Kudos to Bespoke Investment Group for coming up with yet another interesting graphic, which I have reproduced below.
The graphic shows that the average daily volatility of the SPX was 0.72% in 2007, higher than the low volatility years of 2004-2006, but below the 80 year average of 0.75%.
For more information, try the original Volatility? What Volatility? post at Bespoke.
There's so much wrong with what Bespoke has done here that I'm afraid to start typing.
ReplyDeleteSPX didn't exist 80 years ago. If they wanna try 50-ish years, that's cool. Or do the Dow.
Average absolute daily price change isn't "volatility." Imagine a large change that is constant on every day of the year, and you'll easily see the difference between average absolute daily price change and real volatility. "Volatility" is the standard deviation of the daily changes, or how inconsistent these are.
Sometimes I'm surprised at what people who charge for analysis, can call analysis.
Good point on the history of the S&P 500, which goes back to 1957 (see more on the history of the S&P 500).
ReplyDeleteI'll check with Bespoke and see whether they reverse engineered the SPX going back earlier or used different S&P indices -- some of which go back to 1923.
Your other points are valid, but I am content with the Bespoke proxy for volatility as at least directionally correct, if not the identical statistical calculation that is commonly used to calculate volatility.
Cheers,
-Bill
I'm not as content with the proxy, mainly because it's just too easy to calculate StDev(array) or {stdev(if(array1=x,array2))} on a spreadsheet.
ReplyDeleteHowever, I'll have a small spoonfull of crow tartar on a cracker:
There's a real strong correlation between the average of the daily absolute value of pip moves and the actual standard deviation of the daily pip moves, for each year from 1957 to 2007, so doing one or the other gets similar, but not identical, results. Differences in rank order between the two methods are largest in the areas away from the extremes ... and least at the extremes.
The three "least" are the same in rank order, and the eight least are the same (although ordering is different). The two "most" are the same in rank order, and the top nine are the same (but out of sequence).
I'll be curious what they say about the 1927-1957 data, it must have had far higher average daily moves, because the Yahoo!Finance data from 1957 to present averages only about 62 basis points daily.
Seems right to me. Its not clear that we have even moved to a normal credit environment yet. Defaults are still well below average.
ReplyDeleteBespoke provided the following clarification about the pre-1957 S&P data: "For data prior to the 1950s we use the S&P 90 index"
ReplyDeleteLink to Bespoke clarification.