1996, the government kept 80% and made 20% available to investors.

Deutsche Telekom, therefore, was put in at 20% of its market value in BMI, DJGI, and FT/S&P, but 80% in MSCI (a special committee decision to amend the normal 100% rule in the case of huge privatizations). Percentage-wise, the new issue comprised between 1.9% and 2.3% of the first three indices, but a whopping 6.9% of MSCI.

"Any way you slice it, the inclusion of that many unavailable shares in a country index represents a significant distortion," observes  John Blin⁸, a principal of APT, Inc., a New York-based asset management firm and a leading provider of risk-management systems.

Index Levels

Official DJGI index levels are all available in real-time. The other three indices publish their levels just once per day. Real-time pricing could prove very useful in certain hedging and risk management applications.

The distinctions above favor different indices for different usages. There are a few other relevant points which should be mentioned that unquestionably favor Morgan Stanley. These include:

Length of History

For the majority of its indices, more than 26 years of live data exist for MSCI, beginning January 1, 1970. In contrast. the FT/S&P have 16 years of history, five of it backtested. The BMI have just under eight, and the DJGI just over three. Since many of the global markets, notably excepting Japan and Italy, have enjoyed a secular bull market since 1982, this does not provide for much data across market cycles. Therefore, MSCI is currently the only family of institutional indices that provides the opportunity to evaluate historical market behavior in and out of market cycles, although both the BMI and the DLGI report that they have ongoing projects to extend their historical data through algorithmically executed backcasting of relevant data.

Official Fundamentals

Although all three of the other families said that official fundamentals were in their future plans, only MSCI offers official fundamentals today (e.g., P/E, price-to-cash flow, etc.). This can be crucial for tactical asset allocation (TAA) applications that attempt to emphasize data. Max Darnell, director of TAA for Pasadena-based manager, First Quadrant, says, "we need the various valuation ratios along with the return series to construct our developed market allocation models."⁹

Direct Investibility

The majority of MSCI country indices can be bought like stocks by Americans and others through the WEBS on the American Stock Exchange. Another stock-like product called OPALS (not available to U.S. citizens) turns both MSCI country and regional indices into stocks. Although the local market indices are investable on a number of exchanges through listed futures and options, the only one of the three other institutional indices that is similarly investable for U.S. citizens is the Dow Jones Taiwan Index - with both futures and options on the Chicago Mercantile Exchange and the Pacific Stock Exchange.

Nine FT/S&P country indices were investable on the New York Stock Exchange through shares of a series fund called Country Baskets for about 11 months ending February 1997; this ill-fated fund has since been liquidated.

The differences detailed above, both pro and con, may make a difference in the efficacy of the purposes for which these major index families are utilized by investors. The first usage, performance benchmarking, is very important to the plan sponsor and pension consultant communities.

Without a proper benchmark, it is difficult to evaluate the performance of investment managers. Indeed, many plans take this issue so seriously domestically that they use customized style indices to evaluate each of their domestic managers separately. With some exceptions, such rigor is rarely applied internationally. Here, an  EAFE mandate can certainly be considered the standard of the U.S. investment industry.

There are two issues here: Does EAFE represent an appropriate grouping of stocks and nations which all international managers should be measured? Does the MSCI relative dearth of investable stock representation produce a significant return distortion? The former question can be meaningfully evaluated only with respect to an individual plan and the style of the selected manager.

The second question is more universal. Conceptually, if the purpose of a benchmark is to evaluate the performance available to an investment manager, then the BMI would logically appear to be the best choice. Indeed, the BMI does come closest to reflecting the returns of market-cap weighted exchange indices where they exist, providing evidence that the BMI provided the statistically most representative monthly returns.

However, an examination of the differences in monthly country returns on the three indices that have been in existence at least seven years brings to mind the title of a Shakespeare play, Much Ado About Nothing. The correlation coefficient between the common country indices between the BMI and MSCI is 0.993. The average absolute value of the monthly differences was 19 basis points. On a country-by-country basis, there doesn't seem to be a compelling reason to switch out of MSCI because of distortions in country returns.

ASSET ALLOCATION APPLICATIONS

A good example of the asset class proxy usage is the controversial paper in which Rex Sinquefeld¹⁰ questioned the benefits of international diversification. Throughout this paper, international diversification was represented solely by the monolithic EAFE.

Does such usage constitute an oversimplification? Would the benefits of diversification be more apparent if each regio>

Transfer interrupted!

These are among many important questions to be considered when evaluating any issues concerning asset allocation. MSCI has answers for these questions, of course. For example, an important part of Canada's market capitalization is included in the Standard & Poor's 500 Index, while no companies from Australia or New Zealand are included there. Singapore's GDP per capita level is considerably higher than South Africa's. But these explanations are not rules that are systematically applied as conditions for includion within EA.FE.For example, about 40% of the MSCI-Netherlands index comes from constituents of the S&P 500. So each investor must evaluate critically the appropriateness of EA-FE versus other foreign stock asset class proxies to determine whether EAFE meets requirements. On a country-by-country basis, though, investors wishing to use individual nation asset-class proxies have a much more clearly defined decision. With longer historical return and fundamental time series available, the MSCI country indices are easily the most superior alternative for asset allocation purposes.

SELECTION SET ISSUES

The eligibility issue can be one of the thorniest constraints with which an active manager must contend. An investment manager who is constrained to pick among stocks included in a given index needs that index to include the broadest possible prudent selection in order to allow for the greatest potential for adding value through stock selection. Similarly, the ideal situation for such a manager to add value occurs when the index's rules for inclusion are consistent and predictable.

If a plan believes in a manager, it would, in turn, want the manager to be exposed to the type of situation that allows for maximum opportunity to add value. Therefore, a prudent fiduciary would also evaluate the global index family used as a selection set constraint on this basis. Here, the BMI, by design, is clearly the strongest choice while MSCI is the weakest.

Using an index's constituents to define an industry group can lead to u classic case of "garbage in, garbage out." One interesting parallel occurred with Standard & Poor's 500 Index as late as yearend

1993. Exhibit 1 lists the nine companies that comprised S&P's computer software industry group at that time. The overall sales of this group sorely understated the industry's contribution to GDP that year because Microsoft, which accounted for more than 60% of industry sales by itself, was not yet included in the S&P 500 or the derivative S&P 500 industry groups. S&P Industry groups are used not only for benchmarking, but for establishing regulatory standards, forecasting models, and as a basis for passive management. The validity of all such uses was severely compromised by Microsoft's exclusion.

Similarly, the EAFE utilities industry does not include STET, despite the fact that STET as of November 30, 1996, is the third largest stock by market capitalization traded on the Milan exchange. In both cases, the index providers had rationales for excluding the stocks in questions that related to cross-ownership issues. Yet, that explanation begs the question. Suppose that the purpose of having a managed index such as MSCI and S&P, as opposed to a rules-driven index (e.g., BMI, Russell 3000), is to allow the committee to make the benchmark the best proxy for the underlying market.

If that's the case, how can that committee justify the omission of industry leaders that, far from being illiquid as a result of the cross-holdings issue, were among the top five in their respective markets for average daily dollar-weighted trading volume for the time periods indicated? Therefore, for the purpose of defining industry groups and related subgroups, the BMI and DJGI, by their respective designs, appear to be the strongest alternatives while MSCI-bound industry standards have the greatest potential for providing incomplete groupings.

STRUCTURED EQUITY PRODUCTS AND OTC DERIVATIVES

One interesting sidelight to investigating the strengths and weaknesses of the alternative global index families is the number of investment banks involved in the global index business. The ownership and licensing aspects of two of the four index families, MSCI and Salomon Brothers, are controlled by investment banking firms; likewise, the FT/S&P licenses the rights to structured products based on the index through Goldman Sachs & Company - a 33 1/3% ownership partner. In

contrast, not one of the dominant U.S.-oriented index families -Standard & Poor's, Dow Jones, Russell, Wilshire, or Value Line - is owned or licensed through an investment bank. The main reason for the investment banks' desires to be involved in global indexation is readily apparent. Structuring global equity products for the needs of  pension plans and retirement systems is one of the most lucrative derivatives-related activities performed by Wall Street firms. In terms of volume, MSCI is the clear leader in such activities. In terms of the ability to structure the most client specific kinds of customized products, the DJGI with its combination of country, industry, and company exposure would be best suited to the task. According to C. Michael Carty, president of New Millennium Investment Advisors, "a sponsor looking to get exposure in a particular country's industry sector would most likely choose the DJGI since industry sectors in those indexes are more finely drawn."¹¹ However, such a sponsor would currently find the application unavailable. It is currently the only index family of those examined in this article with no current relationships with any Wall Street derivatives desk.

PASSIVE INDEXATION

Indexation is big business. The enormous growth during the past 10 years of the international-index-driven assets under management by Barclays Global Investors (BGI) and State Street Global Advisors (SSGA) provides ample evidence of this simple fact. Both BGI and SSGA utilize the MSCI and the FTIS&P indices for plan sponsor products, but both cite far more demand for MSCI products than FT/S&P products. Both reported little demand for BMI products and no demand for DJGI products. From the manager's viewpoint, a combination of the fewest stocks per index and lowest turnover should make the MSCI the easiest to manage. But, as one manager cautions, "the inability to predict index changes in a systematic manner makes it difficult to avoid substantial implementation shortfalls when sudden surprises occur, forcing the manager to chase the stock." So, all five families of indices have their relative strengths and weaknesses. The index of choice depends upon the individual needs of the investor. And, both investor needs and the makeup of the indices are constantly changing. So, analyses in the spirit outlined in this chapter need to be re-evaluated from time to time. As with virtually every aspect of the investment business, no substitute exists for proper due diligence.

EMERGING MARKETS

No discussion of global indexation could be complete without at least a brief discussion of the emerging markets. And, the two-word phrase, "emerging markets" is, in itself, a source of major conflict. How does one precisely define what constitutes an emerging market? Although each index family has its own published criteria for categorization, no singular industry standard exists. Specifically, as many as 11 countries - including Australia, Brazil, Hong Kong, Indonesia, Korea, Malaysia, Mexico, Singapore, South Africa, Taiwan, and Thailand -have been characterized as both emerging and developed, depending on the index family that includes them.

There also is no general consensus on the index provider of choice for the emerging markets. The most popular emerging market  index families belong to MSCI, the International Finance Corporation (IFC), and ING Barings. Other investment banks and asset managers (e.g., BZW, CS First Boston, etc.) have their own sets of indices, and some are quite well constructed. Nevertheless, applications outside their organizations occur rarely, if at all.

MSCI duplicates its developed market index construction process for the emerging markets, seeking to create indices that capture the general trend in the market without including every single stock in that market above some size threshold. MSCI compiles two emerging market indices: MSCI-Emerging Markets Global (MSCI-EM and MSCI Emerging Markets Free (MSCI-EMF). MSCI-EMF eliminates entirely securities from MSCI-EM that non-national investors cannot own. Since most users, - i.e., plan sponsors, superannuation funds, asset managers, etc, - of the emerging market index families come from outside each emerging market index, the comparisons in the remainder of this chapter concentrate solely on the MSCI-EMF. By year-end 1996, the MSCI-EMF consisted of 1,020 stocks in 26 country stock markets.

A member of the World Bank Group, IFC is owned by its more than 162 member countries. IFC is the world's largest multilateral source of financing for private enterprise in emerging economies, according to the IFC Emerging Markets Stock Factbook for 1996. In 1975, IFC started data collection for its Emerging Markets Data Base,which was introduced to the public in late 1981. Its objective is to serve as a vital statistical resource for IFC and the international financial community in its investment and advisory work. The IFC data base covers 27 markets, providing regular updates on more than 1,650 stocks. Monthly indices are available from December 1975 to the present. The IFC Investable Indexes (IFCI) were introduced in March 1993 to reflect the accessibility of markets and individual stocks for foreign investors. It is this index that is used in this chapter for emerging market index comparisons. By year-end 1996, the IFCI contained 1,225 stocks in 26 country stock markets, 25 of which it shares with MSCI-EMF; the difference is that the MSCI-EMF includes Israel while the IFCI includes Zimbabwe.

ING Barings is an international investment bank wholly owned by ING NV, a major Netherlands-based banking institution. ING Barings publishes several series of emerging market indices. Instituted in October 1992, the ING Barings Emerging Markets Index (BEMI) is intended to be a measure of liquid investment available to non-nationals in the global emerging equity markets. The objective of the BEMI is to create a representative benchmark of markets and shares that international investors can realistically buy or sell individually or as a basket. Accordingly, the BEMI country indices tend to have less breadth than their IFCI and MSCI-EMF counterparts. As of December 31, 1996, the BEMI comprised 449 equities in 21 global markets.

METHODOLOGIES

The BEMI, IFCI, and MSCI-EMF are all market-capitalization-weighted families of indices. BEMI has fewer markets than the other two because it has the tightest rules for country inclusion. To qualify, markets need to have adequate depth and breadth, defined as at least $2 billion (U.S.) in capitalization, no less than 100 actively traded companies, and free of onerous foreign investment restrictions. In addition to not including either Israel or Zimbabwe, Hungary, India, Jordan, or Sri Lanka are not included in the BEMI family; special stand-alone country indices are available for Hungary and India for benchmarking purposes. IFCI and MSCI-EMF both utilize combinations of criteria involving GDP per capita, size, openness to foreign investment, and liquidity; neither applies these criteria to foreign markets being considered for inclusion in any expressly formulaic fashion. Israel's GDP per capita was too high for IFCI standards of inclusion while MSCI-EMF has not yet included Zimbabwe due to general lack of market breadth.

All three families address the issue of foreign ownership restrictions in different ways. MSCI-EMF eliminates entirely share classes of stocks of stocks with foreign investment restrictions. MSCI makes a market level adjustment in Korea and Taiwan, weighting each, somewhat arbitrarily, at 50% of its total capitalization. Until September 1996, Taiwan was completely eliminated based upon foreign ownership restrictions. MSCI does not adjust the market level capitalization of Indonesia, India, and Thailand even though these markets all have similar foreign ownership restrictions. BEMI integrates foreign ownership adjustments with free-float adjustments. Each stock's capitalization is adjusted for the more restrictive of the two adjustment criteria.

For example, in Korea, 75% of the market capitalization of the Cho Hung Bank is excluded from the BEMI due to foreign ownership restriction. On the basis of free float alone, the adjustment would result in the exclusion of 20% of the same number. IFC applies the more restrictive of the following three factors to market capitalization: foreign ownership, adjustments for cross-ownership, and the exclusion of shares held by the local government. IFC applies these restrictions at the country level, sector level, and individual company level. BEMI and MSCI-EMF apply these restrictions only at the country level.

MSCI-EMF is a managed index with general rules and guidelines applied with some discretion by the CIPSA committee.Generally, MSCI-EMF aims to cover 60% of the market while emulating overall industry composition. MSCI states that its selections are made to produce "a representative sampling of large, medium, and small-cap companies, taking liquidity into account." An important distinction between MSCI-EMF is that with very few exceptions, every stock selected for inclusion is weighted at its full market capitalization, irrespective of government ownership or other factors. This can make a substantial difference in market concentration.

BEMI is constructed from a bottom-up perspective on the basis of screens. Only stocks with more than 1% of total available capitalization in each country are included. Stocks with a free-float of less than 10% of total capitalization or with an average daily trading dollar volume of less than $100,000 (U.S.) are also excluded. Contrastingly, IFCI combines a top-down approach with bottom-up screening criteria. The goal is to select a representative group stocks covering 60% of the investable universe for each country' ' market.

First, the investable universe is determined by stocks not eligible for foreign ownership. A statistical snapshot taken of that universe to determine total capitalization and diversification guidelines.

The next step is to exclude all stocks with less than daily dollar trading volume of $40,000 (U.S.). Stocks with investable market capitalization of less than $25,000,000 (U.S.) also excluded unless average daily dollar volume is greater $100,000 (U.S.). Once liquidity and size criteria have been the companies are ranked in order of market capitalization. The largest 50% by this criterion are automatically included in the try index. Finally, the remaining companies are considered in order with an eye toward replicating the industry the overall universe as nearly as possible without excluding heavily traded companies.

INDEX CHARACTERISTICS

Each index family's characteristics reflects its construction and maintenance rules to a large extent. In terms of year-end 1996, the BEMI, almost by definition, was the narrowest of the index families. The BEMI country indices reflected an average of 43% of the underlying market's investable capitalization. The MSCI-EMF covered an average of 53% of the analogous statistic for its index. The IFCI was the broadest of the three, equaling its goal of reflecting 60% of underlying capitalization.

Of the three possible pairs of the three indices, the BEMI and MSCI-EMF are most correlated with a correlation coefficient of 0.92 during the prior eight-year period. By the same measure, the BEMI and IFCI are least correlated (0.86). For reasons relating to breadth, the IFCI demonstrated the lowest volatility with an annualized monthly standard deviation of total returns (20.5%); the BEMI had the highest such volatility (26.4%).

On the other hand, because it is broadest, the IFCI experienced the greatest annual turnover of the three indices and the BEMI had the least turnover, albeit only slightly lower than the MSCI-EMF. In terms of transparency as it relates to predictability of constituent changes, the IFCI and BEMI are both very transparent. Since the MSCI-EMF is "managed," its changes are sometimes more difficult to anticipate. However, according to Jeffrey Davis¹², managing director  of Schooner Asset Management in Boston, "sometimes, having discretion over changes rather than being governed 100% by hard and fast rules, can be a positive thing for an index family; this can be especially true in the emerging markets when highly unusual corporate actions, privitizations, or rules changes suddenly occur."

APPROPRIATENESS FOR INTENDED USES

The MSCI-EMF is the most widely used family of the three emerging market indices discussed in this chapter, although the IFCI also has a quite significant market share. Is the MSCI-EMF the most utilized because it is simply the best? As with the developed global indices, there is no single definitive answer, but the IFCI is certainly not an easily dismissed alternative.

Referring back to the intended uses, the IFCI combines the longest history of available returns with the most representative sample of each market, rendering it from a statistical standpoint, the most appropriate index for benchmarking purposes - unless the plan is so large that liquidity is a major issue. In the latter case, the larger a plan, the more relatively appropriate the BEMI becomes. Nevertheless, for most applications, any of the three families may serve as appropriate benchmarks.

The BEMI has neither the breadth nor the history to be a suitable asset-allocation proxy, either for the emerging markets as a whole or for any of the country regional or sector indices individually. The MSCI-EMF is a qualified candidate on all counts but, statistically, the IFCI would be slightly preferable on all criteria. Its breadth and transparency make the IFCI the obvious best choice for universe selection and for the creation of industry sub-groupings. Almost by definition, the BEMI is easily the least appropriate for these tasks. For most uses of structured products, all three index families would be appropriate. Theoretically, the IFCI would be the most appropriate. But, by far, the most widely used family for structured products is the MSCI-EMF, and it is certainly seems to be adequate for all but the most exacting of purposes. All three commonly serve as vehicles for passive management. The larger the plan, the more appropriate the BEMI is for an index manager. Conversely, in many individual country markets, the IFCI becomes very difficult to replicate in the smaller markets with substantial amounts of assets, especially in markets where daily currency fluctuations are a major issue.

SUMMARY

Most U.S. investors think of the local market indices, such as the FT- SE 100 and the CAC-40, if they think of anything at all, when the subject of global equity indices is mentioned. For institutions, however, there are four prominent families of global indices that cover the developed markets and three that cover the emerging markets. Morgan Stanley Capital International is the only family with significant product in both markets, although 10 of the countries covered in the Dow Jones Global Indices are considered emerging markets in the IFCI. MSCI also has the lion's share of the market, being especially dominant in its developed market penetration.

For the most part, the existence of such a wide variety of alternatives is good news for investors, because all seven families offer quality products. The relative appropriateness of each index family depends upon the specific usage and the size of the investment. The Salomon Brothers BMI supplies the broadest slices and the DJGI has the most history in the developed markets while the MSCI has the most historical and fundamental data available. Since the philosophy of the MSCI is that professional management should have a role in index constituent selection, its rules are somewhat less transparent than those of the other families discussed. The extent to which that is considered a blessing or a curse depends upon the application as well as the needs and size of the user. The BEMI is especially developed for  the needs of the large institutional investor wishing to do frequent trading in the emerging markets. The IFCI gives the best combination of breadth, consistency, and history. The MSCI Emerging Markets Free, however, seems to strike a desirable middle ground for many  investors; it is the most widely utilized such index among U.S. institutional investors

It could also be that the attempt to measure an art form such as emerging markets investment as a science is simply inappropriate.  After all, in the same book where Lord Kelvin gave this chapter its opening quote on the importance of measurability, he also proved mathematically that, "...a heavier-than-air flying machine, ... capable of transporting a human weighing more than 50 stone [about 110 pounds] is impossible."¹³

Footnotes

¹Lord Kelvin (William Thomson), Physics and Aerodynamics, (London: Eton Press, 1877), 63-64.
² Morgan Stanley Capital International Methodology and Index Policy (1996): 6.
³ Ibid., p. 5.
⁴ As of June 1996.
⁵ Mark Makepeace's quoted matter is from a personal interview the author for use in this chapter.
⁶ Thomas S. Nadbielny supplied this quoted matter as text, prepared especially for use in this chapter, with permission to cite.
⁷ Joseph A. Prestbo supplied this quoted matter as text, prepared especially for use in this chapter, with permission to cite.
⁸ John Blin's quoted matter is from a personal interview conducted by the author for use in this chapter.
⁹ Max Darnell, MSCI NewsWatch (October 1996): 4.
¹⁰ Rex Sinquefeld, 'Where Are the Gains from International Diversification?"Financial Analysts Journal (January-February 1996): 8-14.
¹¹ C. Michael Carty's quoted matter is from a personal interview conducted by the author for use in this chapter.
¹² Jeffrey Davis's quoted matter is from a personal interview conducted by the author for use in this chapter.
¹³ Lord Kelvin, Physics and Aerodynamics, 183.

^{Copyright 1997 by Frank J. Fabozzi Associates, New Hope, Pennsylvania}