Traders and regulators would find it useful to be able to measure overall market leverage. So it was with some excitement that I received the latest research report from the quantitative investor solutions (QIS) group at Citi. As the report says in the introduction: “To have a measure of this overall leverage, and to be able to observe its variation through time, would be a useful indication both of market sentiment and general market exposure and sensitivity.”
In Calculating dynamic leverage in FX trading, Citi compares various currency trading indices with a basket of its own unleveraged benchmark beta 1 indices made up of G10 carry, emerging market carry, valuation and trend. Citi then adjusts composition and leverage of the beta 1 basket in order to replicate the volatility of the comparison indices.
But calculation of a leverage factor has some difficult problems to surmount: currency trading indices can differ considerably; and there’s some concern from currency funds about which indices their investors require them to be measured against. This is a subject we will come back to.
A similar problem arises when comparing a beta 1 portfolio with an established index – which is the best barometer of market activity? And a further problem is the dynamic reallocation of leverage between strategies in the market. The mix of strategies in portfolios constantly changes. Saying that, between 2002 and 2007 the mix might have been quite stable – carry was such a win-win proposition that it must have consumed most of the leverage available, but that probably doesn’t hold true after ’07.
Unfortunately, as I understand the report, while it is possible to create the shape of the dynamic leverage curve, it is difficult to extract the magnitude of the leverage. The QIS people won’t need me to point out the parallels with quantum uncertainty.
Citi’s QIS team hasn’t given up trying to calculate leverage and the work done so far enables some conclusions to be drawn. I noticed from the report that volatility for the unleveraged beta 1 portfolio and the CISDM and Barclays indices converged during 2009. I asked Kristjan Kasikov, one of the authors, if this meant that FX funds had, at that time, completely deleveraged. He replied: “Yes, the volatilities of the Barclay Currency Traders Index, CISDM index and HFRX currency hedge funds all dropped to record lows in 2009. The similarity in vol between the unlevered Citi beta and the indices suggests that leverage in the market was probably close to one.”
That was the shake-out complete then: the relative volatilities have since parted.