It’s a common observation that tracker funds outperform managed funds over the long term. The charges levied on managed funds to support their staff are substantial, and make the effective return lower.
What about the underlying portfolios, though? Do any of them outperform the market in the long run (i.e. in a statistically significant manner)?
I suspect that the answer to this is no, but what if it is yes? What would it mean for the efficient-market hypothesis?
Clearly, if someone is earning excess returns, then the strong form is incorrect1.
The semi-strong form, that all public information is baked in to current prices, can only hold under dubious circumstances if someone is earning excess returns. Those earning excess returns must be party to private information. So if the semi-strong form is true, and some managed fund portfolios do outperform the market, then their charges could be viewed as the price of that private information. The market of managed funds is then a market in the effective obtaining and using of private information.
The weak form need not be challenged by any (statistically significant) outperforming portfolio, provided that it is not managed on the basis of past prices.
Assuming that any successful managed funds are not using private information to generate their excess returns, could there be an argument for a further form of the efficient market hypothesis? This fourth form would effectively lie between the weak and semi-strong form forms. Let’s call it the mild form for the moment.
The weak form prohibits successful technical analysis, and the semi-strong form prohibits successful fundamental analysis on public information. A mild form of the efficient market hypothesis, between those two forms, might look like this:
All past prices are reflected in current prices, and some easy-to-discover public information is also reflected.
Some managed portfolios can then outperform the market by reflecting more public information than the market does. The cost of that public information is external to the portfolio, but included in the cost of the fund. Managed funds then compete to use public information more efficiently than each other.
To re-iterate, all of this discussion turns on the existence of managed portfolios that significantly outperform the market. If managed portfolios do not outperform the market in a significant way, then the semi-strong form can hold.
What have I missed? How am I wrong?
1 – I believe that some investors on fundamentals, famously including Warren Buffett, consistently outperform the market. I’m not sure if these outperformances are statistically significant, although I would guess that they are. If they are, then the strong form is empirically disproved.
Things I Thought 