Covid19’s spread is a new market event that has prompted a wave of quantitative investors to pull out of their positions and retrain their models, signalling the start of a new quant pattern, confirming adaptability is still an investor’s real edge.
“When a regime changes so much, you need to give your models time to adapt,” Stuart Bohart (pictured), managing director of FORT LP, said in an interview with Investment Magazine’s Market Narratives.
“That adaptability needs to be systemic, because it’s not your ability to spot disruption that counts, it’s your ability to recover from that disruption.”
Quantitative investors, or those using historical data to model the probability of what will happen in the future, have spent this year retraining their models to understand how markets react during a global pandemic.
“In March, for a brief period of time everything was for sale,” Bohart said. “Gold went down, stocks went down, Treasuries went down, if it had a price and could be turned into cash, it went down.”
FORT partner and co-founder Sanjiv Kumar will appear at Investment Magazine’s upcoming Absolute Returns Conference September 23-24 on the topic ‘Machine learning and improving the investment decision making process’; reserve your digital access to the conference here.
The dramatic selloff followed a similar pattern to what happened in the Global Financial Crisis in 2008, however the wider economic shock of the economy’s abrupt closure meant quantitative investors couldn’t just rely on historical data to spit out the probable market reaction.
“You need to discover whether your signals are still valid, whether the models have learned and adapted, and whether the parameters have changed,” Bohard said.
“If something doesn’t fit past patterns like what happened in March, you need to pullback and de-risk, learn from it, and then push back in, and re-risk.”
The flood of monetary stimulus into the markets, and the steady rise of large index positions, are market dynamics Bohart’s quant fund have had plenty of time to model.
“The issue of the Fed coming in and suppressing volatility, and taking away the downside, we can capture that in our quant models because it’s happened before,” Bohart said.
“We’ve got an idea of what massive liquidity means for models.”
But Bohart doesn’t necessarily buy-in to the hedge fund complaint that cheap money, indexing, and ETFs that are insensitive to price, have ruined opportunities for long/short investors.
“If I was a discretionary investor and had the opportunity to participate in a highly efficient market or in one with indiscriminate buyers and sellers creating an inefficient market, I would choose the inefficient market,” Bohart said, adding inefficiencies over time should drive performance and the problem for short term investors is the time frame which might be longer than they would like.
“When money is free, it’s much easier to keep a poor company running,” he said.
“That might be why short term investors have had a hard time making money in their portfolio picks because cheap money delays the inevitable,” he said.
The understanding that the Fed will continue to support markets through the Covid19 pandemic has seen a wave of investors hoping to capitalise on that momentum, but Bohart says momentum is complex to incorporate into quant strategies.
Quant investors will need to consider variables like their look back periods, the degrees of momentum that trigger their buys, what kinds of stops they have, and whether they have an overlay of mean reversion.
“The trick isn’t identifying momentum, it’s understanding how you identified it, how you captured it, how you managed it, and that comes down to the person executing on the strategy,” Bohart said.
“There are quants who have momentum as part of their strategies and their returns are divergent.”
Working with machines and data has shown Bohart how dogmatic views on the market, whether they are around valuations, earnings in stocks, or GARP, mean investors often fail to adapt to abrupt changes like Covid19.
“If you’ve defined the world and you’re puzzled when it doesn’t fit your definition, you’re likely losing a lot of money,” said Bohart.
“Quants don’t generally have those dogmatic views, because we aren’t coming up with our own views, we just look at data and see whether in these circumstances an outcome is probable, not even likely, just more than fifty-fifty.”
Bohat says one of the arts of quant investing is determining which of your signals are producing better probabilities than others.
For a firm dedicated to sifting through and understanding data streams, Bohart insists the edge in investing still lies with the human beings making the calls, not the machines.
“It’s not the data that leads to the alpha,” he said.
“People talk about alternative data as if it’s a money making machine, but it’s your ability to adapt and use some form of artificial intelligence to map out and change your predictions and parameters that will give you the edge.
“Human intuition leads you to probe and model and really think about what’s going on.”