Tuesday, October 24, 2006

What Historic Low Volatility Means for the Stock Market

My recent Trading Markets article noted that we have seen historically low volatility in the present market. Indeed, going back to 1964 (N = 10,713 trading days) in the S&P 500 Index ($SPX), we have only had 1415 days similar to the current period, in which the average daily trading range has been below 1% on a 40, 200, and 500 day basis. In today's post, I will elaborate on some of my historical findings, which have quite a bit of relevance for the current equity markets.

First off, volatility has been tailing off in recent years. All of the sub-1% average trading ranges over those three time frames have occurred since 1985; none occurred before then. I think this is tremendously significant. With the advent of stock index trading has grown a trade in arbitrage and program trading. This has created greater periods of reduced volatility in the S&P 500 Index. As we see a growing number of ETFs, those opportunities for arbitrage expand to an ever-greater proportion of equities. I would not at all be surprised if we see greater periods of low volatility among NASDAQ, small cap, and mid cap stocks as a result--not to mention sectors such as emerging market issues and international indices.

Second, the notion that range-bound, non-volatile markets are "due" for a breakout just doesn't hold water. The occasions of low volatility have persisted for years (late 1984-early 1986; 1992-1996; 2005-present). Indeed, sixty days after a low volatility period, the average daily trading range is .83%--much lower than the average range of 1.45% for the sample overall. Volatility--and its absence--displays a serial correlation: non-volatile periods tend to be followed by non-volatile periods, not dramatic breakouts.

Third, while low volatility periods have tended to follow market rises, that doesn't mean that low volatility periods are associated with future stock price underperformance. Indeed, 60 days following the low volatility periods in the sample, the S&P 500 Index averaged a gain of 3.37% (1091 up, 324 down)--much stronger than the average 60-day gain of 1.86% (6779 up, 3934 down) for the entire sample. Indeed, buying 5% corrections during low volatility periods has led to very handsome returns.

Fourth, no bear market--including those of 1987, 1990, 1998, and 2000--has begun during a period of ultra-low volatility. Invariably, following such low volatility periods, prices have moved higher with expanded volatility before a bear market has taken hold.

The historical record suggests that low volatility periods persist until macroeconomic events lead to major reallocations of capital, contributing to both volatility and directional price movement. One of those macro events in recent history has been rising interest rates. It is difficult to imagine what would cause volatility to enter the current market in the absence of either radical geopolitical uncertainties or tectonic shifts in rates and/or currencies and commodities.

Finally, a note about high volatility. When the average daily range has been elevated on a 40, 200, and 500 day basis, we have a veritable who's who of excellent buying occasions. These include 1974, 1982, 1970, and 2002. Low volatility does not guarantee a bear market, but high volatility has been associated with the end of many major bear markets. Tracking the ebb and flow of volatility and price change through multiple cycles reveals very interesting relationships that connect nicely with a broader understanding of the business cycle--and the role of institutions in allocating funds to equities relative to competing investments. Low volatility occurs when institutions find no compelling reasons to make radical shifts in reallocations.

2 comments:

James said...

Doc,

How do you calculate 20 or(n)day highs. Do you take all issues or do you include preferred issues and the like? Do you subscribe to a service that calculates the above or do you have it fed into excel and tabulate it yourself. I could probably do it in excel, but i think my computer might blow up. Just a few questions.

the best to you,

jar

Brett Steenbarger, Ph.D. said...

Hi James,

My intraday new highs I calculate with Excel workbooks dynamically linked to my real time feed. Longer term new highs are available via the Barchart.com site, which includes all NYSE, NASDAQ, ASE issues.

Brett