« Sharp Angles | Main | Comments and Random Thoughts »

The Forecasting Follies

September 15, 2005 11:05 AM

Nearly every day an article appears in the financial press "explaining" that such and such a stock fell in price because its earnings did not meet the forecasts of analysts. What nonsense.

What would you say about someone who forecasts the winners of NFL games who got it wrong almost all the time? And why are people so fascinated with forecasts anyway? Election coverage is little more than breathless repetition of polls and long-winded discussions of what they mean for the candidates. Who is going to win the election is a question everyone seems to want to know. But, that knowledge is of little value and no one can predict it unless it is obvious to nearly anyone.

Back to the financial analysts. What the news says when it is announced that earnings are less or more than forecasts is that the analysts got it wrong. It doesn't mean anything about the company whose stock is involved. And the report that the price fell because of the difference between realized earnings and forecasts is just a story. No one knows why the price fell, or even if it did relative to the market portfolio.

There is no reason that one quarter's earnings relative to a forecast should change a company's value significantly. It should make the stock price of the analyst fall, not the company's. Nobody can forecast quarterly earnings reliably. They are random, after all. So they will seldom equal any expectation. The forecasts are also point forecasts of a single value. So any realized earnings will nearly always fail to fall on that single point.

As I said in my Edge Annual Question Center law...

The future is over forecasted and under predicted.

We are fascinated with them and pay people to do them even though they are almost never right. Sounds like the weatherman, but even weathermen do far better than financial analysts and political forecasters. Maybe it is just entertainment, The Forecasting Follies. Or, maybe humans, at least those raised in the present world where politics has elevated social statistics to a kind of belief system, look for any number when they face large uncertainties.

Have a look at this terrific critique of forecasting by James Montier, The Folly of Forecasting published on the DrKW Research site (thanks to Nassim Taleb for bringing this to my attention). My book Hollywood Economics covers the forecasting follies in the movie business.

· Uncertainty

Comments

Posted by: Flower Online [TypeKey Profile Page] at September 12, 2006 7:11 AM

This is too large a topic to be fully explored in this forum, but I don't think that I can agree with all of Professor De Vany's comment, in particular the first three lines, (but in some other aspects too):

1. It is clear that analysts forecasts are generally poor; they are heavily guided by the companies themselves, but they do serve as a means to get information into the market and they do serve as a proxy for the markets expectations.

2. Commentators do seek for explanations of price movements and often they are wrong - especially along the lines of "S&P falls as investors see interest rates rising" in the morning followed by "S&P recovers as investors see interest rate weakness" in the afternoon. However, that does not mean that they are always wrong. I would be fairly confident in forecasting that if there was a major terrorist strike that took out significant amounts of Saudi oil production, stock markets would fall; using the same example, any commentator saying in this instance that "markets fell following the terrorist attack on Saudi oil facilities" would very probably be correct.

3. The fact that the analysts are wrong about a company's earnings does not alter the fact that the news that they are wrong can itself be "new news regarding the asset value of the company" that "would make the price fall". You can argue forever that analysts forecasts are unreliable junk (and I would agree with you), but if a company's earnings come in at half the number they are expected at by the analysts and by proxy the market (and this would probably be after the company had tried to steer the analysts down), then this is market sensitive news and the price will change as a result of it's announcement. Note that as Professor De Vany says, nothing has actually changed at the company as a result of the news itself, but the market's information about the company certainly has changed.

4. Consequently "There is ..." every "reason that one quarter's earnings relative to a forecast should change a company's value significantly". The market sets its price (including its random walk) based on information at hand; if a company announces that profits are unexpectedly down, then that would usually suggest that the information at hand is wrong, hence the need for a price adjustment. It's a subtle difference, but while the profit news announcement hasn't changed the underlying value of the company, whatever has caused the actual drop in earnings certainly has.

5. While I quite agree that "Nobody can forecast quarterly earnings reliably." I can't agree that "They are random, after all". There is an element of randomness, but, for example, I would be comfortable to bet any reader that Microsoft will make a 'large' profit next quarter (that does not mean I would be right, nor does it mean that Microsoft will always be profitable!): if quarterly earnings were truly random, we should expect Microsoft regularly to make large losses, which I don't recall it doing.

6. Obviously I need now to produce proof that not only do earlings shocks affect price, and furthermore by more than relative to the market... However, I would be pretty sure that every investment bank on the planet has a study of this, so I am also sure that the studies do exist.

Posted by: Edward [TypeKey Profile Page] at September 16, 2005 1:32 PM

Ah, but you don't know if it fell relative to the market portfolio. And, you don't know why it fell if it did. I don't know of any solid studies of this. Most event studies of this kind deal with clear and strong public announcements of real effects, not some one's opinion.

Only new news regarding the asset value of the company would make the price fall and, even then, only if it is new information.

The real event is that the analysts were wrong. Eventually, given the clear and abysmal failure of analysts to forecast earnings (or market price now of in the future), the market would ignore analysts. However, if the price does fall reliably (and there is no reason for it to do so), then there is an arbitrage: sell on the news and buy it back tommorrow.

Posted by: Art [TypeKey Profile Page] at September 15, 2005 8:52 PM

............Ahhhhh the peacocks tail feather of rampant pedantry !
I put it to you Sir that you're in need of a good dose of camping or sleeping out under the stars.

Posted by: fellows [TypeKey Profile Page] at September 15, 2005 6:07 PM

You assert that the belief (or explanation) that a stock will fall in price because it does not meet analysts earnings estimates is nonsense. That is surely wrong. It happens all the time so that belief is true. If you mean to say that the fact that this happense is nonsense, that is surely correct.

Posted by: Fugate [TypeKey Profile Page] at September 15, 2005 2:36 PM

Post a comment

Thanks for signing in, . Now you can comment. (sign out)

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)


Remember me?