[Back to FundAlarm Links page]

Some Thoughts on Indexing and Performance
[by Bill Miller, manager of Legg Mason Value Trust; excerpted from the fund's quarterly report, dated December 31, 1998]


Most money managers again failed to keep pace with the benchmark S&P 500,which outperformed over 80% of active managers in 1998 and over 95% of activemanagers over the past five years. Last year was a particularly hard year tobeat the market, since only 28% of the stocks in the S&P outperformed the index,and those were concentrated mostly in the largest names. As evidenced by thevery poor returns of the Russell, to the extent active managers focused theirefforts on smaller companies, by charge or by choice, they had very littlechance to deliver index-beating returns.

The S&P 500 is the benchmark for most active managers; it is thecompetition.1 It is also the cost of capital for active management. Just ascompanies that invest above their cost of capital--which is the weighted averagecost of their debt and equity--add value for owners and those that do notdestroy shareholder value, so too with money managers. To the extent moneymanagers outperform the index, they are adding value for their shareholders; tothe extent that they underperform, they are destroying value relative to returnsthat could be obtained by investing in an index fund. The rapid growth ofindexing over the past decade or so is a direct result of the inability of mostactive managers to add value relative to what is misleadingly referred to aspassive investing.

Money managers are often quick to criticize companies that do not performwell, and to offer management advice on how to do better. Companies that areconsistently beaten by the competition are rightly admonished and urged to adoptstrategies with a greater probability of success. Those companies that areconsistently superior are closely studied and their methods are often copied inorder to improve results.

Surprisingly, scant attention appears to be directed at the methods andstrategies of the S&P 500 despite its long-term record of superiority relativeto most active money managers. Data compiled by Tweedy Browne & Co. indicatesthat since the beginning of l982, the S&P 500 has outperformed over 90% of allsurviving mutual funds. The proportion would of course be even higher if fundsthat subsequently failed or were merged were included. A recent book, Indexingfor Maximum Investment Results2, contains the results of numerous studies ofactive managers vs. index performance. The results are consistent across a wide variety of time frames: index fundsroutinely beat the overwhelming majority of active managers. Index funds doparticularly well relative to active managers in large capitalization stockssuch as those found in the S&P 500. Even with very broad market indices, such asthe Wilshire 5000, the superiority of indexing holds. From 1973 through l995,there were only 8 years that active managers beat that index, and 6 of themoccurred prior to l983.

Two questions suggest themselves: why does indexing work, and can activemanagers improve their results by careful study of the competition?

The answer to the first question is easy: indexing works because the marketis efficient. Academics refer to the notion of market efficiency as ahypothesis, but they are just being overly fastidious. We can dispense withtrying to distinguish, as they do, between various forms of efficiency, and withan unnecessary discussion of the various studies of possible market anomaliesand inefficiencies. For the purposes of real world investing, let's call amarket "pragmatically efficient" if it beats most active managers most of thetime. It is clear that the market for large capitalization US stocks ispragmatically efficient.

An efficient market is one where stock prices reflect all availableinformation. Notice that this description does not say prices are correct, justthat in the aggregate they incorporate what is believed about the present valueof the future. The information available to the market is reflected in pricesthrough the collective behavior of the agents operating in it. These agentsencompass an ecology of objectives, time horizons, risk tolerances, and utilityfunctions. This leaves plenty of room for recurrent, exploitable anomalies suchas those that may arise due to systemic errors in probability assessment. Butany such opportunities are unlikely to be available to the average investor.Average hitters can't win many batting titles.

The most important implication of a market that is efficient is that there isno algorithmic or systematic way to outperform it. More simply, there is noformula which can specify a portfolio that is likely, over time, to outperformthe market. Any such formula or mechanism would be copied and its advantagearbitraged away as it quickly spread among investors. There may be formulationsor principles used by those who have outperformed, but such statements will notspecify a portfolio. We use just such formulations: we buy companies that tradeat large discounts to intrinsic value. Unfortunately, that simple descriptiondoesn't specify any particular portfolio. (When asked the secret to hitting,Stan Musial--I think--said, `wait for a good pitch, and then smash it.')

This means that all of the popular methods of stock selection: value, growth,momentum, sector rotation, thematic investing, etc. are either too vague to beuseful, or if specific, as many purely quantitative methods are, likely to havea short half-life.

Is active management just a loser's game, and passive indexing the onlyviable investment strategy? Not at all. Just as poorly performing companies canoften improve their results by studying winning competitors, active managers canimprove theirs by deconstructing the sources of competitive advantage of theindex, in our case the S&P 500.

The first and most important feature of the S&P 500 is that it does notemploy a passive selection strategy. Managers of index funds employ such astrategy, since they engage in no active stock selection. But the S&P 500 is anactively selected index. Its stocks are chosen from the nearly 9000 publiclytraded securities on the New York, American and NASDAQ markets by a committeeusing specific investment criteria.

The returns of the S&P 500 are the best evidence of the long-term advantageof active portfolio construction. It has consistently beaten broader, passivelyconstructed indices such as the Wilshire 5000, the Russell 2000, and the NYSEComposite. That it has also beaten other active money managers is not anargument against active management, it is an argument against the methodsemployed by most active managers.

The S&P 500 is a long-term oriented, low turnover index employing a buy andhold strategy, which is by nature tax efficient. It lets its winners run, andselectively eliminates its losers. It never reduces a successful investment nomatter how far up the stock has run, and it does not arbitrarily impose size orposition limits on holdings, either by company or industry. Size is fixed at 500names, and new names usually come into the index because of the merger oracquisition of existing companies. Periodically, new names are added and otherseliminated in an attempt to replace companies that are marginal with those whoseposition in the economy or an industry are deemed more important.

The overall index is positioned to represent the broad sweep of the USeconomy. The stock selection committee at S&P consists of 9 analysts, and newnames are meant to be seasoned companies with a history of profitability,financially sound, leaders in their industry or market, with a probability ofbeing in business over the next 10 to 20 years. Index turnover averaged 25 to 30names in the l980s but has picked up to 40 or so in the past few years.

The contrast with the typical active manager is stark. The average mutualfund is short-term oriented, has high turnover, is tax inefficient, and employsa trading oriented investment style. Most funds systematically cut back winnersor rotate out of stocks that have done well into those expected to do better.Position limits and industry weightings are usually rigidly maintained in thename of either investing discipline or risk control. The number of holdings isusually well over 100, but may vary significantly, both among funds and withinthe same fund over time. The overall portfolio is constructed in accordance withsome style the manager erroneously believes is likely to outperform thelong-term, low turnover approach of the S&P 500.

Part of the success of the Value Trust is no doubt due to our not employingmethods that place us at a disadvantage to the S&P. We employ a long-term, lowturnover, buy and hold investment approach that as a by-product is taxefficient. We employ no rigid industry, sector, or position limits, save onlythose mandated by the broad mutual fund rules. We let our winners run andeliminate those positions deemed to have poor long-term economic prospects. Ifwe have companies merged for cash, we will redeploy that capital into new names.

Operating similarly to the S&P in the ways noted above does not help usoutperform the index, although it probably has helped our results compared toother managers whose methods have been different and suboptimal relative to theindex. Our results have improved in the past several years as our turnover ratehas fallen and as we have let our winners continue to run without cuttingpositions back in the guise of risk control or portfolio balance. There is alimit to all good things, though, and our positions in Dell and AOL are muchlarger relative to our portfolio than the largest position in the S&P isrelative to that index. Those positions are at or near their peaks and we wouldexpect them to decline gradually over time.

Since the selection criteria for inclusion in the S&P 500 are not terriblyarcane, it is not superior stock selection that has led to its methods beatingmost other active management styles. There are services out there that arepretty good at identifying likely candidates for the index. America Online waswidely expected to be added, as happened at December 31, and Warren Buffett'sBerkshire Hathaway is another favorite for eventual inclusion.

What is the source of the S&P's superiority over active management styles? Ithink it comes back to market efficiency. The market is pretty well aware of theinformation that may affect the prospects of its constituent companies for thenext year or so. New information arrives unpredictably. Beyond the next year,the complexities, uncertainties, and vagaries are such that no one is vouchsafedany special insight into the future. Most of the activity that makes activeportfolio management active is wasted; it adds no value since it is engenderedby the mistaken belief that the manager possesses information the market isunaware of, or that the market has mispriced. It does impose costs: tradingcosts, market impact costs, and taxes. It is also often triggered by ineffectivepsychological responses such as overweighting recent data, anchoring onirrelevant criteria, and a whole host of other less than optimal decisionprocedures currently being investigated by cognitive psychologists.

Money managers often wrongly decry the growth of indexing and complain thatit is a mindless strategy. While they may be right that pure factor basedindexing strategies, e.g., buy all companies with characteristics xyz, areunlikely to add value over time (the evidence is not clear on this), they aresurely wrong to bemoan either the desire of investors for S&P 500 index funds ortheir own inability to compete with the results of such funds. Investors arerationally selecting an active money management style that is sensible, taxefficient, has a long history, and works.

___________________
1 It is not possible to invest directly in the S&P 500.
2 Neubert Alberts, Editor