[<< | Prev | Index | Next | >>]

Friday, October 04, 2002

Efficient Market Hypothesis

Email snippet:

> If the market is just the sum of all the individual buy/sell transactions,
> why could one not "beat the market" by simply being a better-than-average
> investor?  In other words, some people are losing money, some are making
> money.  Don't you just need to be in the top half of the class to come out
> ahead?  That doesn't seem so hard...
> What am I missing here, according to the EM hypothesis?

1) Most people, over the full range of intelligence, do not invest a great deal of time and effort into investing. Thus, there is a huge pool of people who are not active investors but *could be*.

2) Given any fixed set of information, there is an objective value to a stock, which reflects the probability of all possible outcomes given that information. Note that improving one's estimate of a share value is as much (or more) about acquiring information as it is about raw intelligence. [In fact, since it is not a timed test, intelligence may have little bearing compared to, say, just the pragmatic knowledge gained in acquiring a degree in business or finance.]

Taking these two to the extreme, imagine that the entire market is determined by one uber-investor who has as much information as can possibly be gathered and filtered up to one individual, and who is really really smart. (And everybody just invests in his mutual fund.) This is the ultimate "efficient market" and there's actually nothing inherently bad about it, and in fact it's probably the theoretical ideal in terms of global efficiency. Which is an important point to grok if the rest of my argument is to make any sense, so let me know if I need to clarify.

What that reference point illustrates is that there's a theoretical maximum return the world on average can expect from investing, and it has to do with the actual productivity of the companies that exist to invest in. The only way for an individual to beat that return, other than by chance, is if there are other individuals who are investing poorly.

However, consider: One individual pulls his money from the uber-fund and starts investing it poorly. Immediately, the uber-fund soaks it up and distributes the poor fool's money amongst all of its investors. I.e., it's one bad investor with his own money vs. one good investor with billions of people's money. Eventually the bad investor gets tired of losing, and puts his money back in the uber-fund.

Now, let's imagine the uber-fund didn't absorb the fool's money, which is actually very unlikely assuming the uber-investor is doing his job but just for the sake of argument: Due to (1), there's a HUGE pool of very smart, capable people ready to jump in and take it themselves. That is, as soon as it becomes more profitable to invest their time in investing than in a regular job, they make the switch. Whether it takes one good investor to compensate for one fool, or less or more, it doesn't much matter, since we're very very far from running out of people who could quit their day jobs and start investing for a living. As long as that pool doesn't run out -- which is to say, from your personal perspective, as long is there is one guy out there who understands finance better than you do but isn't applying that ability at the moment -- then you have to figure investing isn't paying notably more than the same effort put elsewhere (like making shoes or something).

Ultimately, the transfer of money from the fool to the good investors is really just a small amount of noise around an objectively determined value. There's nothing that says you have to have fools investing at all, and that they're punished for doing so tends to keep the numbers down.

It's the small percentage of the population who have the highest product of intelligence, interest, and access to information who end up collectively implementing the role of the uber-investor, because they are the ones who turn a profit and are thus both motivated to keep at it, and able to garner control over other people's funds. The ones who lose money can't lose money for a living, so that side is inherently self-limiting!

Anyway, I'm not sure how to neatly summarize this since it came out less organized and clear than I was hoping, but does it seem to answer your question?


Aaron's reply to same:

Some factors you are missing:

1) Transaction costs. Playing the market costs money. Its still a positive sum game, but its a negative-sum game for investors with respect to the average rate of return ("the market") because of transaction costs. You might think these costs are small, but if you trade very often they add up, and can easily negate a small edge. According to poker simulations I've seen, when skill differences are fairly small (or result in fairly small hourly-rate differences), even a modest table fee has a dramatic impact on the proportion of winning players, ie from 50% (playing for free) to 10% or 5%.

2) When computing how good you need to be to be better-than-average, you need to multiply everyone's skill by the amount of money they are controlling. Yes, if we ignore transaction costs, the sum of all winners wins equals the sum of all losers losses. But there is a high correlation between skill at investing and how much money a person controls. Not only because the better people make more money and so have more money, but because the better people get to control lots and lots of other people's money. It takes a lot of $10,000 / year investors on one side to make up for a Warren Buffet making billion-dollar bets.

3) Being "good" is related to how much information you can obtain or analysis you can do, and the big boys can afford to spend more resources obtaining information and doing analysis because they get to make big bets when they find a sweet spot. The cost of the information about a stock is amortized across a larger number of shares. So they have better information than you, which lets them better predict prices.

The above is not at all complete, there are other things going on, but its what I can articulate at 3am. I was gambling with positive expectation in a negative sum game until 1am, and its bedtime :).

And Rich's thoughts:

>If the market is just the sum of all the individual buy/sell transactions,
>why could one not "beat the market" by simply being a better-than-average

I think the best way to comprehend the awful truth of the efficient market hypothesis is to consider this: if you were to look at the long term rates of return of professionally managed mutual funds, how do you think they would compare to the market's own average rate of return? If you think that smart people who are experts trained in finance and in specific market sectors can do 'better than average' you would expect that these mutual funds would have greater than average rates of return. However, when you run the numbers, they do a little worse than average. In fact, since they trade more, once you factor in transaction costs they do considerably worse than average. Of course, some do better than average for a while . . . but can you pick the ones, in advance, who will do that?

>In other words, some people are losing money, some are making
>money.  Don't you just need to be in the top half of the class to come out
>ahead?  That doesn't seem so hard...
This isn't high school where you are competing on a standardized test :)

A *much* more accurate analogy is: wouldn't a person with a lot of knowledge of sports be able to beat the Vegas spread on pro football bets? The answer is no, because *and this is the crucial point* the spread is adjusted based on *how people bet*! Same with investments: once a strategy that returns better than average returns is identified, the influx of money into that investment drives up the cost and brings the rate of return back to average.

BTW I think Andrew*'s and Simon's and Aaron's posts have been right on, but thought you might appreciate one more perspective.


[<< | Prev | Index | Next | >>]

Simon Funk / simonfunk@gmail.com