You might not be as diversified as you thought you were

Sometimes the returns of the equity and debt markets move in a similar direction, and sometimes they don't, according to a new academic paper.

While the findings appear simple, they could be extremely valuable to plan sponsors when drawing up asset allocation models for their funds. That's because most plan sponsors now assume a static correlation between the equity and debt markets when constructing an asset allocation, and do not take in to account that the correlation between the two changes with various market conditions.

Put simply, because the correlation between equity and debt changes over time, there could be periods when the plan sponsors' asset allocations are not as diversified as they originally expected.

The paper, "The Relation Between Fixed Income and Equity Return Factors," was written by Terry Marsh, a professor emeritus of finance at the University of California at Berkeley, and Paul Pfleiderer, Miller Distinguished Professor of Finance at Stanford University. They found that the correlation between equity and debt fell to about -0.4 in 2002, from about 0.5 in 1995. Recently, the correlation has been nearly zero.

Correlation is the statistical measure of the degree to which the returns of two investments move in tandem, and is expressed as a number between one and negative one. If the correlation of the returns of two securities is 0.5 over a given time frame, for example, that would mean that the returns of the two securities moved in the same direction better then 50% of the time.

Additionally, the paper found that significant jumps in the correlation between the equity and debt markets were associated with unexpected "flight-to-quality" conditions, including changes in inflation and interest rates, as well as market volatility and changes in the overall duration of U.S. Treasury securities.

The paper examines the returns of Treasury bonds across all durations, as well as the returns of the U.S. stock market, as measured in a database run by Quantal International Inc., a financial technology firm in Berkeley, Calif. The returns of both asset classes are for the period from Jan. 9, 1996 to July 1, 2005.

"(The paper) is just the beginning of a line of research that digs in to the question of what are the underlying drivers that jointly affect the returns of the equity and debt markets," said Gifford Fong, president of Gifford Fong Associates, Lafayette, Calif., and editor of the Journal of Investment Management. The paper is being considered for publication by the journal.

"The vast majority of plan sponsors assume a standard input of about a 0.3 correlation (between the equity and debt markets) when they set an asset allocation," said Mr. Fong. "That standard input is not appropriate because the correlation is nowhere near stable at 0.3."

Shiv Mehta, a senior vice president and head of asset allocation research at ING Investment Management, New York, agreed with Mr. Fong and said plan sponsors should use various correlation estimations.

"I think this paper addresses in a very eloquent way the question of whether correlations vary. I would agree that they change (with market conditions)," said Mr. Mehta. "A (static) correlation measure doesn't capture correlations during high-volatility periods. You need to capture that because that is precisely when you need diversification. The bottom line is that plan sponsors need to factor in more than one (correlation assumption)."

Some managers said the paper indirectly makes a case for tactical and global tactical asset allocation funds, since TAA and GTAA strategies have the ability to quickly move in and out of asset classes via derivatives. Therefore, when there are significant changes in the correlation between equity and debt, TAA and GTAA managers can more easily adjust the asset allocations of their clients' portfolios.

"The benefits of diversification is greater if markets are less correlated, and it's easier for a TAA strategy to shift between asset classes," said Jeff Knight, a managing director and head of global tactical asset allocation strategies at Putnam Investments, Boston.

According to Max Darnell, partner and chief investment officer at First Quadrant LLC, a GTAA firm based in Pasadena, Calif., assuming a static correlation between asset classes does not account for "shock" events.

"The paper is far more applicable to tactical approaches," said Mr. Darnell. "When you set an asset allocation mix based on correlation assumptions and return assumptions, one of things you could miss are the shocks that create changes in the correlation structures. For example, in 1998 there was the collapse of the Russian ruble and the fall of Long Term Capital Management that created a flight to quality. There are sudden flights to quality that do affect ‘tail' events that can shift correlations significantly."