Secrets of quant investing

By Dean DiSpalatro | June 27, 2013 | Last updated on June 27, 2013
3 min read

Quantitative investing has gotten a lot of bad press in the past few years. An especially egregious example of quant gone wrong was the May 6, 2010 flash crash that saw the Dow Jones Industrial Average tank almost 1,000 points intraday.

Read: What causes flash crashes? People!

But don’t view all quantitative investing in the same light, says James O’Shaughnessy, chairman and CEO of O’Shaughnessy Asset Management (OSAM). His firm sub-advises seven funds for RBC Global Asset Management, which owns 10% of OSAM.

There are three main types of quantitative investing. Algorithm-based high-frequency trading is the kind that grabs headlines. Not only can it cause major market disruptions, it also creates an uneven playing field, says O’Shaughnessy.

A recent Wall Street Journal report says high-speed traders on the Chicago Mercantile Exchange are raking in profits thanks to their ability to see which direction the market is moving a split second before anyone else.

It’s not all bad, however. High-frequency trading “provides liquidity the marketplace might not [otherwise] have,” O’Shaughnessy says.

Read:

High-speed trading un-Canadian, says NY Times

Concerns about high-frequency trading

Will RCMP monitor electronic trades?

The second type “is what the propeller beanie guys are doing. They’re basing everything on physics and disciplines that have little to do with investing. If you can find a pattern using your MIT degree in physics and particle behaviour, God bless you,” he says.

O’Shaughnessy advocates the third: a model-driven approach that combines elements of active and passive investing.

On the active side, it focuses on company fundamentals like earnings quality and growth. “We then passively implement the securities that meet our screening process to completely remove emotion from the equation,” says O’Shaughnessy.

“We agree with the passive argument that says ‘no overrides.’ What passive investing has wrong is [it doesn’t acknowledge] there are many strategies that do significantly better on a risk-adjusted [basis]. You just have to let them do their work,” he explains. “Never once have we overridden an investment strategy.”

Read: Harnessing emotion gives investors an edge

He emphasizes the importance of going where the evidence takes you.

“We wanted to see whether buying stocks that had the greatest increase in dividend yield was a better strategy than just buying stocks with the highest dividends. On the face of it, you’d think a company that’s increasing its dividend yield year in, year out is showing strength. But the data suggests otherwise: they underperform the market by about 100 bps and do significantly worse than the highest dividend stocks,” O’Shaughnessy explains.

Read: Keep emotion in check, portfolio balanced

His models draw data from a wide range of sources. “We scrub the data by comparing the various sources. Say Compustat has a stock’s earnings at $1.50, and Worldscope has them at $2.50. This doesn’t happen often, but when it does it’s automatically flagged. We then have one of our analysts compare it against the Bloomberg data. If we have three separate stories and there’s no logical reason for it, we exclude the company.”

Fundamental analysts often make company visits a cornerstone of their evaluation process. But O’Shaughnessy never does them, because he prefers to look at the hard data.

Dean DiSpalatro