Why monkeys beat fund managers

By | March 10, 2015

Every now and then a bored financial publication goes to the zoo, rents a monkey and pits him against a bunch of mutual fund managers.  Lo and behold, the ignorant monkey who doesn’t know a thing about betas or earnings potential beats the significant majority of active portfolio managers.

These journalists must be pretty good monkey pickers and should probably start a MFoF (Monkey Fund of Funds) charging 2% plus 20%.  It may be that evolution must have robbed us of our stock picking skills, but boy can we pick monkeys!

So what is going on?  Charlie Munger and Warren Buffett both hint at it in various writings and books, my favorite is Charlie’s Poor Charlie’s Almanack. Warren calls it de-worsification, or the fact that many portfolios get worse as more components are added to them. The theory which he pretty well proves via his actions is that the more holdings you have the less likely you are to know a lot about any of them.  He doesn’t say it, but most managers also jump the median, the investment equivalent of jumping the shark.  Buffet and Munger both hint at investing being similar to paremutuel betting.

Stock returns, as exhibited on the graphs below reflecting data from 1983-2007, are log normally distributed having fat tails.

Lifetime returns for equities Cumulative distribution of equities over time

The Mean Ignorant Monkeys.

Ignorance and apathy are key investment traits.  Put another way, know what you don’t know and be patient, activity kills in this game. Ignorant Mean Monkeys beat funds because they allocate naively, ie equally in their portfolios. God forbid the monkey’s start running mean variance optimizers or the game is over.  Mean variance portfolio allocation is probably one of the most costly rear view mirrors of all time.

Suppose, that out of the Universe of 8,000 equities above Mr. Mean Monkey takes on Mr. or Ms. Median jumping manager by throwing his darts.  Most of the these competitions stop right there.  Rarely do the journalists ask the monkey about next quarters earnings projections or where the Dow will be next week. The Monkey’s bets are spread equally across the selected equities and the race with the clowns is on. The clowns performance is usually measured by their funds returns.

The Monkeys should perform roughly the same as a naive (unweighted) index with a small cap bias.  Why?  The monkeys are simply taking a random sample of the markets, so the returns should be roughly equivalent to the chart above.  The smaller cap bias will reflect the greater number of smaller cap firms present in any sampled universe of investable firms.

Imagine that from the 8,000 stocks the monkey selects 8.  One could think of the chart above broken into 8 bins.  That is what the monkey’s return will equate to.  The monkey is an approximate unbiased sample of the market.

Send in the median jumping Clowns.

The active fund manager “median jumper” clown does the same thing with his or her 8 picks, ideas, gambles, allocations or whatever glossy euphemism he or she, uses in their marketing literature.  But the clown is at a disadvantage because they don’t know what they don’t know, so they decide to weight the portfolio according to the “best” picks.  Uh oh, here is where the trouble starts, they just jumped the median and probably picked something closer to the mode, the most commonly occurring sample.

The median is not the mean.

The statistical mean is what most people consider the average of returns.  More importantly is the median and mode, the most likely sample to be chosen and the number separating the higher and lower half of the sample.  It works like this.  You and 7 of your friends find yourself in a room playing bridge with Bill Gates and Warren Buffet.  The mean (average) net worth of the individuals in the room is measured in the billions, but you don’t feel any richer because most likely you are one of the 8 representing the mode or most common sample in the room which is below the median and the mean.

The median and mode for share’s returns is actually well below the cap weighted index which is closer to the mean.  The charts above indicate 64% of shares under perform the index.  This means that managers who “tilt” or weight their portfolio to their best idea are statistically taking a random sample and doing bad things, increasing the odds of picking a sub-index performer and in so doing also diminishing the allocations to the potentially significantly positive outliers that help deliver the mean performance.

This is classic behavior flaw 101, people with more information believe they know more about something.  So the fund manager takes the useless sell side research reports, technical analysis and other hocus pocus that they haven’t back-tested or thought and parks a few more chips on red.

Few managers really understand businesses.  Not many money managers have actually run a competitive business and understand the dynamics of a “market for goods and services”.  When it comes to fund management and equity selection, I will bet on a humble person who has run a successful small business over somebody who can tell me what they think the Fed is doing.

Stock “business” selection skills are different from macro economic analysis and best learned by doing.  People straight from school and fed into the investment banking world are put into an environment that demand answers even when there any.  Smart people who don’t have answers or admit ignorance get weeded out.  This is an environment rarely interested in posing interesting new questions which is a far more interesting and important skill requiring imagination.  Even Buffet ran a gas station into the ground and probably learned as much from it as his Columbia MBA.

Transaction and operational costs etc. also injure manager returns.  This is the case of apathy/ patience proving things out.  Fund managers are like trapped animals.  They get caught, then get nervous and waste energy trading and allocating instead of sitting quietly and thinking about what the dynamics of the game they play truly are.  My suggestion shut off the screen and blackberry and go fishing etc. for awhile and reflect, maybe a few answers will pop into your head, if you are truly lucky a profound question will arise.

The well known reality is that most fund managers underperform a passive representative index.  Managers don’t know what they don’t know.  They are often in the giving answers business as they rise from analysts to portfolio managers, some are just great salesman stock brokers.  Cramer is the worst public example of an analyst gone wild.  He is a bottomless pit of useless answers filling the airwaves to pimp cars, cereal and pills for CNBC. 

Most Mutual Fund Managers could significantly increase their long term performance with an equivalent naive asset allocation across the board and less activity, assuming they have a random sample.  They could at least aspire to mean monkey performance and still get the fun of throwing darts while causing less harm.

$6.24 Trillion dollars of clown self delusion

For fun take a look at your funds top few holdings ideally representing 80% of the portfolio, and allocate them equally.  You might find that a Mean Monkey portfolio approach beat your managers median jumping.

Calling mutual fund managers clowns may be rough, but if a $26 trillion industry under performs the Naive approach by 1.0% and the average fund has expenses of 1.4% that means you or your pension fund loses $26 trillion x 2.4%= $624 billion a year playing the fund game.  The label clown doesn’t seem to harsh and is more fun and light hearted than the anger going around these days.  Over 10 years that lost clown value would equate to $6.24 trillion, not including opportunity costs and compounding.

Send in the Monkeys!

Be Sociable, Share!

2 thoughts on “Why monkeys beat fund managers

  1. Corey Rosen

    Finally, someone is making this point. I don’t get it. Take any statistics course, and the first ting you learn is mean, mode, and median. It doesn’t take a Ph.D. to figure out that if you have a distribution pattern in which there is an absolute bottom (stocks cannot go below zero) and no absolute top, the median almost invariably must lag the mean. Yet, as you say, somehow trillions are ventured on the supposition that somehow it won’t, at least for you.

    The same logic of assymetric return possibilities you illustrate so well here points to a problem with executive pay being based largely on massive option grants. They also have a downside of zero but an unlimited upside. So it makes sense to take outsized risks (if this doesn’t work, I lose some opportunity cost, but no actual cash; if it does I can become very, very rich indeed).

    I suppose the explanation ultimately is we live in a Lake Wobegon world where most of us think we are smarter than average (70% of people surveyed at work say they are in the top 10% of performers). Of course, if it were true, it wouldn’t be true anymore.

  2. Pedro T

    @Corey That’s true statistically if you use stock prices. But most analysis use stock returns, and those can go below zero.

Leave a Reply

Your email address will not be published. Required fields are marked *