Saturday, October 04, 2008

Stock Market Volatility: A Historical Perspective


We are currently living through a period of significant stock market volatility. Let's take both a historical and a psychological look at what that means.

The chart above shows the S&P 500 cash index (blue line) versus its 20-day average daily true range (ATR). The 20-day ATR is a measure of actual price volatility in the index; not implied volatility as calculated by options and not premium levels built into options pricing. The ATR is the larger of the following:

* The percentage price range between today's high and low prices;
* The percentage price range between today's high price and yesterday's closing price;
* The percentage price range between yesterday's closing price and today's low price.

What this means is that ATR captures, not only the size of the range during the trading day, but also the gaps that occur between yesterday's close and today's trading. It is a measure of past price volatility both within and between trading days.

The chart is showing us that actual price volatility has gone into a parabolic rise and that we recently exceeded the volatility levels from the 2002 bear market period.

How significant is this?

I went back to 1962 and looked at periods in market history in which the 20-day ATR for the S&P 500 Index has exceeded 3%.

Here are the market periods and the highest levels of volatility for each:

* October/November 1987: 5.60%
* October/November 1974: 3.95%
* October 2008: 3.84%
* July/August 2002: 3.83%
* June 1970: 3.40%
* June 1962: 3.31%
* October 2002: 3.25%
* December 1973: 3.10%

As we can see, the current level of price volatility is not without precedent, but is at very significant levels: the third highest period since 1962. (By contrast, the median 20-day ATR since 1962 has been 1.40%; the standard deviation over that period has been .57%).

From a psychological vantage point, volatility represents uncertainty in the market's estimations of value. Imagine a market in which there was no volatility whatsoever: market participants would be in perfect agreement as to the value of stocks and hence would have no reason to trade with one another. As there is disagreement in estimations of value, we have incentives to trade, as each participant believes that the market will ultimately move to his or her projected fair value.

When those estimations of fair value become widely dispersed and when participants make great shifts in these estimations, markets become more volatile. During times of economic and political crisis, such as the 1962, 1987, and 2002 periods, we see such volatility in the pricing of markets. A close look at the historical record suggests two conclusions:

1) Periods of historically high volatility have tended to occur during bear markets and have been associated with long-term opportunities for investment. If you look at the periods in which we've had very high volatility, these have tended to correspond with important bear market lows since 1962. The notable exception was December, 1973, which was an intermediate low in an ongoing bear market;

2) Periods of historically high volatility can persist and do not necessarily correspond to precise bear market price lows. Volatility was high through much of the latter part of 2002, for instance, and yet stocks did not bottom until March of 2003. We didn't get a price bottom until December of 1974, after volatility had peaked the month before.

So where does that leave us? As long as volatility is historically high *and* on the rise, as is the case in the current market, it is treacherous to pick bottoms. Periods of elevated volatility can become super-elevated, as in 1987; they can also persist, as in 2002. Panicking and rushing to liquidate your investment portfolio is also treacherous, however, as these elevated periods have been associated with some of the greatest investment opportunities in the past few decades. By following the market like a trader--tracking dollar-volume inflows and outflows on a daily basis, the willingness of large market participants to hit bids or lift offers, the interest in riskier yields vs. safe Treasuries--I find it easiest to keep myself out of trouble as an investor. That saves considerable capital during periods of record volatility; we don't get a bottom until valuations become so low that they represent irresistible bargains for investors. Those trading indicators suggest that we're not there yet.
.

4 comments:

Bill aka NO DooDahs! said...

Here's an even longer-term look at stock market volatility, going back to 1901 with the DJIA.

http://www.marketthoughts.com/z20080403.html

Swami said...

Great post! Clear, objective and concise...sums up my "gut" feel that we haven't bottomed yet.

abel said...

I am trying to find positive divergence somewhere. At this time, absolute New Lows are below peak readings reached at the various selling climaxes earlier this year, (although the ratio of NH's to NL's yesterday matched prior peak readings, .03).

I also track the inverse of the above, the New Low to New High ratio, which is showing a divergence.

This NL/NH ratio reached it's end of day high point to date thus far on 9-29-08, (34.83), with an intraday reading of 42.94 the same day. Yesterday, it closed at 29.89. My data begins from 10-20-05.

This is a thin reed, to be sure. It may help indicate an 'early' slowing to downside participation, even as and if the market may continue to move lower from here.

In a broad sense, the market usually goes after the big dogs late in the downmoves, as the little dogs have already had the stuffing kicked outta them, so the divergence described above may be reflecting that dynamic.

Another form of reflecting the distinctions between momentum lows and price lows.

Mark Wolfinger said...

I find this data fascinating. Thanks for bringing it to our attention.

Dr. Brett, I'm wondering what the data looks like from this perspective: How volatile was the market 16-20 days ago - the days that will be replaced by this week's numbers? If they are relatively low, we can be pretty sure that the 3.84% number is going to move higher.