Product and Strategy Notes: SSgA Passes Fidelity, Research - Active Management, Insufficent Saving for Retirement, Dismal Long-Term Performance, Two Canadian Commodity Index Products

A Turning Point in the Money Management Industry

Institutional Investor magazine has just published its ranking of the largest money managers in the United States, based on their assets at the end of 2003. At the top of the list in this year's ranking was State Street Global Advisors, which surpassed Fidelity as America's largest asset manager. However, as Institutional Investor notes, "the ascension of SSgA marks a symbolic turning point in money management -- the first time that an "indexer" has emerged as the largest U.S. asset manager. Underscoring the point, another indexer, Barclays Global Investors, places second [on the list]." We also note that two other big indexers -- Vanguard and Northern Trust -- respectively rank ninth and twelfth on this year's list. However, there is both more and less in these rankings than meets the eye. SSgA and BGI have moved up in the rankings not only because of the growing popularity of their index offerings, but also because of the money that has flowed into their actively managed products. More specifically, both firms have been strong proponents of a key trend in asset management: the separation of beta and alpha risk.

As readers will recall from previous articles on this subject, beta risk is inherent in an asset class as a whole. The only way to reduce a portfolio's beta risk is to diversify across asset classes. In contrast, alpha risk is specific to a company or group of similar companies (e.g., in an industry sector). Within an asset class, the returns for bearing alpha risk are a zero sum game. Within any time periods, the positive and negative returns for bearing alpha risk cancel each other out, leaving only the return for bearing the beta risk of the asset class.

In a typical "long only" actively managed fund, return is due to a combination of overall market movements (compensation for holding beta risk) and the returns on the specific securities held by the fund (compensation for holding alpha risk). However, the expenses charged by an actively managed long-only fund are typically much higher than those charged by an index fund, which bears only beta risk. This has caused many institutional investors to look for ways to separate the decision to bear beta risk from the decision to bear alpha risk. A practical example of this would be a portfolio that combines low cost index funds for different asset classes (to provide returns for bearing beta risk) with smaller investments in a number of much more expensive hedge funds (e.g., a market neutral fund) whose returns are compensation for bearing only alpha risk.

Both SSgA and BGI have aggressively seized on this trend, and their asset growth (and improvements in the Institutional Investor rankings) reflect not only the success of their index fund offerings, but also the growing attractiveness of their hedge fund and enhanced index products.

However, the growth of strategies that separate beta and alpha has taken place in an environment in which indexing has continued to gain market share. Institutional Investor cites a study by the consulting firm Greenwich Associates that estimated that as of June, 2003, "forty percent of U.S. institutional equity assets were indexes, up from thirty percent [ten years earlier]." Comparative numbers for the U.K. were 31% versus 22%. All in all, we find this to be very encouraging data.

Independent Research: It's Still Active Management

As you recall, the settlement of charges related to conflicts of interest in equity research reached with the U.S. Securities and Exchange Commission by various Wall Street investment banks and brokerage firms required them to spend over $1 billion dollars to support the production and distribution of independent investment research. The basis of the settlement was that independent research would help active investors to make better decisions. In its July 26, 2004 issue, Business Week magazine tested this hypothesis by comparing the performance so far this year of 121 independent research providers. They found that only three of them -- that's about 2.5% -- had managed to beat the return delivered by the S&P 500 Index. While independent research may be an improvement over its predecessor, indexing still seems the more prudent choice for most investors.

How Effectively Are People Preparing for Retirement?

Ask yourself two questions. First, how well do you think you have prepared for your own retirement? Second, how well do you think most people have prepared for their retirements? We would guess that most of our readers would say they’ve done a better job than the majority of their peers. A new research report from the Principal Group, "The 2004 Global Financial Well-Being Study" provides more data on this issue.

The survey is based on at least five hundred interviews in each of twelve different countries, and was conducted earlier this year. At the 95% confidence level, the margin of error for the studies findings is approximately four percent (e.g., if a finding is that 25% of the people surveyed believes X, you can be 95% confident that the true value of X for the population as a whole is between 21% and 29%).

The study’s authors recognize that its key findings are a paradox: "although people worldwide say they are concerned about their financial future and recognize that putting money aside for retirement is increasingly important, a surprisingly small number have done anything about it. Many have not even begun planning for retirement…Few have tried to figure out how much money they will need to retire and how many years their nest egg will have to last." It is therefore not surprising that the study also found " a deep-seated pessimism about retirement prospects in almost every country. People are concerned their standard of living will be lower after retirement and are afraid they may become a burden to their families. Even more alarming, a significant number of [those surveyed] are concerned that they won’t be able to pay for basic expenses after retirement."

Two survey questions asked (1) how well people felt they were doing "in performing your role to ensure that you have a financially secure retirement", and (2) whether they have tried to estimate the amount of savings they would need to accumulate to live comfortably in retirement.

A March 2004 survey of 2,000 Canadians by SEI Investments produced similar findings in that country. They found that only 15% of those surveyed considered themselves to be "ver knowledgeable" about retirement planning, while 34% reported that they "did not feel knowledgeable at all." Moreover, a separate question found that only 30% of those surveyed "considered themselves to be actively involved in their retirement planning."

Could these findings be explained by a widespread belief that benefit payments received from national social security programs will provide an adequate standard of living in the future? Another question on the Principal Group survey addressed this issue. It asked people to rate their confidence that their national social security programs would continue to provide the same level of benefits in the future as it does today. The percentages of people saying they were very or somewhat confident were actually quite low: Japan, 18%; France, 28%; Germany, 26%; Italy, 31%; United Kingdom, 19%; and United States, 35%.

Given that they haven’t planned and don’t have much confidence in their national social security programs, it is no surprise that people are concerned about the quality of life they will have in their retirement. Another survey question asked people whether they thought their standard of living in retirement would be better, worse, or about the same as it is now. The percentages answering worse were as follows: Japan, 53%, France, 62%, Germany, 44%, Italy, 40%, the United Kingdom, 34%, and United States, 49%.

Another way of looking at this issue is the percentage of people who reported that they feel "very confident" they will have enough money to take care of their basic expenses during retirement. In Japan, only 3% gave this answer; in France, 13%, Germany, 25%, Italy, 8%, the U.K., 28% and the United States 24%

These findings are both depressing and provocative. Clearly, they indicate a crisis is brewing, and that it will become increasingly acute as more baby boomers reach retirement and confront the stark reality of their insufficient savings. But they also raise an intriguing question. How is it that so many people can still report a belief that they are doing a good job of preparing for their retirements when their answers to other questions suggest that this objectively is not the case? At minimum, it suggests that many people have a much more short-term perspective than is assumed by most economic theories (technically, they have much higher time discount rates). This would also explain why people are now much more willing to use borrowing to finance consumption than they were in the past. It might also explain why so many people prefer active investment management, and tend to "chase performance" by investing in those funds with the most impressive recent returns. However, the even more interesting question is why this is the case. What economic or social changes can account for the shortened time horizons and preference for current consumption over long-term security that we observe across such a wide range of countries? We don’t pretend to have a definitive answer to this question. However, we suspect that it is one that future historians will spend a lot of time trying to answer.

Another Dismal Report on Long-Term Performance

Ilia Dichev of the University of Michigan Business School has just published a fascinating paper titled "What are Stock Investors’ Actual Historical Returns?" It does not make for pleasant reading, particularly if you are believer in active management. The study begins with an observation: "stock investors’ returns are determined by two factors: the returns on the securities they hold, and the timing of their capital flows into and out of these securities." Dichev notes that many studies "typically use buy-and-hold returns to assess the return experience of stock investors, essentially ignoring the effects of capital flow timing." Dichev’s study attempts to take capital flows into account, to provide a more accurate estimate of investors’ returns.

The way he does this is to use an internal rate of return methodology to calculate the actual return experienced by investors at the level of the aggregate stock market. He treats initial public offerings as cash outflows (from investors to the aggregate equity market) and dividends and repurchases as cash inflows (from the market to investors). Dichev compares this IRR to a value weighted buy-and-hold return (i.e., each year the proportion of each share held is adjusted for its relative market value). This buy-and-hold return is, as the author notes, the return that would be achieved by an investor who simply bought an held a broad-based equity index fund over the time period covered by his study.

As Dichev notes, there is no inherent reason why buy-and-hold and IRR returns should differ. This only occurs "if there are material correlations between the timing of [investor cash outflows and inflows] and past and future stock returns." In particular, the IRR-based return will be lower than the "buy-and-hold" return if cash outflows (i.e., net investor purchases of equity) are followed by lower than average equity returns, while net inflows to investors are followed by higher than average equity returns. This is exactly the phenomenon that Dichev discovers in his data. He finds that "dollar-weighted [IRR] returns are lower than buy-and-hold returns for both the NYSE/AMEX and Nasdaq firms, suggesting that in the aggregate actual U.S. investor returns are lower than buy-and-hold returns…The difference between [IRR] and buy-and-hold returns is (1.3%) for NYSE/AMEX firms…The truly remarkable evidence is for the Nasdaq, where the difference is (5.3%)." Since the buy-and-hold return for the Nasdaq was roughly equal to that for the NYSE/AMEX, Dichev notes that "one could reach radically different conclusions about investors’ performance depending on the return metric used. If one uses the conventional yardstick of buy-and-hold returns, one would conclude that Nasdaq investors have performed roughly in line with the overall market. If one uses dollar-weighted [IRR] returns, one would conclude that investing in the Nasdaq has been disappointing, yielding returns in line with the risk-free rate, but with much higher volatility."

The author also finds evidence that "investors tend to move into the market after high past returns and before low future returns." He notes that in the United States, aggregate cash inflows and outflows "show long-swing patterns over the years, with [cash inflows] dominating the time series, but also with three clusters of frequent [cash outflows] in the 1920’s, 1968-1972, and 1990-2002." He notes that "these clusters tend to coincide with strong bull markets, and are followed by the worst bear markets in the century, during the Great Depression, 1973-1974, and 2000-2002."

Dichev also tests his U.S. results using shorter data series from non-U.S. markets. While his NYSE data series covers 1926-2002, his international series are shorter, and begin only in 1973. He finds that "dollar weighted returns are lower than buy-and-hold returns in 18 out of 19 countries; the only exception is Canada, where the difference is not statistically significant." The following table presents the results for some of these countries:

Finally, based on the evidence he presents, Dichev concludes that "passive investment strategies [i.e., indexing] are likely to do well because they avoid both transaction costs and the negative effects of timing."

Two Interesting Canadian "Commodity Index" Products

In December 2003, SEI Investments Canada introduced a retail product (Class P Shares) of its Futures Index Fund, which had previously only been available to institutional investors (via Class O shares, which have a minimum investment of C$ 150,000). Per its prospectus, the annual expenses charged on the fund can be no higher than 1.50%. These include both investment management and operating expenses. By comparison, on the Class O shares investment management and operating expenses are charged separately. The former is negotiated with each investor, while the latter was a flat 0.27% in 2003.

The Futures Index Fund tracks the Mount Lucas Management (MLM) commodity index, which is quite an interesting benchmark. First, it is based on 25 equally-weighted futures contracts. These include not only the usual suspects (e.g., metals, energies, foodstuffs), but also U.S. Treasury securities (5, 10 and 20 year maturities) and currencies (A$, C$, Euro, Yen, UK Pounds and Swiss Francs). Second, it also includes a momentum-based trading rule: the index is long contracts that are above their 12 month moving average, and short those below the 12 month moving average. Because of this, it is another one of those odd ducks we wrote about last month in our article on "active indexing" -- in effect, the "index" it tracks actually reflects the results of a consistently (i.e., mechanically) applied active management strategy. In terms of its historical returns, the MLM Index has a low correlation with returns on the Goldman Sachs Commodities Index (a long-only index that we used as our proxy for the commodities asset class when developing our model portfolios). Given this, we consider the MLM Index to be more accurately classified as a subset of the larger class of hedge fund indexes, rather than as a commodities asset class index. There is nothing inherently wrong with this classification; our only objection is the potential confusion caused by calling this product a commodity index fund.

Year-to-date through June 2004, the Futures Index Fund had returns of 0.89%, versus a 1.73% return on the MLM Index. By way of comparison, the institutional (Class O) shares of the fund returned 1.70% over the same period. Presumably, the difference in returns is due to the difference in expenses; however, the fund's reports don't make this clear.

Finally, SEI Investments also offers another interesting fund based on the MLM Index. While thus far it is only available to institutional investors, it is also worthy of note. The fund is called the 3XL Futures Index Fund, and it uses debt (in addition to investors' money) to try to deliver a return that is equal to three times the return on the index. In short, it is a momentum trading strategy to which leverage has been added to make things more interesting, and hopefully more lucrative. Through the end of June 2004, the 3XL fund had returned 2.16% -- about 125% of the index return.

Three Follow Ups From Last Month's Articles

An alert reader wrote to note that www.scottrade.com offers online trades for U.S. $7.00, even less than the $10.99 fee at Ameritrade we used in our mutual funds versus exchange traded funds analysis. At the margin, this makes exchange traded funds more attractive.

Another reader wrote with a question about last month's article on the advantages and disadvantages of taking tilts toward small cap stocks. "Is it not true," he asked, "that what counts is expected portfolio risk and return? And, if this is the case, might not a tilt toward small caps still make sense in light of allocations to other asset classes?" We responded that he is right on both counts. However, in writing our article we did not include this additional level of complexity. To do so, we would have had to run our simulation optimization model substituting the risk, return, and correlations for small cap stocks with other asset classes in place of the broad asset class statistics we normally use. The challenge here is how to limit the possible permutations of this process (which, as you can well imagine, give rise to substantial computational challenges). For example, should we also include small cap data for foreign developed market and emerging market equities? Should we include other tilts (e.g., toward mortgage or high yield bonds, or towards one type of commercial property security)? Should we also include domestic equity sector indexes, some of which (e.g., utilities) have very low correlations with small cap stock returns? For better or worse, there is not right or wrong answer to this question, apart from the caution that including highly correlated tilts (e.g., large and small cap stocks) as investment options usually gives rise to very unstable optimization solutions (i.e., situations in which a very small change in expected return has a drastic impact on the portfolio weights given to different asset classes).

Finally, a reader from South Africa wrote "is it not the case that a total market index (or index fund) like the Wilshire 5000 is not 'neutral' but is, in fact, tilted toward large cap companies relative to the total market of listed and unlisted stocks? If this conclusion is correct, wouldn't it follow that a 'neutral' investor should invest in the Wilshire 5000 and additionally into a mid-cap and a small cap index fund?" This is a very interesting question. Our first step in addressing it was to look at flow of funds data from the Federal Reserve. The most recent report (statistical release z.1, table L.213) shows that at the end of the first quarter of 2001, the estimated value of the equity issued by U.S. companies (both private and public, financial as well as non-financial) was $13.7 trillion (using a methodology based on the replacement cost of assets). We then compared it to the end of June, 2004 market value of the 5,079 companies included in the Wilshire 5000 Index: $12.7 trillion. At first glance, this comparison would suggest that there is roughly (because of the different dates, and the methodology differences used to calculate market value) about $1 trillion in equity market value not captured by the Wilshire 5,000.

Our next step was to develop a better picture of the key differences between the companies included in the Wilshire and Federal Reserve data. We looked at two key indicators. The overall market/book ratio for the Wilshire companies was 2.86, and 1.67 for those that make up the microcap portion of the overall index (these are the companies whose market capitalization rank is less than 2,500). The market/book ratio in the Federal Reserve data was 1.34. The annualized dividend yield (dividends/market value) for the Wilshire 5000 was 1.61%, and just 0.58% for the microcap sub-index. For the Federal Reserve data it was 2.65%. Based on this evidence, we reach our first conclusion: in so far as "value" companies are characterized by relatively low market/book ratios and relatively high dividend yields, the data suggest that many companies not included in the Wilshire 5000 are probably in this category. Logically, this makes sense: if the market is not going accord your company a high valuation, why go public

To get a more detailed picture of the companies that are not included in the Wilshire 5000, we looked at 2001 corporate tax return data (the most recent available) from the United States Internal Revenue Service. The following table summarizes this data:

After examining this data, our best guess is that a substantial portion of the equity value included in the Federal Reserve data but not the Wilshire 5000 is in wholesale and retail trade, and different types of services. Moreover, a comparison across the three data columns (percent of returns, assets, and net worth) suggests that within these industry sectors, this equity (net worth) value is distributed across a large number of small, relatively unleveraged (i.e., high net worth/total assets) companies. If many of these companies are paying out to their owners a relatively high proportion of their earnings, and growing at a relatively slow rate (or not at all), it would account for both the higher dividend yield and the lower market/book ratio of the Federal Reserve relative to the Wilshire 5000 data.

In sum, while there are no doubt very attractive private companies in which one might like to invest, we infer from the data that most of them are not in this category. Given this, we do not believe it makes sense to tilt a portfolio's U.S. equity allocation toward small cap public equities in order to better represent the true distribution of companies in the overall economy.

| The Summer Financial Humor Column: "Irreconcilable Differences" | Global Asset Class Returns | Product and Strategy Notes: SSgA Passes Fidelity, Research - Active Management, Insufficent Saving for Retirement, Dismal Long-Term Performance, Two Canadian Commodity Index Products | Equity Market Valuation Update | Who Is the Real Fool? | Model Portfolio Update | This Month's Letters to the Editor: Equal vs Market Capitalization Based Index Weighting |