Exploring 'systematic macro' as a hedge fund strategy

AsianInvestor has written about discretionary macro as a hedge fund strategy. That’s where a very clever economics-trained person is able to think through all salient issues and formulate a trading strategy in their own noggin.

AsianInvestor has written about managed futures and CTA funds. That’s where a computer works out trends and momentums across a myriad markets and trades on them.

However, what AsianInvestor has never written about before is systematic macro. That's a hedge fund strategy hybrid combining both human and model – where blackbox meets cheesebox.

Dr Sushil Wadhwani, of Wadhwani Asset Management, has been operating a quant fund-style strategy since 2005, and he recently visited Hong Kong to talk about this strategy.

He has partnered with fund manager GAM to create a Ucits 3 vehicle in which GAM is the investment manager, and he is the delegate investment adviser. For compliance reasons, he does not want to name the fund or discuss performance. The fund launched in April 2010 and invests in equities and fixed income.

Asian investors have a long tradition of investing with CTAs. Systematic macro is a related strategy, but it is not the same thing. It does have some things in common, such as the diversification benefit and the ability to do well in higher volatility environments.

A common complaint about trend-following CTAs is that they tend to have big drawdowns and are not sufficiently agile around significant turning points. A systematic macro approach uses momentum strategies like a CTA does, but combines that with a variety of macro information, with the aim of doing better around those turning points and playing a much better defence.

“I have a ‘box’ that uses information a CTA might use, such as price volume and open interest,” says Dr Wadhwani. “My box also uses as inputs a variety of information on macroeconomic fundamentals, using a satellite model that produces, say, a forecast for growth or inflation, by country and for the world.  It then compares those to consensus forecasts and looks for surprises. I’ve been working on these models for twenty years.”

In April 2010, his models told him that economic growth was set to slow, especially in the US, so he used that as an input into the models, to go against the trend-trade which was to be long equities. The box then told him it might be better not to do the momentum trade, but to go the other way.

So in May, when the correction came, trend-followers lost money. If they had factored in some macroeconomic inputs they might have done better.

Sometimes the inputs are proven right, and sometimes wrong, but he finds that if you use enough of them and they are sufficiently diversifying, then performance will be better than just relying on momentum.

“We produce forecasts based on historic data. Our model is predictive,” he explains. “In addition we use sentiment data, derived from the options market or from surveys. It is useful around turning points, and we’ve found that in a variety of asset markets and a variety of countries.

"If you have a set of investors with a persistent bias, you find that their optimism or pessimism still varies over time. By focusing on the rate of change in their view and that movement over time, you can effectively strip out their bias. There’s no linear relationship nevertheless between sentiment and future asset prices.”

His approach also uses predictive valuation models. He gives the example of March 2009 when the S&P stood at 666, valuation models were screaming that stocks were a buy but most trend-following CTAs were still short the S&P because it was the momentum trade. If you were using valuation models, you would have been loading up then.

However, sometimes markets behave erratically, he finds, and have a tendency to overreact to certain events and data releases on the day itself.

“That gives an opportunity to fade that on subsequent days,” he says. “We have quant strategies that exploit that tendency which we build into our models. We've found instances where indicators are either of little use to us, or where you should be fading them. I prefer not to say exactly which ones they are, in case it offends someone. We would disclose it to our investors off the record though. “

Asian investors are familiar with both discretionary macro and trend-following CTAs. Dr Wadhwani’s experiences on this visit to Asia lead him to believe investors here seem intrigued in finding out more about systematic macro as a third strategy.