Spikes in the VIX: Harbinger of a Coming Crash… or Not?

[dc]G[/dc]iven last week’s volatile spikes, there has been a lot of talk about the VIX and volatility in general. This is a good time to take a deeper look at the VIX and make sure we understand what it is, what it measures, and what it might say about the future direction of stock prices.

What is the VIX?

The VIX is sometimes called the fear gauge, is seen as an accurate forecast of future volatility, rising VIX means the market is going to crash—none of these things are really true. In fact, far too much attention and far too much time is spent talking about the VIX. What is the VIX? Well, without going into the gory details, VIX is simply a measure of the price of options on the S&P 500 index, nothing more.

If you want to get a bit more technical, the VIX is the square root of the annualized forward price of the 30-day variance of the S&P 500 return based on a replicating portfolio of options delta-hedged with stock index futures. (You probably didn’t want to get that technical.) The calculation is quite different from backing an implied volatility out of an options price via, say, the Black-Scholes model, but the concept is similar in that the VIX is a measure of expected volatility over a 30 (calendar) day window.

The VIX is quoted as an annualized percentage. If you want to know what the number is actually saying, you must “de-annualize” it by dividing the current VIX by the square root of 12. For instance, Friday’s VIX was 17.03. 17.03 / sqrt(12) = 4.92, meaning that the S&P return is expected to vary by 4.9% over the next month. So, the VIX is simply a measure of expected volatility—nothing less and nothing more.

Seasonality in the VIX

image001

Seasonality is a squirrely thing, at best a slight influence and subject to severe distortions. The chart above shows a seasonal index of the VIX generated from the past 10 years of data. Historically, VIX has a tendency to bottom sometime in July and then to increase into October. In that context, was last week’s increase surprising? Nope—this is simply the VIX adhering to its typical seasonal pattern.

Daily changes in the VIX

image003

Many people discuss the day to day changes in the VIX, but I have never been able to find any useful information in the daily changes of the VIX. If you graph the daily changes in the S&P against changes in the VIX, you will find that there is remarkable stability in the relationship. Only very rarely will you find a move in the VIX that is larger or smaller than you could have predicted by the daily change in the S&P. If you are going to look at the VIX every day, you need to know what the historical relationship has been. The chart above shows one way to display the relationship. The daily return of the S&P 500, expressed as a standard deviation of the past 20 trading days’ returns, is on the vertical axis, plotted against the percentage change in the VIX on the horizontal. Over a very long history (the chart only shows a few recent days), the relationship tends to stay within the shaded area on the chart. Last week saw some big declines in the stock market. The VIX went up. Your reaction should be “so what?”

Long History

8-2-2014 3-11-26 PM

There has also been talk of the VIX being at unprecedented levels and how this points to a coming market crash. This is simply not true, as a look at the last two decades of the VIX’s history shows. Recent levels have been low, but they are not the lowest ever recorded. Levels have lingered this low for three years in the past. Yes, low VIX levels preceded the 2007-2009 Financial Crisis, but low levels also set up the dot com boom of the 90’s.

Also, a look at this chart puts last week’s VIX spike in perspective and shows it for what it is: a non-event. We’ve had similar VIX spikes over the past couple of years, and, so far, there is no cause for undue alarm.

What Usually Happens After the VIX Spikes?

When we analyze prices of financial assets, the first task is usually to convert those prices to returns, or percent changes. We need to be careful with the VIX though, since it is already a percentage measure. In this case, differencing (i.e., subtracting the prior period’s measure from the current) may be better since low values of the VIX will lead to possible distortions in the percent change. In my analytical work, I’ve favored talking about point changes in the VIX, rather than percentage changes, though you could do either.

So, what happens after the VIX has a large spike? Last week’s action was at least moderately interesting: based on VIX data back to 1/5/1990 (1,283 weeks), last week’s increase was in the 94th percentile for both weekly and monthly VIX point changes. (If you look at percent changes rather than differences, then both of these are 98th percentile.) Perhaps because of the popular simplification of thinking of the VIX as the “fear gauge”, people tend to think that spikes in the VIX are an ominous sign for the market. Nothing could be further from the truth.

image007

The chart above shows that stocks have lagged over the next month following a large spike in the VIX, but have actually outperformed over the next quarter, six months, and the following year. The big difference between median and mean values should warn us that there might be some large outliers in there, and there are: 2000, 2007, and 2008 did have some large negative returns over the following twelve months, but, overall, buying stocks following a large spike in the VIX has been a winning strategy. Of course, it could be different this time, but the lesson of history is clear.

I suspect what is happening here is a function of mean reversion in stocks: the VIX increases when stocks go down, and large spikes in the VIX have always been associated with stocks selling off. Large spikes (up and down) in stock prices (individual and indexes) tend to be reversed, so buying dips in stocks is a good strategy, if you can handle it with correct risk control. The VIX may add nothing to this equation except pointing out spots that stocks have sold off hard and might be set for a sharp recovery.

Last, there is one more point to consider, and we’ve been pointing it out in my daily research for many, many weeks: Volatility clustering is the tendency of volatile moves in the stock market to lead to more volatility. We had some large shocks in recent history, so have been on guard for selloffs and corresponding increases in the VIX, and we should continue to expect more volatility in the near future. We are certainly in a challenging trading environment and it seems everyone is talking about the VIX. Now, more than ever, it is important to understand what the VIX is, what it measures, and what it might say about the future direction of stocks, which is that all of the hyperbole and fear is simply not justified by the historical record.

AdamHGrimes

Adam Grimes has over two decades of experience in the industry as a trader, analyst and system developer. The author of a best-selling trading book, he has traded for his own account, for a top prop firm, and spent several years at the New York Mercantile Exchange. He focuses on the intersection of quantitative analysis and discretionary trading, and has a talent for teaching and helping traders find their own way in the market.

This Post Has 5 Comments

  1. alrik3

    So if I understand you correctly you are saying that a temporary massive spike in the VIX is typically a market overreaction to the short circumstances at hand due to crowd following ie if price is up with increased VIX the chance of a drop is emminent and vice versa. ie mainly profittaking or entering positions. However if the VIX starts to trend from a low ranging position and price is at an extream this might acctually be the turning point to a new trend in price… Or did I get it right….:) best

    1. Adam Grimes

      There are quite a few points in your comment that aren’t supported by what I wrote. You may well be right (I don’t know), but you’re bringing some very different thoughts to the discussion. Be careful with labeling market moves “profit taking”, “short covering”, etc. We never really know that. Also, it’s exceedingly rare to see a large spike in the VIX tied to a sharp rally in stock prices. Volatility evolves asymmetrically in stocks.

  2. joana

    Hello, how is a spike calculated in this case? Same as a vol spike? Thank you

Comments are closed.