ebook essays pieces of the puzzle
Tuesday, June 5th, 2012
correlation roller coasters

One of the sponsored sessions at the CFA annual conference was presented by Axioma, about how to survive the “correlation roller coaster.”  It chiefly concerned the degree of correlation among stocks in an index — and you can find some of Axioma’s conclusions about that below.

I borrowed Axioma’s imagery for the title of this chart, because it reinforces the notion that correlations fluctuate, sometimes dramatically.  Our models (think of that staple of most financial planning, the mean-variance model) and our perceptions often fool us into thinking that’s not the case.

The chart shows 26-week rolling correlations with the S&P 500.  Top panel:  As you’d expect, the small-cap S&P 600 has a very high correlation with bigger stocks, but it did dip a couple of times in the 1990s.  And look at how the correlation with non-U.S. stocks has marched higher.  (So much for diversification benefits.)  Middle:  Correlations for these key commodities have really looked like roller coasters, although they have generally moved higher as well.  Bottom:  We are in a vastly different regime as to stocks and bonds than we were in the 1990s.

The general question is, of course, what are your beliefs about correlation and how do they affect your investment decision making?

While stated in regards to the correlations among stocks, many of the points made by Axioma apply more broadly as well.  The importance of correlation in the scheme of things varies significantly by investment strategy, even within an asset class.  Correlation effects are usually swamped by volatility effects, and the dispersion caused by the interplay of the two dictates much of what drives the alpha of various strategies.  While long-term correlations move quite a bit, in the short term they are more predictable, so strategies can be adjusted in response in ways that minimize periods of underperformance.

A couple of other interesting points:  All hedge fund categories other than short funds do worse in high correlation markets.  For active equity managers, periods when correlations are high and volatility is low are the “danger zone,” while low correlation and high volatility provide the dispersion that allows good managers to outperform.  (Chart:  Bloomberg terminal.  For more on Axioma’s work, here’s its latest quarterly risk review.)

See the rest of the series about the CFA conference in this PDF.

investment conversations

Three questions summarize the recent postings on the research puzzle.  What conversations drive your investment decision making?  When talking to investment clients and prospects, are you marginally less oblivious than the next person?  Do you understand the network dynamics that feed your world of information?