Quantitative investing needs to change, and should do so by scaling up to produce more proprietary data, reducing excessive numbers of signals and becoming more “market savvy”, according to the global head of equity research at BlackRock, Ronald Kahn.
Mindful of the terrible press that quants received when most practitioners recorded substantial negative returns in August 2007, Kahn now seeks to differentiate ‘scientific investing’ from ‘quantitative investing’.
He laments that the latter has come to mean “optimising portfolios with forecast returns proportional to a few well-known, publicly available financial ratios – book-to-price, earnings-to-price, price momentum and analyst estimate revisions…in its worst implementations, it mindlessly searches for patterns in historical data, to extrapolate into the future.”
The simultaneous collapse of these four standard quant signals in August 2007 is shown in the graph.
That’s not to say all generic signals should be ignored – Kahn points out that book-to-price was a great predictor of global market recovery in March 2009 – but he believes that ‘scientific investing’, as a superior sub-set of quant, should focus on ”identifying new investment ideas and continually improving their implementation”.
However, a “new idea” should not be confused with “just another signal that captures a value premium in a slightly different way to all the others”, Kahn says.
“Scientific investing is not just about maths, you know, inverting a matrices. If algebra could be converted into alpha, quants would always outperform because we can all do it. The key is coming up with ideas, grounded in economic sensibility, and running a variety of empirical and analytical tests against them.”
Economic sensibility was a priority for Kahn’s ‘Scientific Active Equity’ team at Barclays Global Investors, long before it became a part of BlackRock following the big merger last June.
A classic example of a new idea which “grew from a hypothesis grounded in economic sensibility”, according to Kahn, was a ‘quality of earnings’ signal which broke up a company’s reported earnings into a ‘cashflow’ piece and an ‘accruals’ piece.
“Richard Sloan had done some great work on this in 1996, yet everyone but us was ignoring it, and looking at earnings in totality.”
Sloan had shown that the higher the proportion of the cash component of earnings to the accrual component, then the greater was the persistence of earnings performance.
Economic sensibility is one thing, however the quantitative manager performance crisis, from which Barclays/BlackRock was not immune, had shown that it needed to be accompanied by market savvy.
“You need to be aware of the prevailing market environment and whether it supports the ideas you’ve got,” Kahn said.
Any signal tied to analysts’ revisions, for example, needed to recognise that sell-siders were “slow to update their expectations” in more volatile markets such as those recently experienced.
“The classic example was two days before Lehman Brothers collapsed, the analysts revised down their financial year one estimates for Lehman earnings – but not for financial year two”.
Kahn said the BGI merger with BlackRock had helped his scientific team gain this vital market savvy, encouraging interaction with fundamental analysts and broadening perspectives.
“Quants are no different from any other investor, in that in order to model a particular company’s future earnings, you also have to model its customers and competitors around the world,” he said.