Comments on: How correlations change

By: Th.M

Th.M — Mon, 23 Apr 2012 09:39:42 +0000

This video (made in 2011, and with fewer assets) shows the dynamics; the risk-off phenomenon is particularly clear during crisis, when the whole matrix turns red.

http://www.youtube.com/watch?v=LBO2bYH_K eA

By: BrPH

BrPH — Fri, 20 Apr 2012 16:44:12 +0000

What exactly is meant by “correlation” in this case?

By: MrRFox

MrRFox — Fri, 20 Apr 2012 15:22:47 +0000

Of all the asset classes on the chart, I believe gold has been the best performer over the period, yet it is neither risk-on nor risk-off, but middle of the pack on both lists. Same is largely true of all the “real world” stuff and commod-currencies.

It’s the financial stuff that is at the extremes, and where the changes in rankings take place – the $ being the most striking example.

So, what’s the lesson for us from those 5 tumultuous years? How about – financial engineering is a dangerous thing to get mixed-up in?

By: M11

M11 — Fri, 20 Apr 2012 14:27:45 +0000

(In my 2nd point I was referring to the overall amount of correlation among assets, not the ranking)

By: M11

M11 — Fri, 20 Apr 2012 14:24:39 +0000

I would say that this has a lot to do with the spread of statistical arbitrage trading.
Also, comparing those correlations calculated using some simple Pearson coefficient might be a complete fallacy. There could have been entirely same inherent correlations with just a bit noisier sampling representation (due to less amount of automated or just lower frequency trading), which is not that easy to filter out and simple correlation measures fail to do that.