Yeah, But Are You "Sure-Sure"?

Ivory soap is famously 99.44% pure. I like that extra 44/100s. It gives me peace of mind.

If only financial forecasters would follow the Ivory model in their predictions. I've been hearing/reading/seeing a lot of expert predictions these days. New calendar years and volatile markets seem to attract them.

Now, let's be clear, I don't have a crystal ball. (I do have a Magic 8-Ball. But when I asked it if the Jets would make the playoffs it told me "Signs Point to Yes." So I'm thinking it's busted.)

The only predictions I will make with any confidence are these:

1) All consensus predictions will be too narrow in scope.

2) People will overuse artificial parameters in the form of round numbers and calendar years when formulating those overly narrow predictions.

Okay, I cheated.

Those aren't predictions. They're observations of human behavior that are among the most reliable you will ever find.

How reliable? Research into the area that behavioral finance folks call "overconfidence" indicate that when people are asked to predict a range in which they are 99% confident results will fall (i.e., a 99% confidence interval) they are correct 80% of the time.

Now at first blush, that may not seem so awful. 80% vs 90-something%...what's the big deal?

But it is awful.

Truly, horribly, make-you-want-to-toss-your-cookies awful.

Why?

Think of the corresponding behavior in light of such predictions. When we're 99% of something, it's basically as close to saying we're absolutely certain as we're going to get.

You could go Ivory Soap and say 99.44% certain but when we blurt out, "I'm 99% sure that won't happen", we're essentially saying, "No shot in hell."

That's dangerous even when it's TRUE.

Once in a hundred years was the standards to which they built the New Orleans levees. That works fine... right up until your neighborhood has to be airlifted off the rooftops.

But with market predictions, it's 20x worse. Events that people - and this includes experts, mind you -- say would happen every 100 years (1%) - happen EVERY FIVE YEARS (20%).

Let's say you listen to a more conservative expert predictor. He/she is twice as good and are accurate 90% of the time.

That STILL means every 10 years we're going to experience something that "nobody" saw coming.

Nassim Nicholas Taleb wrote a book called The Black Swan. (It's not as good a book as Richard Peterson's Inside the Investor's Brain, but it's certainly worth reading).

Where are we seeing such predictions these days?

Oh... everywhere.

"Where do you believe the S & P will be a year from now?"

"How high do you think unemployment can go?"

"What are the chances you will have to cut your dividend, Mr. CEO?"

Remember, fellow investors, fight the danger of narrow framing and don't be drawn into sharing the outlook of those who look at the horizon through a key hole and tell you wide it is.

We have no reliable way of knowing how bad (or how good) it's going to get.

The key is to expand the scope of expectations and to have plans in place for even the most unlikely-seeming scenarios.

Think "Ivory Soap".

And good luck.

-Frank

(If you are interested in a MarketPsych seminar, please feel free to contact us at info@marketpsych.com. I'm 99.44% sure you will find our seminars valuable.)

CEO's do it.

Fearless Forecasting: How Low Can You Go?

It's official. The DJIA dropped 20% from its highs last October.

In other words, the Bear is back.

Whenever we hit a nice round number (e.g., "Dow 10,000) or experience a round number move (e.g., "Down 20%) it leads to a big picture discussion of where the market has gone... and where it will go next.

That means "market predictions".

In an earlier post, I observed that employing a black-tailed marmoset to throw darts at a board would prove just as useful an exercise (and an infinitely more entertaining one.)

It may be useful in at this time to review two major causes of precisely why.

One major cause is something called the Gambler's Fallacy, a miscalculation that ironically tends to afflict more market savvy investors (pros) than casual investors (amateurs).

Quick Example: Say you're at Mohegan Sun (where I was last week) and you're observing the roulette table. The wheel turns up "red" results 7 times in a row. These results don't fit with our mental schema. We know that the odds of a ball coming up "red" vs "black" at a roulette table is roughly 50%/50% (47.368/47.368 to be more precise). Our brain says something to the effect of "Black is due"! And we feel the urge to bet (overbet?) on a black result next time. Of course, the odds of the wheel yielding a "black" result are the same as ever - roughly 50/50. But it feels like it should turn up black, and that feeling overrides our rationality.

This is the classic manifestation of the Gambler's Fallacy - the notion a series of independent events yield useful information about predicting future independent events.

Pretty elementary stuff, I grant you. So why should something so obvious effect even top Wall Street Strategists?

Because the same tendency reveals itself in Market Predictions.

Hersh Shefrin, in his landmark book, Beyond Greed and Fear, provides a relevant example. At the beginning of 1997, Barron's interviewed chief strategists from top Wall Street firms, requesting 12 month market predictions.

On June 20, the market had risen 19.7% for the year to 7796, well above every strategist had forecasted.

A chief strategist for Smith Barney said in response, "We've all been humbled".

When Barron's asked the strategists for revisions predictions in late June, the average prediction was for the DJIA to drop 10.3% by year end.

Point of fact, the DJIA close slightlty higher for the year at 7908.

So despite all we know about market tendencies to move higher, the experts predicted a steep, upstream move in the opposite direction.

Why did they do it then and why do they continue to do it?

The answer is the investing version of the Gambler's Fallacy, that template driven interpretation of regression to the mean. We know the Dow tends to go up on average 9% or so every year. And we have a strong desire to fit predictions into that template.

But there is nothing magical about a calendar year - it's just a handy way of charting time. And if stocks tend to go up 9% or so every 12 months, than regression to the mean demands we predict that stocks will go up 9% or so every 12 months - not that they will reverse themselves according to our schedule in order to provide yearly averages.

Now, I'm not throwing stones here. Believe me, I'm not. I'm wrong constantly. And certainly all the participants were wise and learned professionals whose opinions are worthy of respect. But that's part of what makes this so fascinating.

Even they (especially they?) are not immune from the same impulses that drive roulette players to overbet because they think "red" is overdue or because a single digit number hasn't popped up in a while.

And - I can't help myself, I'm gonna say it - the other factor is no, (gosh darnit) they were not humbled, despite declarations to the contrary.

Wrong? Yes. Embarrassed? Perhaps. But humbled? No way.

A crucial component to being humbled is admitting you are wrong.

By prediciting a 10% reversal, the experts adjusted their predictions to support their original predictions.

Trying to prove you were right all along is not humility. It is the opposite of humility.

So with a bear market here and the inevitable market predictions to come, what are some things for investors to keep in mind?

1) Stay ready.

2) Stay humble.

3) Recognize the mathematical illusions inherent in regression to the mean.

Happy Investing.

Frank