August Newsletter: Have we reached the saturation point of the indexing strategy?

Posted by Alex Frey (@alexhfrey )

It's a common aphorism that beginning investors spend most of their time thinking about why they might be right, while world-class investors spend the bulk of theirs imagining why they could be totally wrong. 

Since our approach rests on a foundation of automated passive investing and index funds, a common thread we like to periodically return to is: under what conditions might indexing cease working?

This question takes on added importance because the popularity of indexing has continued to explode since this site was launched. It is well-known that many financial strategies have a strange property where they stop working as soon as some critical mass of people start to believe in them – could indexing be the same?

In addressing this, we'll need to review some material that may be familiar to many readers – we'll start with why indexing has become so popular in the first place. 

Index funds are guaranteed to give returns equal to the market average before fees, which are usually very small, while the average actively managed fund is mathematically nearly guaranteed to do worse. There is one small caveat to this statement that we will discuss below. For active managers who operate in the same arena (i.e., choose stocks from the same market), the average asset weighted return must also equal the market average before fees. But the active manager normally charges much higher fees than the passive (index) manager, so the average active manager will have lower returns than an index fund after fees are deducted. 

There are some caveats here that can cause confusion.

If the active and passive managers are choosing from the same fund arena, and if all active managers operating in that arena are included, the conclusion stated above is a mathematical fact. However, if the active and passive managers pick from different stock arenas, the conclusion as stated above doesn’t necessarily hold. This may seem obvious, but we have seen discussions of this issue where active “apples” were compared to passive “oranges”. So for instance, if the active managers are choosing from the arena of all world stocks, and the passive managers are using only U.S. stocks, then it is possible for the average active manager to outperform or underperform the passive manager depending on whether foreign stocks outperform or underperform domestic stocks. Of course passive investors have the option of adding an international index fund to their portfolio to overcome this problem.    

Also, we must acknowledge that all active managers are not mutual funds. The average of all active managers in the market must equal the market average before fees. However, in a subset of the active managers, for instance active mutual fund managers, it is possible that the average manager in that subset will exceed the performance of the market average. This could happen if the average mutual fund manager is a better stock picker than the other active investors in the arena (pension funds, hedge funds, foundations, individuals, etc.).

So what does the data show? 

As an example of the data available, Figure 1 from Morningstar was included in our November 15 newsletter. For funds operating in various Morningstar categories, the chart compares the performance of actively managed funds to an average passively managed (index) fund operating in the same Morningstar fund category. The performance of the average index fund was obtained by averaging the performance of real index funds operating in that category and includes their management fees. The numbers given in the chart are the percentage of actively managed funds that beat the average index fund over the time period indicated. For the 10 year period, the results for the actively managed funds are reported separately for the lowest and highest cost quartile of funds. From the numbers, it is clear that relatively few active funds are beating the index funds regardless of the time period or the fund category. It’s also clear that the lower cost active funds do much better than the higher cost active funds.

Aug 1

Figure 1: The percentage of actively managed funds that beat an average index fund operating in the same fund category and over the same period of time. Data from Morningstar.

Why is the data so unequivocal? The basic principle is that the active managers, by constantly researching stocks and buying and selling according to their perception of value, keep prices at a “correct” level given the perceptions that investors have at any given time. The passive (index) investors get to take a free ride on all the work done by the active investors. 

Nothing that has just been said implies that some active managers might not have the talent to outperform the market averages. However, if they do, some other active manager must underperform. So an investor has a choice. He/she can try to pick an excellent manager who will beat the market, or he/she can buy an index fund knowing that the index fund will beat the average active manager

The problem is: How do you identify the excellent manager as opposed to the average or below average manager? The obvious answer is that you look at past performance. Unfortunately, that doesn’t seem to work very well. Managers who have beaten the market averages in the past may have been skillful; or they may have been lucky; or they may have had a market strategy that was correct for one time period but is wrong for another, or they may have retired or moved to another job. 

Several researchers have studied the persistence of mutual fund performance. Carhart (reference 1) wrote a well-known paper on this topic in 1997. Figure 2 is taken from his paper. Funds were grouped into deciles based upon their performance in one year, and then the performance of each decile (a portfolio including the funds in that decile) was tracked for five successive years. Figure 2 shows the performance of each decile group in the succeeding years. The higher group continued to outperform for the first year, but the performance of the different deciles converged for longer periods. By the fourth year, the bottom decile had moved slightly ahead of the top decile. 

More recent studies have tended to confirm this picture of very short term persistence in mutual fund performance. A recent Morningstar study (reference 2) compared funds that were sorted into quintiles based on their performance during a sorting period of 1 to 10 years. The study then looked at the subsequent performance of each quintile for a time equal to the sorting period, starting at the end of the sorting period. Figure 3 shows the results. The numbers in the chart are the difference in the percentage return of the top quintile and the bottom quintile. Where the chart says 1 Yr, the sorting period and the subsequent evaluation period were one year. Likewise, where the chart says 2Yr, the sorting and evaluation periods were two years, etc. The chart shows some persistence of performance for one year, but for longer periods the persistence is mostly lost. It is interesting that for the ten year period (sorting and evaluation), there seems to be evidence for some performance persistence. However, taken as a whole, the data indicates that performance persistence is very weak.

Aug 2

Figure 2: Carhart (reference 1) data on the persistence of mutual fund performance. Funds were grouped into deciles based on their performance for a one year period. The performance of each decile was then followed for the following five years. The vertical axis is the monthly return in excess of a short term Treasury bill rate.

Figure 3 Aug 2016

Figure 3: Morningstar data (reference 2) on the persistence of mutual fund performance. Funds were sorted into performance quintiles based on their performance for the period shown in the column header. Then the performance was followed for the same period of subsequent years. The numbers in the table are the differences in the performance of the highest and lowest quintile during the subsequent years. 

If mutual fund performance persistence is weak, then the case for indexing is strong, but the question remains: when is too much indexing too much? At the extreme – if everyone was indexing, it seems there would be little competition in the active arena, so large scope for active investors to generate excess returns off the fruits of their research. So how much is too much?

We recently read a blog that had an interesting approach to this question. The blog is philosophicaleconomist.com. We have commented about this blog in the past. The articles that appear there are thoughtful, well written, and often have nuggets of wisdom that aren’t available elsewhere. We don’t know who writes this blog. Apparently the author doesn’t want his or her identity to be know, so we will call him or her PE for short. 

PE recently wrote four pieces about indexing, and one of them, “The Paradox of Active Management”, had a new approach to this question. PE suggested that as active managers lose funds to passive managers, a time may come when the active managers don’t have enough fee income to properly research all of the stocks. Because of their lack of knowledge, they will demand a larger bid-ask before buying or selling in the market. In addition, the number of active managers may decrease to the point that the competition for buy and sell orders is reduced, and this would also increase the bid-ask spread. So PE suggests that one of the real functions of active investors is that they provide liquidity to the market; i.e., they keep bid-ask spreads narrow so that other (passive) investors can buy and sell without taking big loses. Passive funds don’t trade much, but they may have periods of net withdrawal or net contributions, and they have to buy or sell at these times. When they do, they lose the bid-ask spread, so this is another drag on their performance in addition to fees. As bid-ask spreads gets bigger, passive funds will tend to fall behind their intended target index. PE notes that this hasn’t happened yet. The Vanguard S&P 500 Index fund tracks the index almost perfectly (Over the past five years, it has trailed the index by 0.03% per year, even less than its fees).

We would add one thing to PE’s argument. As the fee income of active managers falls, many of them may lose their ability to research all stocks, and therefore, they will have no basis for assigning relative values to stocks. To take the extreme (and unrealistic) case, let’s assume that many of them, lacking adequate funds for research, decide to pick stocks with a random number generator. Investors will eventually figure this out, but in the meantime, the manager can collect fees. However, if there is an active manager who retains the ability to research all stocks, he/she should be able to buy underpriced bargains and to sell overpriced stocks. Under these circumstances, the manager who retains the ability to do adequate research should easily outperform the other active managers, and he/she should be able to do it year after year. 

In short, another way to know when indexing has become too popular is to observe that there is significant persistence in mutual fund performance. Tracking mutual fund performance persistence is more difficult than tracking how well an index fund keeps up with its index. We don’t know how often Morningstar will repeat the exercise it reported in reference 2. However, Standard and Poor’s has a “Persistence Scorecard” that they issue periodically, and one can look at this report for evidence that performance persistence has increased.  

The upshot: we don’t think that we are anywhere near the point where indexing is too popular to become effective, and we now know two clear signals to look for if or when this point does approach -- bid-ask spreads will increase, and it will become difficult for index funds to accurately track their index. For now, with bid-ask spreads miniscule (effectively less than a cent per share on commonly traded stocks) and most index funds having no issues tightly tracking their indexes, indexing continues to look like the responsible and mature approach for individual investors who want to tune out the noise and act like financial adults.

Reference 1: Mark M. Carhart, “On the Persistence in Mutual Fund Performance”, Journal of Finance, Volume 52, Issue 1, March, 1997.
Reference 2: Alex Bryan and James Li, “Performance Persistence Among U.S. Mutual Funds”, Morningstar Manager Research, Jan. 2016.


Get our next article delivered to your inbox.

Sign up below and be the first to know about our freshest data-driven thinking on the markets, and investing. We will send you no more than one email a week. This is free.

Ready to start putting this into action?

Take a free two-week trial to IvyVest premium -- our premium subscription service. You'll get access to our rules-based dynamic asset allocation model, tools that will show you exactly what you need to buy in your own discount brokerage account (and when to re-balance) to implement it for yourself, and an insightful monthly newsletter that will keep you on abreast of the most important things going on in the markets. There is no credit card required. Get Started Now!


By Alex Frey