In a bull market, hedge funds run high net exposures, and might hedge using any old index future, never expecting to make much on a hedging policy outside of alpha from its stock specific shorts and merely hoping just to break-even from its short-side hedges.

In the current volatile markets (and letÆs remember that hedge fund managers originally wanted these rough seas in order to prove their mettle), the specific hedge selected becomes important. Even breaking even on a particular hedge is not necessarily easy.

Comparing the effectiveness of hedges is a task that has received attention in North America, but has yet to gain traction in Asia. Lehman Brothers has recently shipped in Ryan Nelson of its equity synthetics business in New York to tackle this problem with AsiaÆs hedge funds.

Lehman has brought a machine to Asia with the artificial intelligence that can figure out which of the many hedges on offer is most correlated to a long position.

"These types of models are being used to create customised hedging strategies in the US and Europe, but they are relatively new here," says Nelson, head of equity synthetics structuring for Lehman Brothers in Asia-Pacific. "Our hedging strategies are of great value to our clients, especially in today's markets, because we can help them understand the sources of risk to be hedged in their portfolios."

The system, called 'customised hedging solutions', employs quant models to pick the best hedge, even if the relationships are indirect, such as for example, by trying to hedge a technically unshortable Shanghai A-share portfolio via an H-share stock basket listed on the Hong Kong bourse. Lying behind all this is a gigantic suite of analytics, called WebBench, that works out the correlation of the hedge, plus via a back-test, the amount of hedging income it would generate compared to the long position.

The portfolio manager can then choose which hedge he likes best, and if he wants to execute, click on an icon that handles the trade. ThereÆs no extra cost for the analytics, the fees are just the normal agency brokerage fee plus the stock borrow cost for the hedge.

There are lots of ways of hedging any position, but they arenÆt all equal. Some correlate well, others might look satisfactory, but then when the sums are added up at monthÆs end, the offsetting profit from the hedge might fall disappointingly short of the losses from the long position. Some æhedgeÆ that was. In current market conditions, pre-trade hedging analytics could become a whole new career industry.