Taking A Global Look at Risk and Correlations

Risk management 21 Feb 2012

Axioma, a research and analytics company, has published its “Quarterly Risk Review,” a reference work on the state of investment markets around the world drawing on data for the fourth quarter of 2011. The review is available in seven editions, distinguished by region: U.S., China, Australia, Europe, Asia Pacific ex-Japan, Emerging Markets, and Developed Markets.

Comparing these editions leaves some fascinating insights. For example: it is becoming more difficult over time, in much of the world, for investors to create significant diversification within the (domestic) equity portion of their portfolio, because the correlations of stock pairs have been increasing. China’s stocks are an exception to this rule.

Figure 14 in the U.S. edition of Axioma’s report on is quite striking. It shows the rise of both 20 day and 60 day correlations (blue line and red line below) from 1999 forward. Although the trend lines in both colors look like wild zigzags, the undeniable over-all movement is up and this was quite obviously the case for most of 2011.

The 20-day correlation line peaked in early 2010 near, though still below, 0.8. It immediately fell off that peak, all the way to 0.1. But it spent 2011 rising, getting above 0.8 in October. It began another decline late in the year

Axioma assures us, though, that the fact that stocks’ returns are more and more similar is “not necessarily a bad thing.” Indeed, since pairwise correlation almost hit 0.0 at the time of the dotcom bust, mid 2000, one would be hard put to it to believe that high correlation is necessarily bad or low correlation good. Still, the recent increase of correlation to stratospheric levels could not help but have complicated risk management.

As a broader observation: risk in the U.S. market through 2011 was significantly higher than it had been in the middle of 2009, even though it risk fell a bit toward the end of 2011.

Quarterly Risk: European edition.

Europe is now “front and center in discussions of global equity markets.” But Axioma indicates that a fever may have passed. Risks ended 2011 well below their peaks of the year.

Axioma uses four distinct models of risk: medium-term fundamental; medium-term statistical; short-term fundamental; short-term statistical. Figure 2 in the European edition shows the level of risk, given all four models, both over a five-year stretch and over the final quarter of 2011. In the five-year chart, there was a collective peaking at the end of 2008 and early 2009, a second (much lower) peak in mid 2010, then a decline into 2011. But there was a sharp rise in risk, measured by any of these models, in the third quarter of 2011, reaching above the 2010 peak, then a slight falling off.

In the second graph within Figure 2, we see the details of that falling off. In the final days of the year, the lines representing the two middle-term models began showing greater risk than the two short horizon models.

The authors of the Europe report side with the more optimistic side of this end-of-year split. They write, “If risk is changing rapidly, it may be picked up sooner by our short-horizon models,” so if the trend perseveres, “we may see lower risk in developed Europe this year than we have seen in a while.”

In Europe as in the United States, “stock-stock correlations rose substantially throughout 2011 and near record highs,” though in each case also they “eased downward” by year’s end.

What about China?

Compare China’s “realized pairwise asset correlations” to those of the United States. For the period 1999 to 2011, the ups and downs of China’s correlations seem to add up to … nothing. Almost all of the movement of those two lines (for the 20-day and the 60-day correlation) has been within a range from 0.2 to 0.6, and both lines remained comfortably within that range at the end of 2011. This contrasts with the U.S. graph of the same period, which as seen above shows a constant expansion of the range, and a general move toward greater correlation.

Axioma concludes, “The risk profile of China was somewhat different from what we observed in the rest of the world, but … opportunities to use risk models as a key tool for managing portfolios” are similar.

Be Sociable, Share!

2 Comments

Leave A Reply

← American Healthcare: Under All the Noise, Psilos Finds Cost-Effective Solutions The Volcker Rule: Return of the Fabulous Fab →