Monday, January 28, 2008

The Market Prediction Game: Here We Go Again...

There's been a lot of activity in the markets so far in 2008. We've seen uncommon (though harldly unseen) volatility. And with volatility comes one of "The Street's" favorite pastimes; The Market Prediction Game.

But how do these predictions tend to pan out? With talking heads doing their talking thing everyday, it's hard to keep track of the daily (hourly?) deluge of prognostications.

But when we do collect the information, it is telling. The Wall Street Journal surveyed top economists semi-annually, to get forecasts on what bonds were going to do over the next 6 months. The data go back to 1982.

The experts (intelligent people all, to be sure), were wrong in the predictions of the direction bond yields 66% of the time. That is to say, when asked 6 months from now will the yield on a 10 Year Treasury be A) Higher or B) Lower... they got it right 1 out of 3 times. (Source: Davis Advisors)

Do you realize how bad that is?

Employing a black-tailed marmoset to throw darts at a board marked "higher" and "lower" would be a better predictor!

MARKETPSYCH LEGAL COUNSEL DISCLAIMER: does NOT promote or otherwise endorse the practice of marmoset dart throwing. Sure, it's fun. But that's beside the point. Arming small, wiry primates with sharp objects for throwing is dangerous and most likely illegal in the US (with the possible exception of licensed establishments in the state of Nevada). Marketpsych partners are NOT responsible for damages suffered by those engaging in this activity.

The fact is, human beings are notoriously lousy predictors of future market events. A study by George Wolford and associates at Dartmouth College found that even rats and pigeons outpeform humans in short-term market prediction. (No word on marmosets).

This does not mean the market doesn't have cycles, or that patterns don't emerge. Indeed, to be wrong 2 out of 3 times (as the economists were on bond yields) lends credence to the notion that the predictions are NOT random. It points to the central theme of short-term reactivity that seems to dominate investing patterns - something we call Whack-A-Mole Syndrome. (TM)

My colleague, Dr. Richard Peterson, has written about it here in his superior book, and even developed the Marketpsych Fear Index which tracks how investor emotion is often an inverse predictor.

But the point is you don't need a crystal ball to be a succesful investor. You need a few simple but undervalued qualities. 1) The ability to recognize companies with proven records. 2) The ability to recognize when their stocks are at an attractive valuation vs. earnings. 3) The discipline to invest your money in them... and not monkey with it. (pardon the pun).

But we can't help ourselves. With so much information available, with so much money on the line, we love to engage in the Prediction Game. (By the way, Pats 34 - Giants 14 - you heard it here first!).

The Market Prediction Game reminds me of the end of the classic 80's flick, War Games (starring a young Matthew Broderick), when "Joshua", the American military super-computer aborts a nuclear launch on the Soviet Union because it realizes that it would result in mutually assured destruction. The computer learns the folly of the eponymous "War Game".

"Strange game. The only winning move is not to play."

Indeed, Joshua. So don't play.

How about buying some great companies cheap?

Or perhaps a nice game of chess, instead?

Monday, January 21, 2008

Is Capitulation Here? Advice for plunging markets

The million/billion dollar question is when will we hit market bottom. A lot of traders and short-term investors are feeling the heat right now.

Ironically, bottoms are often hit when investors capitulate. If you're fully invested, the usual benchmark for when the bottom is hit is when you finally give in to that feeling that you're going to lose everything, and it's so painful that you just can't take it, so you throw in the towel and sell everything. I don't see capitulation yet. The sell-offs still feel somewhat orderly and gradual (though rapid).

If you're waiting on the sidelines with cash, it's often better to not "catch the falling knife", but to wait with buy-stop orders that trail the market, so when the rally occurs, yours is one of the first orders hit. Of course, when any rally appears, there will initially be a trickle of buyers, and then a stampede. It won't be easy to get a great price on a stop order that fills at the market price during an explosive relief rally.

No matter where you are, it's never useful to sell into a panic unless you sell a little bit ("throwing a maiden in a volcano") to give you mental breathing room.

Now is a good time to assess what powder you've got dry, and look for the opportunities. The opportunities will usually not be in the previously hot sectors. A stock screen using low P/E won't catch the drop-off in earnings expected in many companies that were booming last year. However, if you add low price to book as another screening criteria, then you might find better quality issues.

Volatility rewards those with patience and a clear mind ...


Friday, January 18, 2008

Negative Market Expectations at a High

If expectations drive stocks, then this graph should be of interest. As opposed to the MarketPsych Fear Index, this is a plot of the relative percent of negative expectations (subtracting out positive expectations, such as for market "recovery" or "rebound"). It looks like investors are expecting very bad news going forward. As you can see, investor expectations were relatively more positive in June and July 2007. The relative percentage is displayed on the left y-axis. A negative value actually indicates a positive balance of investor expectations.

As in our other graphs, this is a candlestick chart (in this case of the QQQQ - Nasdaq 100 proxy). The brown line is a 30-day exponential moving average of the balance of negative-positive expectations. It is derived from the results of a linguistic analysis of the financial press. Essentially, you are seeing the frequency of reported negative expectations attributed to investors.
Does this ugly graph mean it's a good time to invest? Well, we haven't crunched the numbers on this one yet, but we will soon....

Monday, January 07, 2008

Hail to The Redskins! (Curse); Hail Victory!! (For the incumbents)

Did you know that when the Redskins LOSE their LAST HOME game, the INCUMBENT party has LOST every presidential election since 1936? Isn't that spooky!

Well, not really. It's more silly than spooky, I think we can all agree.

Plus, George W. Bush ruined the streak in 2004, so the "Redskins Curse" is over.
For those who are interested, Snopes (the site dedicated to debunking or validating urban myths) has more on the subject.

But it's still fun to see how our superstitious minds can craft tales of curses and omens and lucky charms that predict the future.
Let's take a look at the Redskins curse in greater detail for a moment and see how Behavioral Finance would explain the development of this myth?

Q: Why the Redskins?

A: The Availability Bias. We tend to use the information that is most handy when we make decisions/predictions. Washington is a political town where people pay more attention to elections. The 'Skins are the local team. It stands to reason that they'd notice a political or 'Skins related anomaly. (In Green Bay... not as much.)

Q: Why the last home game?

A: The Recency Bias. In a series of events we tend to remember the events that occurred in the beginning and, even more so, those that occurred at the end. (Let's face it, who could remember the statistic if it occured in week 5 of the season?) Plus, we remember events that carry greater emotional weight. Home games are more likely to be attended by the politically interested fan base. When you leave a stadium, you remember a win - as well as a loss. (Epecially if that loss was to the &*$#*@! Cowboys).

Q: Why did Redskins Curse exist at all, why the anomaly?

A: Probability. We know that the incumbent party has a natural advantage. We also know the home team in football has a natural advantage (usually at least 3 points according Las Vegas). It should come as no surprise that two events will tend to occur simultaneously when the probabilities are greater than 50%.

The question of predicting future events based on past events is an important issue for today's trader given the popularity of "back-testing" strategies (plugging in your future strategy to past events to see if it would have worked). Numerous online brokerages currently touting this method as tool for validating trading models. And it can be. But there is a fine line between back-testing and data mining. The key is recognizing that correlation does not equal causation.

Since the Redskins Curse is dead. I think it's time to come up with a new, cool curse. Doing so means engaging both sides of the brain. MarketPsych has provided a model below.

The Pittsburgh Steelers Curse for Democratic Candidates

STEP 1: Mine The Data (Left Side of the Brain - The Correlation)

First, you're going to need a stat, a several standard deviation event that makes for an interesting coincidence. Fortuntately NFL records provide a mountain of data in which to go mining. As you would with a stock screen, sort through every Steelers season on election years since 1936. Eventually, the screen will turn up some anomaly - a particular week, a particular stat - that has consistently correlated with Democratic Party victories. Let's say that this particular data holds up for week 7. (I.e., When the Steelers LOSE the 7th game they play in a season, the Democrats always LOSE the presidential election.) Got your stat? Good. Proceed to step 2.

STEP 2: Create A Narrative Around It (Right Side of the Brain - The Causation)

The left side of the brain will do the math. But the right side will "tag" it with a story. (More on this and other fascinating brain explanations in my colleague Richard Peterson's brilliant book.) The story needs to create some sort of plausible context that would support a potential reason for the anomaly. And the spookier, the better. Sometimes, the numbers do this for us. (Lucky number 7, hooray! The number 13, booo!). But there are many ways to create a deeper meaning for the numbers. Use your imagination! Tie it to a disgruntled player (maybe he wore #7 !) who's uncle was the Republican nominee. The Steelers traded him on week 6 and ever since that fateful day....

Personally, I like this one: The owner (Who owned the team before Art Rooney?) was an enthusiastic supporter of Franklin Roosevelt, and invited FDR into the locker room to address the team on week 7 back in 1936. FDR, fine orator that we was, gave the 'boys a major pep talk. He ended it with a promise. He told the assembled players that "If you win today, I'll guarantee you a victory in November... as well as lower taxes on steel products, moustache wax, and Polish Sausages!" (Hey, all politics is local, y'know?) Well, don't you know the Steelers rallied to defeat a powerful Chicago Bears team with a miracle last second pass. And ever since that fateful day...
Feel free to create your own. Just don't bother using the Jets. The whole franchise is already cursed.