The common sense truth about managed funds

There are two types of managed fund. One type, the traditional "active" managed fund has the goal of producing returns superior to an appropriate benchmark index, and they attempt to meet this goal with a combination of stock picking, market timing and asset allocation decisions. The other type of fund is called a "passive" or "index" fund. These funds don't try to beat the index, they just try to match it as closely as possible. The main object of an index fund is to produce returns in line with that of the asset class, minus their very small fees.

It is obvious why active funds are so popular. It is hard to bring yourself to accept average returns. Why settle for average returns when you could be above average? It seems when you look at a list of managed funds, the great majority seem to have beaten the index, a testimonial to the triumph of professional management over the deliberate mediocrity of index funds, or so it seems.

Now a bombshell, "mediocre" index funds have on average, in a wide variety of markets, over very long periods of time, with great consistency, outperformed roughly three quarters of all active managers after fees. Some index funds appear to be harder to beat than others, for example in the United States the main benchmark index is the Standard and Poors 500 index, and this has beaten virtually every managed fund in America bar only a few for decades. Australian share funds do seem to have had some success, here index funds beat the majority, but apparently not a large majority of active funds.

Over the period of March 1995 - March 2002, most retail funds underperformed their benchmark indexes in most asset classes. Possibly Australian share funds are an exception, but if so they are the only one.

Retail funds beating major benchmark indexes

(Source: Vanguard, using data provided by ASSIRT)

The median performance data over the same period shows the underperformance of active funds compared to indexes:

Retail fund performance vs major benchmark indexes

(Source: Vanguard, using data provided by ASSIRT)

When you plot the performance of every managed fund vs the index, the message is particularly dramatic. Here is the performance of every US large cap fund with a 15 year history vs the S&P500 and CRSP 1-10 indexes over the 15 Years ending 31 December 2001 (285 Funds). As you can see, while a small number of funds did outperform, they didn't outperform by very much. On the other hand most funds lagged by a very substantial amount. The success of the winners did not compensate for the failures of the losers. A diversified portfolio of active funds underperformed the S&P500 index by about 3%pa.:

US large cap funds vs the S&P500 index

(Source: http://www.tamasset.com, picture originally by Dimensional Fund Advisors)

Active funds don't seem to be capable of reducing risk either. Even if they can't perform any better surely one would think active fund managers would be able to avoid the risky stocks. Index funds may hold hundreds or even thousands of stocks, but most active funds would hold about 50. As a group, active funds are actually more volatile than index funds.

US active funds volatility vs the S&P500 index

(Source: TAM Asset Management, Inc "Investment Policy Guidelines & Strategies Within the Context of The Prudent Investor Rule", "average active fund" = average of 7125 US domestic equity mutual funds in the Morningstar Principia Pro Database July 1991 - July 2001.)

It may seem strange, that all those highly paid portfolio managers as a group don't seem to be able to display any consistent ability to perform better than a diversified portfolio of stocks chosen on the basis of nothing more than their size, but all of this is exactly as it ought to be.

 Active funds vs probability theory

First I will state something that may appear obvious, but it lies at the heart of a very successful investment strategy: it is mathematically impossible for everyone to be above average. No matter what the average is, there must always be a distribution such that half of all people (or in the case of investing, the managers of half of all invested dollars) would be below average.

There are two ways that an active investor may seek above average returns:

And of course that is just what most investors are trying to do. There is a nagging problem with this though that investors as a group need to be aware of. If you want to sell a stock, someone has to buy it off you. Another glib truism, but if you'll bear with me this is going somewhere.

If you want to sell right at the top, someone is going to have to buy right at the top. You will perform much better than average, and whoever buys that stock off you will have below average performance as a result. The same goes with inherently "superior" and "inferior" stocks, for you to choose a portfolio containing only the best issues, someone else must be left holding the bad ones. Clearly as a group, investors are all going to have average performance, it is impossible for any amount of trading or stock selection to improve on that.

Now in investment, the average is measured with an index. Most indexes are market value weighted, meaning the total value of the company on the market accounts for the weighting that company is given in the index. This is sensible because clearly market value weighted indexes are the best (though not necessarily perfect) way of tracking the performance of all investors.

If an index is the average, then like it or not half of all investors are going to do worse than the index, before costs, and there isn't a thing that can be done about it. Of course half will do better, but no amount of trading or research will stop half of all investors failing to beat the index.

Today's markets are better researched and more "efficient" than at any time in history. The dominant players, accounting for most of the dollars flowing around the markets are institutional investors. Whatever your opinion on managed funds, one thing you cannot deny is that they do try very hard to improve their investment performance, if only so as to impress customers and gain more funds (and hence fees) to manage. Managed funds employ the brightest and best graduates in economics and finance, and will pay huge salaries to retain good quality analysts and traders.

The industry is highly competitive, if one managed fund is going to pay top money to get the top staff, then they all will. While undoubtedly some management teams may be superior to others, widespread head-hunting makes these competitive advantages difficult to keep up. The fact is, these highly talented individuals make investment a dynamic profession, and make it difficult for less informed investors to secure any kind of "edge". As clever as these analysts are though, the laws of probability can't be overcome by wits alone: half of all these teams of dedicated and hardworking analysts and portfolio managers are going to underperform the index.

To use an analogy, no matter how hard the athletes train, and how good they are, even at the Olympics somebody has to lose. Half the athletes will come in the second half of the field. It makes no difference at all if the athletes as a group are superb, terrible or somewhere in between, the actual ratio of winners to losers was fixed all along. Investment is the same.

So you might presume that you have about a 50/50 chance of choosing a manager that will beat the index, but you'd be wrong. One thing we haven't considered so far is costs. There are a variety of expenses in investment, there are those big analyst salaries for a start, but they are nothing compared to the costs of buying and selling shares on the market. The costs of trading include brokerage, the buy/sell spread, stamp duty and more. As well as the costs of research, marketing, tax reporting and compliance and distribution for managed funds, it simply costs money to buy and sell shares. The Plexus Group, a US based investment research house estimates that it costs about 0.8% for institutions to buy or sell stock, on average, so a manager that turns over 100% of the portfolio each year will lose around 1.6% in costs.

What is the average turnover of managed funds? At the low end of the scale, there are funds that turn over around 20% of the portfolio each year. At the high end, several hundred percent. It is hard to find local data on what constitutes average, though American funds average around 80%pa.

Remember that while trading can be profitable, it comes at the expense of the inferior trader. If funds are just trading with one another then the costs of this activity are going to drag down the performance of the whole group. The better traders may incur a marginal advantage, but the majority of managers are equally talented, and it is hard to find one that is so good that any long term advantage can be gained from their high turnover strategy, eventually the laws of averages catch up with managers and high costs overwhelm trading activity.

In fact managed funds researcher Morningstar, in a review of 3,560 American stock funds, found that funds with low turnover ratios generate superior long term returns to funds with a high turnover. Over a ten year period, funds with an annual turnover below 20% outperformed funds with an annual turnover above 100% by a margin of 1.58%pa - exactly what you'd expect it to be given the Plexus estimate that 100% turnover costs 1.6%pa! This has been confirmed by many other studies, fund managers as a group do not gain anything from market timing and trading, the highest turnover funds perform worse than the buy and hold funds.

Similarly, fund managers seem to be unable to gain an advantage over their rivals through stock selection. There are only a very small number of funds that have demonstrated a significant degree of success as stock pickers, and even then rarely for extended periods of time. Remember, managed funds all have great stock pickers on staff, professional managers compete with each other and as a result competitive forces (and lucrative head-hunting bonuses) keep the very best stock pickers well distributed around the industry.

Academic studies have found little evidence that superior stock picking stays in any one managed fund. Last year's hero fund manager, with the colossal outperformance of the index is barely any more likely to do well this year than any randomly picked fund. Past performance, despite the massive attention it gets, is practically useless in predicting future results.

The following two pictures tell the story pretty well. In the first picture, based on a study by Frank Russell Australia using Morningstar data, you can see the subsequent performance of all the funds from the top quartile (top 25%) in 1997. Few of them were top quartile again the next year, none in the two years following.

The second graph shows the performance of 1999's top quartile funds. As you can see, top quartile funds did worse in the year before and the year after than random chance alone would suggest. I don't recommend blind "contrarianism" (picking last year's loser just because it was a loser and you think it will bounce), but statistically that is a better strategy than choosing last year's winner.

Studies have in fact found that there is some reliability in longer term data, but it seems to be best employed as a purely negative filter. Really poor performance seems to persist, untalented stock pickers with high costs and sloppy trading don't seem to get any better with time, though good performance isn't as persistent.

Studies of the effectiveness of fund ratings systems such as those used by Morningstar have shown that the systems are reasonably good at identifying very poor funds but not as good at identifying very good ones. Out of the 5 star systems so commonly employed, the 1 and 2 star funds are often dogs, and often continue to be dogs, but there is little evidence that a 5 star fund will do better than a 4 star fund or a 3 star fund. If anything, there is a little bit of evidence at least that 3 and 4 star funds outperform the 5 star funds!

There is one factor that has been shown to have a major impact on results, and by now you may have guessed what it is. Researchers looking for the holy grail of managed fund picking have found that one factor seems to explain most of a fund's long term performance advantage or disadvantage compared to an index: costs. Because everyone incurs costs, the average manager does not outperform the index. The only reliable way to achieve an above average performance is to achieve below average expenses. Passive funds are great for that, lower fees and more tax efficient.

 How to win a "loser's game"

In game theory, there are games called "loser's games" and "winner's games". A loser's game is a game where the outcome is determined by a mistake from the loser. A winner's game is won by the skill of the winner. Good examples of losers games include "noughts and crosses", where it is impossible to beat an opponent unless he or she makes a mistake, and amateur tennis where it makes more sense to play conservatively and not try anything fancy, so you'll let your unskilled opponent whack the ball into the net.

The opposite of a "loser's game" is of course a "winner's game", where points are won by skill. If amateur tennis is a loser's game, then professional tennis is a winner's game. To beat a professional tennis player you can't rely on waiting for him to make an unforced error, you must be more skilled and hit the ball out of his reach.

Charles D. Ellis, in a classic book Winning The Loser's Game explained that due to the difficulty of scoring "points" as an investor, that is, outperforming a crowd of hard working and highly skilled investment managers, investment is best played as a Loser's Game. If you want to beat your opponents you could play a winner's game strategy - try to time the market or pick the best stocks, but you are up against some pretty serious competition here. A more reliable strategy would be the strategy of letting the others make the mistakes, and in investment the biggest "mistake" is high costs. Ellis argued that the best way to beat the crowd would be to adopt a low cost investment strategy.

The great thing about indexing is that it is very cheap indeed. Index funds don't need expensive analysts, they barely lift a finger to sell themselves, incurring practically no expenses in marketing or distribution (ie, when was the last time you saw index funds advertised on TV, and have you ever compared the beautiful glossy brochures of the average managed fund with the spartan black and white printouts most index funds use?). Index funds don't trade much either, so they don't incur the 1.6% cost of buying and selling.

If all funds could operate completely free of costs, index funds would probably sit right on the 50th percentile, with the successful active funds beating them, and the unsuccessful active funds doing worse. But as I've said over the last few paragraphs, active funds are much more expensive than passive funds, which explains why index funds tend to beat between 60 and 80% of active funds in most markets.

So to summarise: competitive forces keep the skill levels about equal in almost all managed funds, which explains why it is rare for any one fund manager to perform much better than its peers for any extended period of time, this also makes it practically impossible to pick winners in advance. Also, since index funds are simply a whole lot cheaper than active funds they tend to perform much better than active funds as a group.

 So why do active funds all seem to outperform the indexes?

The above all makes perfect sense, yet if you look at a list of funds it seems the great majority of them outperform the index. What is going on here?

There are a number of ways that active fund managers have been able to promote the illusion that as a group they are adding investment value. The investment management industry has many tricks to make it appear as if everyone is doing better than average.

For a start they can load the dice by opening to the public a lot of historically profitable funds. Fund managers introduce creation bias into the equation by starting a lot of aggressive new funds. The way this trick works is they give seed capital to a number of promising young portfolio managers every year. These managers then invest and trade aggressively for the next year or two, establishing a track record. The fund managers that didn't do well get the flick, their money gets plowed into the more successful fund and then the investment company opens the new fund to the public and puts enormous marketing hype behind it. In this way managers can ensure that all new funds (that the public hear about) have excellent track records.

There are studies that have found that brand new funds often underperform their own track records, and as a group seem to do worse than more established funds, this is probably why. Note that these new funds don't do any better than average following their launch, but they do get to brag about impressive past performance.

The second trick is to bury the evidence if one of their public funds ever falls behind. Survivorship bias is introduced when fund managers close or merge their less successful funds with more successful ones. By continually weeding out the weaker funds, a fund manager can present a prospectus showing the entire range of investment options being market beaters. Similarly, when a fund is culled it is usually deleted from most databases, so history is rewritten by the winners. Professor Burton Malkiel, author of the excellent book A Random Walk Down Wall Street studied this phenomenon and estimates survivorship bias could add as much as 1.5%pa to the performance of the median fund manager. Investors do not benefit from survivorship bias, real world investors do lose money on funds that are deleted. All that improves is the historical average performance of fund databases.

Third trick is to throw a lot of hype behind high performing funds. There is little evidence that winning funds are able to sustain their high performance over the long term, so this can only be seen as a cynical marketing exercise, cashing in on last year's luck. Naturally funds that don't perform well don't advertise much, so all you see are ads showing high performance.

The fourth trick is as scurrilous as the rest, even a very poor fund that has underperformed over the longer term can have a good year or two, so as long as these funds only talk about recent past performance they can avoid the prickly question of longer term past results. Similarly, funds can always brag about past glories, proudly displaying the fund manager of the year ribbon they won 3 years ago on every advertisement, and brag about a high long term performance even if the last few years have been dreadful.

It is difficult to get hold of data that is free of survivorship bias and these other problems, data by ASSIRT, Morningstar, Investorweb and other commercial researchers are invariably affected by survivorship bias. One way to lessen the effect of survivorship bias would be to look at shorter periods of time, not ten years or more when survivorship bias can be huge.

TD Waterhouse calculated that in the three years to June 30 2002, only about a third of Australian fund managers beat the ASX200 index. In stark contrast to the chart at the start of this article, here is how Australian actively managed share funds have performed over the last three years:

Australian managers are not beating the ASX200 as is so often claimed.

Academic researchers, using data that is "clean" (free of survivorship bias), and weighting the performance of funds according to size (so we get a better picture of what the average investor is getting, rather than give equal weighting to large funds with tens of thousands of clients and tiny boutiques with only a dozen) have shown that there is nothing special about Australian fund managers, they have in fact been beaten by the indexes - just as you'd expect them to if you assume that the laws of arithmetic have not changed on our side of the pond.

And on top of all that, active funds are usually less tax efficient than index funds because of their high portfolio turnover and frequent capital gains distributions. I'll explain this in the next article.