The Case For Indexing

January 01, 1999

Barclays Global Investors commissioned a study of the index fund phenomenon. The reason for its success was clear enough: In general, index funds have worked well for investors. The study also examined and rebutted some of the major criticisms of these funds and took a look at their probable future. It sees more massive growth ahead, and not just in the US and Europe.

When the first index fund was launched 25 years ago, predicting that indexing would one day be the core of all institutional portfolios and the foundation of investment innovation might have been greeted with skepticism by the investment management community.

Few are skeptical today. Indexing's growth over the past quarter century reveals how fully accepted-indeed, fundamentally important-indexing is to the management of large and small portfolios. On the institutional side, an estimated 23% of equity mandates in the US and 20% in the UK are under index management. On the retail side, 3% of US retail equities are indexed compared with just 0.5% of that category as recently as 1994.

Despite these powerful numbers, research on indexing's impact has been limited. So Barclays Global Investors commissioned PricewaterhouseCoopers to take an independent look at indexing. The resulting report-25 Years of Indexing: An Analysis of the Costs and Benefits-found that indexing has saved US institutional investors $80 billion to $105 billion since 1973, and currently saves them an estimated $14 billion to $18 billion per year. Such savings can mean a lot to the performance of individual portfolios. An institutional investor in the US who made a tax-exempt investment of $100 million at the start of 1973 would have gained over $650 million as a result of investing in a S&P500 index fund instead of an actively managed large-capitalization fund. (The $650 million may be conservative. From 1/1/73 to 12/31/98 $100 million invested in the S&P 500 would have grown to $2.729 billion or 13.56% annually. Reducing the dollars by $650 million reduces the average annual return to 12.38% or a reduction of 1.18% -not much, given the higher investment management fees associated with active management funds and the higher turnover and therefore higher transaction costs.)


By their very nature, index funds are more economical to manage than actively managed portfolios. For a start, an index fund merely seeks to match the performance of a particular benchmark. It can do this by buying all of the securities in the benchmark in their appropriate weights or by buying a subset of benchmark securities that broadly matches the risk characteristics of the benchmark. In either event, the effort involved in creating an index fund is significantly lower than creating an active portfolio with a similar target benchmark. Consequently, the fees charged for indexing are generally much lower than for active management.

Since security selection is largely determined for an index fund manager, the key to successful index fund management is controlling transaction costs. Alarge index fund manager can substantially reduce trading costs through crossing, in which a security is traded without a bid/offer spread and the trade is executed based on a price at a specific time (frequently the market closing price). As a result, cross trades are generally executed at little or no cost. Additionally, because index fund managers trade a security to simply complete an index portfolio, their market trades are deemed to be "informationless." Consequently, index fund trades are executed within a narrower bid/offer spread than might be available to active managers, who trade on information uncovered through research.


During the period 1987-1997, large-capitalization active portfolios underperformed the S&P500 Index by an average of 1.2% per year in seven out of the 11 years. After adjusting for costs and survivorship bias of active portfolios, this difference increased to 2.3% per year. Adjusted for risk, index funds dominated a large proportion of actively managed funds (showing the same or better returns with lower volatility), although some actively managed funds secured higher returns with higher volatility.

Small-cap active managers on the whole fared much better in the study. During the period 1987-1997, active small-cap managers outperformed the Russell 2000 by an average of 2.4% per year in eight out of the 11 years. However, according to most consultant universes, small-cap active managers tend to maintain portfolios with larger average market caps than their benchmarks. Given that larger-cap stocks have generally outperformed over the 1987-1997 period, active small-cap managers who maintained portfolios with average market caps larger than that of the Russell 2000 would have been helped by this bias.

On the international front, while some data indicate that active managers have outperformed the IFCI Index of emerging markets, the results for individual countries were very different, with the median US-based active international manager under-performing by 50 basis points over a five-year horizon on a country-by-country basis.


Why do so few active managers (particularly large-cap) outperform their benchmarks? After all, a fundamental premise underlying active management is that the superior investment insight of the active manager will generate outperformance versus a relevant benchmark. The answer to the question lies in

"The Arithmetic of Active Management" by Bill Sharpe (published in the January/February 1991 edition of the Financial Analysts Journal). Sharpe essentially argues that, if indexers collectively own a slice of a given market, then active managers collectively must also own a slice of that market and therefore collectively must resemble an index fund. Furthermore, active managers trade among themselves while indexers pursue a buy-and-hold strategy. After subtracting transaction costs, the average actively managed dollar must underperform the average indexed dollar. There is one other point to note. Sharpe's argument does not require that markets be efficient for indexing to be an effective strategy. Although the effectiveness of indexing is widely accepted in developed capital markets like those in the US or UK, Sharpe's argument is universal and also holds in small, inefficient emerging markets like Zimbabwe or Sri Lanka.


Although indexing is quite popular, particularly in the US and UK, questions sometimes still arise about the potential dangers to market performance if indexing were to reach higher levels. Many of these concerns, according to the PricewaterhouseCoopers report, are unfounded:

Indexing contributes to price distortions: One argument says that indexing reduces the pool of funds available to invest in securities too small for the index, and these assets consequently will be underpriced. Conversely, assets within the index will become overpriced. This argument, however, is not substantiated. Simple competition among active managers in the market would ensure that this did not happen. Any underpricing of stocks outside the index would result in a higher dividend yield. Moreover, underpricing of small stocks outside the index is likely to result in a higher-than-average volume of takeover activity. Both of these factors should ensure outperformance for stockholders, attracting an inflow of investment from active managers, which will offset any underpricing. Additionally, empirical evidence does not support the assertion that stocks outperform after they have entered the index. In the UK, for example, stocks tend to underperform after promotion to the FTSE 100 and to outperform after they are demoted.

Indexing hurts research quality: Some also argue that indexing contributes to price distortions because it reduces the level of research and analysis of the market. Active managers conduct expensive research to support their investment choices, while indexers make no such effort. Ashift away from active to indexed management could reduce research and therefore contribute to price distortions. For example, under-researched stocks could have unjustifiably inflated prices simply because they are in the index. However, it is also equally possible that as indexing grows, active managers will increase, not decrease, their research because the pressure would be even greater to beat the index. Moreover, it is difficult to prove that the current level of research is efficient or contributes to market efficiency. Indeed, the market might currently be over-researched. Indexing is also unlikely to ever grow so much that a serious deficit of research would emerge to undermine pricing efficiency.

Indexing could create market instability: Another argument holds that if a large number of investors choose indexing, the market will be driven by the allocation decisions of just a few active managers. The more highly correlated the trading of these managers, the more volatile the market. There is an implicit assumption in this argument that if a larger percentage of the market were actively managed then active managers by their differing strategies would act as a stabilizing force on the markets. (Therefore reducing the pool of actively managed assets serves to increase market volatility). There is no evidence to support this proposition. In fact, indexing weighs against instability by promoting an approach to investing that is broadly buy and hold.

There is also an argument that to the extent index fund managers do indeed buy and hold, their investment strategy might reduce market liquidity. Assuming this is true, the reduction in liquidity does not manifest itself exclusively during volatile market periods but exists at all times. This is a big (and not very accurate) assumption. Although an index fund portfolio does not normally trade very much for index changes, compared to an actively managed portfolio, clients contribute and withdraw monies from index portfolios all of the time, and index fund managers need to buy/sell index fund slices to accommodate this client activity. Additionally, index fund managers when rebalancing portfolios frequently trade patiently over time looking to provide liquidity to the market rather than demanding it.

Indexing could cause markets to collapse: This concern relates to the notion that the entire market cannot be indexed because it would lead to "locked" assets and prices. The study, however, said "the proportion of funds indexed is unlikely to reach a level that could threaten market collapse." Indeed, "system stabilizers" are at work, and as indexing grows, an equilibrium eventually will develop between indexed and active management. Why is this? "The growth of indexing is likely to increase the net value of market research (thereby raising the incentive to obtain private information)," the study said. "This is likely to increase the return on active management before costs, increasing the proportion of funds under active management. Over time, the percentage of indexed funds would reach a ceiling."

Indexing impacts adversely on corporate governance: There is a perception that, because they are only trying to match the index and not do in-depth research on the companies they hold, indexers do not care about corporate governance. However, today index fund managers are taking an active role in corporate governance, and have been at the forefront in advocating shareholding rights and systematically voting proxies (and tendering shares) to maximize value for their investors. Active managers can impose the ultimate penalty against corporate management by selling the stock. Indexers do not have that option, which is why corporate governance is important to them.

Indexing underperforms in bearmarkets: It is sometimes argued that active managers perform consistently better in bear markets because index funds have no choice but to follow the market downward, while active managers can shift assets to cash or other defensive investments. However, at best the evidence to support this claim is mixed. Including the market downturn in the 3rd quarter of 1998, the average active manager beat the index in only one of the three most recent bear markets. In fact, the very poor relative performance of active managers during the bear market of 1973-74 was a key impetus to the development of indexing. There is no empirical evidence to substantiate the claim that active managers do better in down markets and no logical reason to expect that they would. Additionally, even if active managers did outperform in down markets, what investor needs a strategy that will only work when the market is down - less than 1/3 of the time?


Where does indexing go from here? Has the level of indexed assets topped out or could-indeed, should-it go even higher? How will indexing affect investment options and strategies for the next 25 years? The study's findings and current trends suggest that not only is there room for pure indexing to grow, but indexing will likely be the foundation on which investment strategies of the future will be based. The popularity of indexing is growing at the retail level. Additionally, most indexed dollars, even 25 years after the start, are still concentrated in relatively few benchmarks. There are many additional categories of indexing to tap.

Though indexing theoretically could take as much as 70% or 80% of the market without having a negative impact, investors should really focus on the continuum between indexing and active management. As investors recognize that the ends of the continuum are limiting, they will examine the various quantitative active and enhanced indexing strategies that lie between either end. Evidence of this is being seen slowly as traditional active managers apply quantitative techniques to manage their portfolios.


The PricewaterhouseCoopers study identified three immediate areas where indexing will likely grow. First, European Economic and Monetary Union eliminates exchange-rate risk and thus will lead to the emergence of pan-European benchmarks. A more integrated European equity market is likely to lead to an increased use of indexing as a company's country of domicile becomes much less relevant. Additionally, as country exposure becomes less relevant, industry exposure becomes more relevant and a number of indexed funds based on pan-European industries may appear.

Government sponsorship of indexing may also drive future growth. In several countries, governments are investigating the option of compulsory pension provision to supplement state and occupational pensions. Index funds have great appeal to such plans because of their straightforward and low-cost nature. Additionally, index funds have the added benefit that they produce market-like returns, which mitigates some of the risk associated with underperformance. (Governments will still have to educate pension participants as to appropriate asset allocations, but at least will not run the risk that active management may compound any shortfall.)


By now, most institutional investors are aware of the conflicting information circulating about index versus active management. On one hand, active managers continue to promote their funds as outperforming their indices. However, studies have consistently proven that few active managers beat their index and the PricewaterhouseCoopers study merely reaffirmed this conclusion.

One key obstacle to the PricewaterhouseCoopers research was obtaining accurate information about actively managed funds. Specifically, active managers have little incentive to deliver information about underperforming funds or funds that failed. This adds to "survivorship bias" in the available data for actively managed funds-because the universe of funds incorporates only surviving funds, actively managed funds (as a group) are able to post artificially better results than if full disclosure had been the rule.

(Survivorship bias is the statistical bias that results from examining only investment managers that "survive" long enough to figure in a data series. In this context, only investment managers who are still in business at the end of any period would be included in the manager universe. Managers who have gone out of business, presumably because of poor performance, are excluded, thus likely overstating the performance of the managers in the universe.)

Estimating survivorship bias is a daunting problem, particularly for the small-capitalization market, which was also examined in the study. The small-cap market has higher volatility with more funds coming into and dropping out of the market, making it more difficult to analyze than the large-capitalization universe. This survivorship bias leads to the creation of misleading universes which active managers can then use to their advantage. If there were greater transparency in the data, it would make it easier for investors to compare performance based on facts.


Despite these issues, the study has validated indexing as a core position in any institutional investment fund. Indexing has led to more competitive pricing and driven most active managers toward better risk management. Additionally, indexing forces investors to look at the value of active managers and make the necessary tradeoff between the incremental cost of active investing and the risk of outperformance.

The study concludes with the notion that "eventually, an equilibrium between indexed and active management seems likely, with indexed and active funds increasingly being seen as complementary." In other words, the old "active versus passive" debate is no longer an "either-or" proposition.


To arrive at these numbers, various assumptions were used. First, costs and benefits were defined as those for the investor in using return and expense data (so-called private costs and benefits). Second, the study compared a universe of actively managed funds against a comparable universe of indexed funds, rather than against an index. This provides a proper like-for-like comparison and captures the tracking error of indexed funds. In making this comparison, the study attempted to ensure that the appropriate benchmark index was used for both the active and indexed funds.

The study did not distinguish between the different indexing methodologies, such as full replication and optimized sampling. Given advances in quantitative techniques in the past few years, the differences between the methodologies are getting smaller in terms of returns and turnover. In comparing actively managed and indexed funds, cash holdings of the actively managed funds were not deducted. This reduces the average performance of the actively managed funds when stock markets are performing well. However, it is a valid comparison that replicates the real-life choice in deciding whether to award a mandate to an active or indexed manager.


Following a careful review of the available data sources, the study used Lipper Analytical Services and Callan Associates for the US, and the WM Company, CAPS and Micropal for the UK. For the US large-cap institutional sector, performance results were cross-checked against those available from Callan Associates and Frank Russell to ensure that the results are not dependent on the precise database used.

For the purposes of the study, the US data from Lipper was constructed in accordance with AIMR guidelines. Returns were gross and equally weighted; the average returns of the active and indexed universes were compared. For the performance measures, a universe of equity growth and income funds was used; for the more detailed work on stock turnover, expense ratios, and risk and return, the study concentrated on the equity growth universe where Lipper Analytic's coverage is more comprehensive. The UK CAPS and WM databases calculate median (rather than average) returns, but do so on an equally weighted basis.


Find your next ETF

Reset All