The Missing Piece for Quants

As quant shops struggle with complex and underperforming factor-based investments, a new study offers a simple way to create an outperforming portfolio.

Illustration by II

Illustration by II

Managers who choose stocks based on characteristics like value, size, and profitability are struggling. New research claims the problem is all about implementation — and that a simple formula can solve it.

“There is an increasing awareness of factor implementation. It’s an issue of some factors [value and small size] not having done well, managers not paying attention to how factors interact, and using rules of thumb to make decisions,” said Steven Thorley, a finance professor at Brigham Young University. Thorley, alongside co-authors were Roger Clarke and Harindra de Silva, analyzed the best way to combine investment factors like value and momentum for a new paper, “Risk Management and Optimal Combination of Equity Market Factors.” The paper is expected to be published Wednesday in the third quarter 2020 issue of the Financial Analysts Journal.

According to the study, a portfolio constructed using what’s called forecast risk management can deliver significant excess returns. In back-tests over 54 years, the research portfolio generated annualized returns of 10.79 percent, compared to vs. 7.77 percent for a similar portfolio, the authors said.

“When you construct a multi-factor portfolio it is a complicated process, with lots of decisions such as which stocks to include and how to weight them if they are included,” Thorley said in an interview with Institutional Investor. “Some of the methodologies are ad hoc, with arbitrary decisions being made. That just creates noise. Lately, those implementation issues have been a drag.”

He explained that one of the major contributions of the paper is an actual formula that determines the weights of an optimal multi-factor portfolios with risk management. “There are new methodologies that we apply that make it simple and straightforward,” he said.

The paper relies on some recent innovations in portfolio construction, including so-called pure factors — versions of factors that don’t overlap with each other. That means a value factor, for example, is neutral to other factors, such as size. Thorley and his co-authors first published research on pure factors 2017 in the Journal of Portfolio Management.

Their new research also looks at performance from 1966 to 2019 of the five factors (value, momentum, small size, low beta, and profitability), using the pure factor methodology and adjusting for risk. It shows that the performance of the value factor, for instance, has essentially been flat for 20 years, far longer than many experts note. The good news, for factor enthusiasts, is that low volatility and profitability have done well.

[II Deep Dive: Why Value Investing Sucks]

“Value has had flat performance since about the turn of the century,” Thorley said. “It did great in 70s, 80s, 90s, but it really has done almost nothing for 20 years. Then on top of that, in this Covid-19 market drop, and partial recovery, value has had even worse performance. It was already under siege and suspicion, now this may be the nail in the coffin of value factor.”

As for the paper’s larger findings, Thorley was hopeful that asset managers would adopt the new methodology. “People are looking for some answers,” he said. “That might be enough to make its adoption more widespread.”