|Benchmark Index Comparisons |
By: Hugh Wilson
Date: June 30, 2002
Table of Contents
Treatment of Shares Outstanding
Correlation, Volatility and Characteristics
In the US, plan sponsors, consultants, and investment managers have thoroughly dissected the many popular and well-known US equity indices with the goal of determining the most appropriate benchmark for asset allocation and performance measurement. However, in the case of global equity indices, this exercise is less common. This article seeks to compare and contrast the four most commonly used indices: the Morgan Stanley Capital International World Index, the FTSE All World Index, the Salomon Smith Barney Global Equity Index, and the Dow Jones Global Index.
The MSCI World Index is the dominant benchmark followed by the US investment community. It was originally developed in the late 1960s by Capital International Group and was purchased by Morgan Stanley in 1985. Because the Index has existed since 1969, Morgan Stanley can provide 25 years of valuation data and price and return history. The FTSE All World Index is second to Morgan Stanley in popularity, and is widely used in Europe. It is compiled by FTSE International Limited and was introduced in 1987. The Salomon Brothers World Equity Index was developed in 1989 in a collaborative effort with the consulting firm Frank Russell, which ended the partnership four years later. Lastly, the Dow Jones Global Index from the US provider is one of many index series they provide. There is a history dating back to 31st December 1991.
All four indices’ constituent weights are determined by market capitalisation, i.e., market price multiplied by shares outstanding, with an adjustment for the proportion of shares which are not freely available to investors. (Free float adjustment, is referred to in more depth below.) Country inclusion criteria are all similarly based on the size of the equity market, the freedom of capital movement, and the ability to repatriate dividends. As a result, the countries included in each index are the same, for the most part, although there are a few notable differences.
In creating the MSCI World Index, MSCI use a bottom-up approach that aims for 85% of free float adjusted market representation within each industry group, within each country. The companies included in the indices are intended to replicate the industry composition for each market. The chosen list of stocks is composed of a representative sampling of large, medium, and small-cap companies from each local market, with liquidity being an important factor in the selection of index constituents. Stocks of non-domiciled companies and investment funds are excluded from the individual country indices. The goal of the MSCI’s methodology is to create a benchmark which is highly replicable and investable, and provides a broad and fair market representation. At the end of December 2001, 1,516 stocks were included in the MSCI World Index across 23 markets, with a total capitalisation of US$16 trillion. The MSCI World Emerging (26 countries) contained a further 704 stocks with a capitalisation of US$0.5 trillion.
Table 1: Developed Market/Capital Distribution at 31st December 2001
|MSCI World Index ||US$tr ||Names ||% |
| || || || |
|>$10b ||12.390 ||325 ||77.79 |
|$1b-10b ||3.340 ||842 ||20.97 |
|$500m-1b ||0.153 ||212 ||0.96 |
|$250m-500m ||0.037 ||97 ||0.23 |
|<$250m ||0.007 ||41 ||0.04 |
| ||15.927 ||1517 ||100.00 |
| || || || |
|FTSE All World Index || || || |
|>$10b ||13.530 ||355 ||77.46 |
|$1b-10b ||3.762 ||936 ||21.54 |
|$500m-1b ||0.129 ||176 ||0.74 |
|$250m-500m ||0.032 ||88 ||0.18 |
|<$250m ||0.014 ||100 ||0.08 |
| ||17.467 ||1655 ||100.00 |
| || || || |
|Salomon World Equity Index || || || |
|>$10b ||12.220 ||329 ||64.96 |
|$1b-10b ||5.050 ||1638 ||26.84 |
|$500m-1b ||0.705 ||1012 ||3.75 |
|$250m-500m ||0.451 ||1273 ||2.40 |
|<$250m ||0.387 ||2946 ||2.06 |
| ||18.813 ||7198 ||100.00 |
| || || || |
|Dow Jones Global Index || || || |
|>$10b ||13.037 ||351 ||67.72 |
|$1b-10b ||5.276 ||1679 ||27.41 |
|$500m-1b ||0.616 ||856 ||3.20 |
|$250m-500m ||0.240 ||645 ||1.25 |
|<$250m ||0.082 ||546 ||0.43 |
| ||19.251 ||4077 ||100.00 |
| || || || |
Sources: Salomon, FTSE, Dow Jones 12/31/01
The FTSE All World Index was launched in 1987 and aims for at least 90% coverage of the equity capital available in each market. Countries are allocated into one of three categories, Developed, Advanced Emerging, and Emerging, depending upon their level of development, investor perceptions, and other subjective factors. Developed plus Advanced Emerging is known as the FTSE World Index, and Advanced Emerging plus Emerging is known as the FTSE All-Emerging Index. An investable universe is constructed and ranked on a country basis, based on full market capitalisation. Any company falling into the bottom 10% is removed, and investability weights are applied to the remainder. A sample of 85% to 95% of the investable universe is used to select individual stocks, allowing for economic groups and a minimum 5% free float. The FTSE All World Index is free float adjusted using a banding method, with intervals of 10% except the bottom band of 15%, and the two top bands of 50-75% and 75-100%.
At the end of December 2001, 2,261 stocks were included in the FTSE All World, of which 1,655 were in the 23 developed markets, (capitalisation of US$ 17.5 trillion), with a total capitalisation of US$18 trillion including the 25 All-Emerging markets. (All capitalisation sizes are free float adjusted.)
The Salomon Smith Barney World Equity Index uses a top-down methodology that extensively covers the universe of available stocks. At the end of each May, all listed shares of companies with a total available market capitalisation of at least US$100 million are included in the Index, while any stocks in the Index with market capitalisations that have fallen below US$75 million are removed, effectively representing over 95% of the total available market capitalisation. SSMB claim that all segments of the market, including all industry and economic sectors and capitalisation ranges, are fully represented. As at 12/31/01, the SSMB Broad Market Index (BMI) was composed of 8,151 stocks with a total market capitalisation of US$19.4 trillion (after investability adjustment). SSMB split their index by capitalisation, and offer developed/emerging markets subgroups. From the BMI, SSMB have created two mutually exclusive subindices for each country: the Primary Market Index (PMI), with a total of 900 large-capitalisation stocks, and the Extended Market Index (EMI) with 7,251 small-capitalisation securities. As at 12/31/01, the total capitalisation of the PMI was US$15.5 trillion, while the total capitalisation of the EMI was US$3.9 trillion. Each July, SSMB reconstitute the two subindices by ranking the stocks within each country index by total market capitalisation. The PMI universe is defined as those stocks falling within the top 80% of the total market capitalisation in each country. The remaining issues, or the bottom 20% in market capitalisation, form the small-cap EMI universe. In order to reduce the amount of turnover due to the annual reclassification (June 30), SSMB institutes a buffer of 5% around the 80/20 split.
The Dow Jones Global Index aims to cover 95% of market capitalisation at country level. As with FTSE and MSCI, there are the same 23 developed markets, but with greater coverage many more names are included. Large cap is defined as the top 70% of the investment universe, mid cap as the next 20%, and small cap as the bottom 10%.
The DJGI universe is created by first aggregating all stocks traded on the local exchanges of eligible countries and assigning each company to the country of its primary market listing. After screening out non–common issues and illiquid stocks, each remaining company is assigned to an industry group according to its primary line of business. The selection process then proceeds at the country level (for developed markets excluding Europe) or at the aggregate level (for all Europe and all emerging markets).
The indices are designed to be investable by including only sufficiently liquid stocks. Issues under consideration for index inclusion are screened to ensure that they are actively traded and are readily accessible to investors. The indices are free float adjusted to exact percentages, with a minimum 5% free float requirement.
Table 2: Total World/Market Coverage
|Index ||Market |
|Number of |
| || || |
|Salomon World Equity Index ||95% ||8,151 |
|Dow Jones Global Index ||95% ||4,867 |
|FTSE All World Index ||90% ||2,261 |
|MSCI World Index ||85% ||2,220 |
Sources: Salomon, FTSE, Dow Jones 12/31/01
At the end of December 2001, the DJGI was composed of 4,867 stocks across 34 markets, with a total market capitalisation of US$20.1 trillion (US$19.2 trillion in developed markets) after investability adjustment.
Treatment of Shares Outstanding (Free Float)
The greatest difference in construction methodology among the four benchmarks is the interpretation of total shares outstanding used to calculate a stock’s cap-weight in the index. There are four primary factors that can affect the float (or percentage of shares freely tradable) of a company:
1. Corporate cross-holdings: One company is a partial owner of another company, either within the same country or domiciled in another country. In Japan, this practice is called Keiretsu. According to Salomon Brothers over 40% of Japan’s total share capital is unavailable because of Keiretsu.
2. Private control blocks: An individual, family, or trust holds a significant percentage of an issue’s total shares outstanding.
3. Government holdings: The government is part owner of a private company.
4. Legally restricted shares: Companies are limited by either corporate charter or government restrictions as to the amount of shares that can be owned by foreign investors.
Morgan Stanley band in 1% intervals to15%, then 5% intervals. FTSE band in intervals of 10% except the bottom band of 15%, and the two top bands of 50-75% and 75-100%. Only Salomon and Dow Jones apply an exact factor, but they do not publish how this is arrived at. To prevent excessive turnover all providers require a significant move in free float percentages before adjusting the factor.
The SSMB free float is determined each July 1st, at the annual reconstitution, using company handbooks, annual reports, local stock exchange information, and press announcements. SSMB determine the split between their respective large-cap and small-cap indices slightly differently from other providers. They first rank the stocks in each country based on total, unadjusted market capitalisation. SSMB determine the lower boundary by splitting the rank order of each country at the stock where the cumulative available capitalisation reaches 80% of the total available capitalisation. In this way SSMB classify securities based on the total size of the company that issued the security, rather than the available float of the stock.
Thus, the ranking accurately reflects the fundamental characteristics of the company. The main differences between the providers arise from the interpretation of the free float. In the past MSCI simply had no free float adjustment, and FTSE only adjusted for partially privatized companies and foreign ownership restrictions. Both have now changed their stance. FTSE has the largest banding, with anything over 75% being rounded up to 100%. This gives rise to them being the only provider to include Microsoft at 100%. This single difference lifts the top end of the FTSE market cap by US$50 billion. Interestingly, a piece of 1930s US legislation regarding ownership of radio stations accounts for a large anomaly between Salomon and the other providers. SSMB’s interpretation gives a 25% free float on SBC Communications and Verizon (amongst others), whilst the other providers attribute a 100% float. This produces a huge US$200 billion difference between the SSMB World Equity Index and the other indices.
In general, the country weights are quite similar. All providers include 23 developed markets, and only Salomon add South Korea and the Czech Republic to developed. There is little difference in the weight of each developed market, the US being around 60% of them all. The biggest relative difference is Japan, where 7.5% is the norm, and DJGI has 9%. This is caused by the higher free float weights that Dow Jones attribute to many Japanese securities, with an average of 80%, compared with 60-65% for the others. Within Europe, France has a slightly higher representation in FTSE, and Dow Jones view Europe (including the UK) as a single entity, giving it a lower overall weighting.
All providers include around 25 emerging markets, except Dow Jones who only have 13. This is because Dow see Emerging Markets as a single entity, rather than defining which countries to include. They include any company falling within 95% of the market capitalisation of this entity.
| ||Salomon |
|FTSE All |
|Dow Jones |
|Austria ||0.05% ||0.04% ||0.05% ||0.03% ||$0.01 ||$0.01 ||$0.01 ||$0.00 |
|Australia ||1.43 ||1.54 ||1.47 ||1.58 ||0.27 ||0.25 ||0.26 ||0.30 |
|Belgium ||0.46 ||0.40 ||0.42 ||0.28 ||0.09 ||0.06 ||0.07 ||0.05 |
|Canada ||2.20 ||2.14 ||2.12 ||1.64 ||0.41 ||0.34 ||0.37 ||0.14 |
|Switzerland ||0.03 ||0.03 ||0.03 ||0.03 ||0.51 ||0.47 ||0.49 ||0.50 |
|Republic ||0.01 ||0.00 ||0.00 ||0.00 ||0.00 ||- ||- ||- |
|Germany ||2.71 ||2.74 ||3.01 ||2.65 ||0.51 ||0.44 ||0.52 ||0.51 |
|Denmark ||0.28 ||0.28 ||0.30 ||0.24 ||0.05 ||0.04 ||0.05 ||0.05 |
|Spain ||1.24 ||1.24 ||1.30 ||1.20 ||0.23 ||0.20 ||0.23 ||0.23 |
|Finland ||0.86 ||0.96 ||0.85 ||0.82 ||0.16 ||0.15 ||0.15 ||0.16 |
|France ||3.66 ||3.70 ||4.07 ||3.68 ||0.69 ||0.59 ||0.71 ||0.72 |
|Greece ||0.17 ||0.15 ||0.20 ||0.16 ||0.03 ||0.02 ||0.03 ||0.03 |
|HongKong ||0.84 ||0.65 ||0.96 ||0.88 ||0.16 ||0.10 ||0.17 ||0.17 |
|Ireland ||0.33 ||0.35 ||0.33 ||0.30 ||0.06 ||0.06 ||0.06 ||0.06 |
|Italy ||1.44 ||1.38 ||1.45 ||1.43 ||0.27 ||0.22 ||0.25 ||0.28 |
|Japan ||7.39 ||7.58 ||7.58 ||8.90 ||1.39 ||1.21 ||1.32 ||1.71 |
|Korea ||0.67 ||0.00 ||0.00 ||0.00 ||0.13 ||- ||- ||- |
|Netherlands ||2.12 ||2.26 ||2.30 ||2.05 ||0.40 ||0.36 ||0.40 ||0.40 |
|Norway ||0.16 ||0.18 ||0.20 ||0.15 ||0.03 ||0.03 ||0.03 ||0.03 |
|NewZealand ||0.04 ||0.05 ||0.05 ||0.04 ||0.01 ||0.01 ||0.01 ||0.01 |
|Portugal ||0.15 ||0.15 ||0.19 ||0.15 ||0.03 ||0.02 ||0.03 ||0.03 |
|Sweden ||0.86 ||0.87 ||0.98 ||0.86 ||0.16 ||0.14 ||0.17 ||0.16 |
|Singapore ||0.35 ||0.31 ||0.30 ||0.26 ||0.07 ||0.05 ||0.05 ||0.05 |
|UK ||10.47 ||10.91 ||10.69 ||10.19 ||1.97 ||1.74 ||1.86 ||1.96 |
|USA ||59.40 ||59.13 ||58.38 ||59.90 ||11.18 ||9.42 ||10.17 ||11.69 |
|Total ||100.00% ||100.00% ||100.00% ||100.00% ||$18.81 ||$15.93 ||$17.42 ||$19.25 |
Sources: Salomon, FTSE, Dow Jones 12/31/01
Correlation, Volatility, and Characteristics
Despite the construction methodology differences among the indices, the bottom line risk/return attributes of each of the indices has been quite similar. However, turnover and trading cost estimate differences among the indices are more meaningful and may have practical implications for investment performance of passively managed portfolios. Since the country weightings and fundamental characteristics are quite similar, it is not surprising that the indices are highly correlated to one another and have similar diversification benefits.
Predicted tracking errors (BARRA) of each developed benchmark are as follows. (Note that South Korea and Czech Republic are excluded from the SSMB index here in order to ensure that the comparison is on a like-for-like basis.)
BARRA Predicted Tracking Errors
|Index ||Dow |
|Salomon ||FTSE |
| || || || |
|Salomon ||0.62 ||- ||- |
|FTSE ||0.48 ||1.34 ||- |
|MSCI ||1.08 ||1.45 ||0.43 |
Sources: Salomon, FTSE, Dow Jones 12/31/01
The SSMB BMI is the most comprehensive index, covering over 95% of each country’s market, and adjusting each stock’s capitalisation precisely for available float. It also has very simple and objective selection criteria - capitalisation is basically the sole determinant. The cost of this broad market coverage can be higher, though, both in terms of the number of holdings in the portfolio (8,151, each with a ticket charge) and also costs incurred in trading the EMI stocks. The only liquidity screen applied is to eliminate stocks with less than a US$75 million market cap and add stocks with a cap greater than US$100 million. Consequently, it is not unusual for some stocks to trade with a bid/ask spread of several percent, making EMI replication costly. Turnover costs can also be high if tracking the PMI only, due to the PMI/EMI annual reconstitution when turnover often approaches 25% or more.
Like the SSMB BMI, the DJGI is very comprehensive. The lower number of names (4,867) is due to the inclusion of Europe and Emerging Markets as single entities, rather than attempting to cover each individual country as SSMB do.
The FTSE All World Index offers broad market exposure (approximately 90%) but applies a variety of screens to ensure the stocks are truly investable to the institutional marketplace. One of the FTSE’s unique attributes is that policies and objectives are established by the World Index Policy Committee (“The Panel”). The Panel is composed of not only the three member firms of the consortium, but also representatives from the actual end-user groups including investment managers.
MSCI attempt to capture the lowest percentage (85%), and is the only provider to target by industry group, rather than by size, country, or region. At a detailed level, the MSCI methodology is the one most at odds with the other providers. MSCI have addressed this by recently raising the industry coverage from 60% to 85% to include stocks that they once excluded. However, notable exceptions remain, Berkshire Hathaway (also excluded by DJGI), Schering, and Colgate. Within Europe familiar names such as BMW, Credit Lyonaise, and Olivetti are excluded by MSCI.
In terms of governance, Salomon and Dow Jones are rigidly rules-based, FTSE is rules-based but overseen by a committee, and with MSCI the emphasis is on committee decisions within a code of rules. All providers review the names, shares in issue, and free float factors on a quarterly basis, except Salomon who review the names annually. With a much lower number of names FTSE and MSCI are easier to invest in and replicate, but will not give the smaller cap depth of Salomon and Dow Jones. The initial costs of replicating each index vary, though not by much for the execution (due to large cap dominance), with ticketing costs favoring indices with fewer names.
Characteristics of a good benchmark include:
1. Relevance: The index should track the target asset classes.
2. Comprehensiveness: The index should include all opportunities realistically available to investors.
3. Replicability: The index should be investable/buyable to all market participants.
4. Stability: Turnover should be kept to a minimum, and changes should be well understood and predictable.
5. Simple and objective selection criteria: There should be a clear set of rules governing security inclusion and exclusion.
Each of the four index providers meets these criteria to varying degrees.
The MSCI Indices are relevant since each country index reflects the performance of approximately 85% of the market. They are also stable, with turnover averaging less than 5% per year, but have the greater committee influence over inclusions. The FTSE Index offers a broad market and applies a variety of screens to ensure the stocks are truly investable to the institutional marketplace. The rules are clear, and changes are predictable. The SSMB and Dow Jones Indices are much broader, but in being so many smaller stocks may not be easy to invest in and may take some time to acquire. Rules are rigid, and changes are predictable.
Choosing which index to track has more to do with client preferences than any huge variations between the providers. Clients may wish to focus on the major structural differences such as the industry approach of MSCI, with a lower target percentage, or the regional grouping of Dow Jones, which gives a slightly lower UK weighting, but a deeper global coverage. Entry costs to all indices are very similar at around 30 basis points for market impact and commissions, with little additional cost from buying many more names, due to the dominance of larger cap stocks (there will, however, be a higher ticketing cost). Turnover costs will vary due to the differing rules and may also be impacted by a client decision to track only a portion of a provider’s index. The biggest difference of the past (free float vs. full float) has now disappeared, leaving the interpretation of a company’s free float as the cause of the biggest variations in weightings. When passive management emerged as a powerful international investment alternative in the 1970s, Morgan Stanley was virtually the only comprehensive index available. As a result, most index managers, including State Street Global Advisors, are able to offer the greatest liquidity and cost savings potential to investors in the MSCI Indices solely because of the total assets under management in these indices.
The information contained herein does not constitute investment advice and it should not be relied on as such. It should not be considered a solicitation to buy or an offer to sell a security. It does not take into account any investor’s particular investment objectives, strategies, tax status or investment horizon. Past performance is no guarantee of future results. The views expressed are the views of Hugh Wilson only through the period ended June 30, 2002 and are subject to change based on market and other conditions