Saturday, March 17, 2007

Oversold Stock Market, Or First Leg Of A Bear?

A question I'm hearing again and again from traders is whether we're starting a bear market or completing a normal bull market correction. Few current questions are more relevant for traders and investors alike.

Let's take a look at what the data are telling us.

Suppose we describe each day's price change in the S&P 500 Index (SPY) as a function of its average daily price range over the past 200 days. We thus adjust price changes for the volatility occurring at that time and measure change in terms of "range units". A market that goes up .50 range units thus rises by an amount equivalent to half the average daily price range over the prior 200 days. A market that goes down by -.50 range has fallen by an equivalent amount.

Such a measure enables us to more directly compare price changes from historically volatile market periods to those from markets, like the present one, that have been non-volatile.

Now we will construct a very simple overbought-oversold (OB-OS) indicator by summing the daily range unit changes over the past 20 trading sessions. The result, plotted from 2004 to the present, can be seen in the chart above.

Several patterns are evident in the chart. First, we have seen important market bottoms when the OB-OS measure has dipped below -3.0. Second, we've seen sharp rises off the market lows confirm that we, indeed, have seen lows. Those sharp rises indicate that buyers are jumping into the market to obtain bargain prices. Third, we see that upside momentum in the OB-OS tends to peak ahead of price, as indicated by the blue arrows. That tells us that we tend to see smaller upside thrusts as bull swings age.

At an OB-OS reading of -6.0 recently, we touched oversold levels not seen since the important June/July 2006 market lows. Note that each oversold reading in OB-OS has occurred at higher price lows. That, by definition, occurs in bull markets. Nothing thus far in the current market decline has violated that pattern.

Indeed, going back to the start of 2004 (N = 785), we find 84 occasions in which OB-OS has been below -3.0. Twenty days later, SPY averages a healthy gain of 1.88% (64 up, 20 down). That is much stronger than the average 20 day gain of .44% (442 up, 259 down) for the remainder of the sample. If the bull market is intact, we should see higher prices shortly.

Ah, but here's the rub: If we examine the data going back to 1998 rather than back to 2004, we see that 20-day returns following OB-OS readings of -3.0 or lower are actually negative. What gives?

The answer lies in the difference between bull and bear markets. What makes a bull market is that relatively mild downturns in the OB-OS are taken as buying opportunities by investors. During bear periods such as late 1998 and the 2001-2002 period, those same downturns in the OB-OS become much deeper. To give just a few examples, OB-OS reached -14.0 in July, 2002; -10.0 in September, 2001; and -11 in late August, 1998.

Notice that we can't attribute the difference between those periods and the recent ones simply to market volatility. Each decline is measured in units derived from the volatility of the market as it traded then. No, what makes bear markets--even adjusted for volatility--is that what had been a normal, corrective buying opportunity in the bull market now no longer attracts the same buying. This lack of demand, in turn, stimulates further sales and deeper OB-OS lows.

So what does this mean for the current market?

It is the quality of the next market rise--not anything we can crystal ball at this moment--that will probably tell us whether or not the bull market is alive. If this is a normal correction in an ongoing bull market, we should see investors jump into bargains and drive prices sharply higher. If we get only a tepid bounce from these lows that cannot make new price highs, that will embolden the bears for another round of selling.

In practical terms, that means I take the Missouri approach to the bull market: show me that something has changed since 2004. Until I see distinct evidence of change in actual market patterns--not just fears of mortgage lender bankruptcies, inflation, or a runaway Yen--I will assume that the current regime (trend) is intact. The market will have to show me that something has changed before I alter my investment strategy. It can do that in one of two ways: by producing a subnormal bounce from these lows or by resuming a decline to distinctively more negative OB-OS levels.

At the June/July, 2006 lows, we saw over 50 S&P 500 stocks make fresh 52-week lows. After the recent market tumble, we've seen 8 S&P stocks make annual lows. This time may indeed be different, but the current market will need to show me more before I come to that conclusion.

In an upcoming post, I will update my assessment of Dollar Volume Flows into the large cap stocks to see if institutions are using these lower market prices to pick up bargains. After the initial large market drop, if you recall, Dollar Volume Flows told us that we were *not* seeing bargain hunting. That turned out to be a helpful indication that further weakness was ahead. If we continue to see weak money flows into stocks, even at continued oversold levels, that would trigger my Missouri skepticism about the bull.