“Statistics are like bikinis. They show a lot, but never everything.” – Lou Piniella, Cubs Manager
I am always surprised as to the posts readers find most interesting. This post on Volatility Clustering was one of my favorites, but didn’t elicit a single comment. I think it is fairly timely with the pickup in volatility in the stock markets now that we are below the 10-month moving average.
Notice the only market where the volatility is higher when above the moving average – commodities – which makes sense given that commodities are much more likely to spike up due to supply/demand issues. Take a look at a chart of wheat or corn for an example. Volatility for asset classes, on average, is about 25% higher and returns about 60% lower when under the moving average.
Reminds me of the behavioral reasons listed by Lo in this earlier video.
—-
A new book on the way soon from Robert Schiller – The Subprime Solution: How Today’s Global Financial Crisis Happened, and What to Do about It
—-
KKR is going public by merging with their listed fund in Amsterdam. Hopefully (albeit unlikely) this will bring some needed attention to the listed hedge fund and private equity fund space.
—-
I don’t use the mean reversion technique I mentioned in an earlier post, but it would buy US Stocks, Foreign Stocks, and REITS in August for a two month hold. I’ll track the performance with SPY/IWR, EFA/EEM, and IYR/RWX. Other particularly awful performers that could see a bounce are:
Financials (XLF)
Netherlands (EWN)
Sweden (EWD)
India (INP)
Infastructure (MG)
Private Equity (PSP)
Foreign Private Equity (PFP)
I liked that post…it’s usually the time from when I read one of your posts to when I start playing around with it where I miss putting a comment.
The volatility vs. 200dma correlation ;aka clustering is the big message. I bought some SRS because of it some time ago to experiment. (2x REIT index ETF). It’s been a good decision.
Garch type models are interesting. Another interesting attribute of Vol is that it persists over weekends, indicating the market moving “information” reflected in Vol is constant over weekends. The clustering effect is pretty interesting especially when seen in the absence of an exogenous shock to the market. Vol does get a little over played as a predictor of direcational movement. The measurement does vary based on duration of the metric. Shorter interval equals higher vol. The interesting story is correlation spiking and clustering. That is where mean variance asset allocation gets knocked around. Damn those black swans