The efficient-market challenge

The serious objection to everything in this course is not that value investors are unskilled. It is that the information they work from is public, that public information moves into prices as soon as it exists, and that a method which reads filings anyone can read should therefore earn nothing beyond the market.

State it properly before answering it

The position is not a claim that markets are always right. It is a claim about what is left over for you. If thousands of people with the same filings and better resources are already pricing a security, the part of its price that your own reading can improve on is small, and the cost of finding it is real.

That argument has a second half which is the hard one. When someone measures a persistent return difference between cheap and expensive securities, the position does not have to deny the measurement. It can accept it and say the difference is compensation: the cheap group is riskier in some way the measurement did not capture, and you are being paid for bearing that risk rather than for spotting a mistake.

The measurement that made this an argument

Basu's paper is titled as a test of the efficient market hypothesis, which tells you what was at stake when it was written. He ranked New York Stock Exchange industrial companies, about five hundred a year, by price to earnings at each year end from 1957 to 1971, sorted them into groups, and measured what one year of holding each group returned.

The figure above is the result. The cheapest group returned 16.3 percent a year against 9.3 percent for the most expensive. One million dollars in the cheapest group became 8,282,000 dollars over the fourteen years; the same amount in the most expensive group became 3,473,000.

The column that matters most

It is not the return column. It is beta. The most expensive group carried a beta of 1.11. The cheapest carried 0.99, and the two groups in between carried 0.97 and 0.94.

The risk defence needs the cheap group to be riskier. On the risk measure the defence itself uses, it was not: the cheapest group earned seven percentage points a year more while moving slightly less with the market than the most expensive group did.

Where this leaves the tradition

Not refuted, and not vindicated. The measurement survives, the interpretation is contested, and the contest is about what risk means rather than about arithmetic. Fama and French's 1992 study pushed the same question from the other direction, finding that once size and book-to-market were accounted for, market beta had little remaining relation to average returns in their sample.

That result is usually quoted as a blow against beta. It is equally a problem for anyone using beta to argue that cheap securities were riskier, which is the reason it belongs in this lesson rather than in a lesson about cost of capital.

Check yourself