On Monday, Bespoke Investment Group, who probably has one of the most impressive stock market databases around (similar to our ODDS Online, that has one of the most impressive options databases around), put together an interesting report and graphic showing how the market has performed during different earnings seasons. What caught my eye is that when I glanced at the numbers, I got the impression that if you added them up, it would turn out to be a whole lot of nothing. So, I took the average of all the months, and indeed the average return of the S&P 500 during earnings season was a meager -0.11%.
I then compared that average return to the average monthly return for the market, as they are both about 30 days in duration. Knowing that the S&P 500 is pretty much right where it was in October 2001, I knew that the market’s return wouldn’t be that much either. Turns out, it’s a minuscule -0.02%. So there’s really not that much of a difference month-to-month … at least when you look at the average.
Which begs the question: Why are stock and even index options generally more expensive going into earnings season?
The reason can be explained by looking at how far each period’s returns deviates from the average. Hopefully, these graphs will illustrate what I mean.
The first is a histogram of the S&P 500’s month-to-month change since October 2001 – the period in which Bespoke’s analysis began. The x-axis is the monthly percent change. The y-axis measures the frequency of the monthly change. For instance, in the 0 to 2% category, the y-axis is 25%. That means for 25% of the months since October 2001, the S&P 500 gained between more than 0% but less than 2%.
You can see that it has a slight bell-curve shaped distribution. The peak is skewed to the right, telling us that there are more up months than down months. But there are a handful of little blips on the left side of the graph, indicating three huge monthly declines that exceeded -10% each. Fat tails, Black Swans, whatever you want to call them, the bottom line is that this chart does not deviate so far from the bell curve that calculations based on an adjusted normal distribution/Guassian have to be tossed away.
Click image to enlarge
Now here’s a histogram of the earnings-season performance data from Bespoke.
Click image to enlarge
As you can see, there is very little resemblance to a bell curve whatsoever. That hump on the left is a major problem. Also, the bars on the extreme right are much taller than one might expect if the distribution was normal. Historical volatility, being a measure of standard deviation, and therefore, a measurement based on a normal distribution, would not provide anything remotely close to an accurate portrayal of actual market behavior during earnings season. Instead of having a typical bell curve, or even a skewed one like the graph at the top, you get that mess of blue columns that have no simple representative equation!
But here’s the real kicker: Every professional who trades options knows this! They may not know that this is the exact shape of the histogram. They just know that during earnings season, the market makes moves that are different than at other times. And they price options accordingly, not on some formula that they know is based on an incorrect assumption for the earnings reporting period. Incidentally, this is a key factor in our “Measure, Don’t Model” analysis that lies at the heart of ODDS Online.
In sum, while the average monthly performance during earnings season may be nearly identical to other time frames, underneath the surface there’s a big difference. During earnings season, there are a lot more moves that deviate substantially from the average than normal. Option professionals expect there to be more frequent large moves. That expectation of more big moves is the reason why options get a lift going into earnings season.