A working paper that is nicely summarised on the Empirical Finance Research Blog shows how markets react to information rapidly, that they are not perfectly efficient, and how those setting the efficient market price are rewarded.
The key finding is that returns are lower after information is disclosed, and higher in periods when information is not disclosed. The main implications of this are:
- The market does react rapidly to information (evidence for semi-strong efficient markets)
- The market is not perfectly efficient, because then there would be no reward for fundamental stock picking, which is the mechanism through which the (efficient) prices are set. This is similar to the way in which arbitrage opportunities are limited by arbitrageurs.
- You can make a profit by short selling shares following disclosures, and buying shares on which there is little new information.
- You probably should not buy immediately after new information is disclosed (e.g. not immediately after the results)
There is an alternative, behavioural, explanation for the the first of these: that the market consistently under-reacts to bad news and over-reacts to good news. Given that it is obvious that there must be a market mechanism to reward those who invest based on fundamental research, in order for prices to adjust, the behavioural explanation seems redundant: by Occam’s razor I would reject it pending further evidence.
The main finding is that it confirms a theoretical prediction based on efficient markets. To quote from the blog post:
In the end, one of the basic predictions of the Grossman and Stiglitz theory is that when information is cheap, accessible, and easy to understand, prices will be very close to 100% efficient. Whereas, when information is expensive, has restricted access, and is complicated and time consuming to decipher, prices will likely stray from their “fundamental value.”
The nice thing about this is that it gives us semi-strong efficient markets and a mechanism that compensates those who generate fundamental research.
My understanding of EMH is that efficient markets are essentially random, and therefore it’s not possible to profit from superior interpretation of already public information. However, you’re saying it is possible to profit from public information (though that seems to be denied definitions of the semi-strong form you mention).
If all that the efficient markets hypothesis tells us is that, other things being equal, the more timely, accessible, and understandable information is, the more difficult it is to calculate a more realistic price for a share than the market price then I agree.
But as the author of the paper says, real world markets aren’t informationally efficient:
“I fi?nd that alphas are even higher among smaller stocks (up to 7% per year), where information production would be more costly. These numbers, in all likelihood, underestimate the true information premium. The portfolios use only stocks from the NYSE, known to have one of the most strict listing requirements in the world,
and hence is more likely to be informationally efficient.”
Also the main finding of the paper doesn’t confirm the Grossman/Stiglitz theory you quote, the author proposes further study to determine whether its true or the alternative behavioural interpretations you mention, but reject, are true.
I suspect both are true, s, for example, investors seeking to beat the market will find larger returns in smaller value stocks, but larger value also offers the opportunity for profit (in the long run of course).
Despite all this your interpretation of the EMH seems to be much more nuanced than the one promulgated by the industry (e.g. index tracking) sections of the financial press, and academia (e.g. Burton Malkiel) who ‘for their own good’ advise investors they can’t beat the market because the market is somehow right. At it’s worst that encourages investors not to think about price, which is very dangerous and quite possibly contributes to investment bubbles and therefore the inefficiency of the market!
Once you’ve got to the point where you admit that markets are inefficient enough to reliably gain excess returns though, it seems to me you might just as well say they are inefficient as efficient.
Just to add, you might call them imperfectly efficient – as Andrew Smithers does.
And a final thought. We know there are inefficiencies i.e. opportunities for excess profit. The question is whether the cost of analysing them is worth the benefit in excess returns. I’m sure it doesn’t in theory, because ‘there’s no such thing as a free lunch’. However, at the level of the individual investor, it depends entirely on how slick, ‘efficient’ even, your method is. There’s a thought. I think I’d better get back to work :-)
Hi Graeme, Richard,
I’ve recently taken an interest in investing, and Richard’s blog has pointed me towards various sources of interesting reading material.
I’ve just finishedMontier’s book, and found it pretty convincing.
At the moment I’m trying to understand the counter arguments.
Greame, I didn’t fully understand this passage, any chance you could help me to?
“There is an alternative, behavioural, explanation for the the first of these: that the market consistently under-reacts to bad news and over-reacts to good news. Given that it is obvious that there must be a market mechanism to reward those who invest based on fundamental research, in order for prices to adjust, the behavioural explanation seems redundant: by Occam’s razor I would reject it pending further evidence.”
Also, isn’t there an element of Pascal’s wager to all this? If markets are efficient, then, as long as you’re sufficently diversified, all strategies are equal. You might as well use the strategies that some claim can beat the market.
The issue we are dealing with is that returns are lower in the periods following good news. The behavioural explanation he is suggesting is that, for example, the price rises a little too much on good news, so the share performs a little worse in the following month.
My objection to this is that given that we know there must the a mechanism by which the prices adjust, and the lower returns can be explained by this mechanism itself, we do not need the behavioural explanation.
It is hard to imagine any strategy that will not make you less diversified: in order to beat the market, you must pick winners and concentrate on some part of the market.
However, as I have already agreed with Richard that value effect seems to undermine modern portfolio theory I have to say that value investing looks like it might give better returns than the market at the same of lower risk. Its a conservative approach so it is not going to be very risky anyway.
However, if you are a value investor in a period when growth out-performs, you could still do badly – I have blogged on why it might be a good time to buy growth. That said, the recovery is far from certain, so value might still be safer.
Thanks for the reply, Graeme. It’s clarified your position for me. Basically it’s that a rotational strategy is probably better than a pure value strategy, right?
“My objection to this is that given that we know there must the a mechanism by which the prices adjust, and the lower returns can be explained by this mechanism itself, we do not need the behavioural explanation.”
Still not entirely clear on what this mechanism is. Obviously the mechanism involves investors spotting imbalances and correcting them, but I’m still not sure how this accounts for the fact that the good-news shares underperform in the following month.
By the way, what do you think to the iShares value and growth ETFs (idjv and idjg)?
Do they represent a hand picked value/growth basket of shares fairly well?
In general I would prefer a rotational strategy, but it is more risky as you may get the timing wrong. It is perfectly possible say whether the market is in a bull, bear or bubble phase, but it needs detachment. If in doubt, go for value and income.
The mechanism works because the market tends to undervalue shares about which there is little information: they are more difficult to research and value.
When the amount of information increases (for example, when the results are released, or a merger is announced), the share price tends to rise.
This rise is bigger for shares about which there is generally little information.
This results in an out-performance for people who bought before the information become available. This rewards those will to do research, and gives bigger rewards to those who do more difficult research.
I personally prefer to put together a portfolio myself. I am not familiar with those particular ETF, but I assume they track style indicies, in which case I would prefer them to a market index tracker: I would mostly keep my money in the value one and switch to growth during bull runs: Richard had a nice graph of the best timing recently.