The Dow's 300-point Gains

The Dow's 300-point Gains - So What?

Wednesday, August 6, 2008

One of the pieces of data that's been floating around trading desks today came from a CNBC interview with David Rosenberg from Merrill Lynch.

I don't watch CNBC so I can't vouch for the quote, but his contention was that because we are seeing 300-point gains in the Dow Jones Industrial Average, we're in a bear market because that's what we saw during the last one (and market lore states that the sharpest rallies occur in bear markets). It's a controversial statement, but easily tested. There are a couple of things we need to adjust, though.

First, a 300-point move in the DJIA is meaningless. 300 points today is a lot different than 300 points twenty years ago, so instead of looking at that size point move, we'll look at any day with at least a 2% gain in that index.

Second, we need to figure out how far back to look. For these purposes, we'll sum up the number of daily 2% gains over the past month. That will give us a good feel for how many "bear market rally days" we've seen in a reasonable period. Also, because markets were much different (and volatile) creatures in the 1930's and before, we're only going to look at data since 1950.

The table below shows all periods that are similar to our current one, in which we've seen 3 days with 2% or greater gains over the past month. We've eliminated dates that were within a month of each other so as not to distort the results with multiple days during similar periods.

In the table, we show how the Dow performed in the three months and six months leading up to the clusters of 2% gains. If Mr. Rosenberg's assertion is correct, then all of those cells should be red, showing negative, bear-market returns.

From the table, we can clearly see that clusters of 2% gains did not exclusively occur during bear markets, but it was most common to see negative returns leading up to them. This makes sense, as volatility is higher during down-trending markets than up-trending ones.

Not only that, I don't really care about it. We've operated under the assumption that we're in a bear market since early January. I don't need to know descriptive types of things, I need to know predictive ones.

And these clusters were fairly predictive - with a bullish skew. A month later, 75% of the instances were positive, and six months later 83% of them were. In fact, the only negative six-month returns were during the 2000 - 2002 bear market, which as far as I know is the only one to which Mr. Rosenberg referred.

My point isn't to do a hatchet job on a Wall Street analyst, but my job is to try to find us actionable edges, and debunk media sensations that most often lead us astray.

The fact that we're seeing multiple 2% gains in the Dow now is probably an indication that we're in a bear market (which we already assumed six months ago). But that doesn't tell us anything about what's going to happen - the table above does that, and overall it paints a more bullish picture than bearish.

If you have to listen to CNBC or other mass financial medium, please use a discerning ear.