OppenheimerFunds’ Head of Product for Beta Solutions, Mo Haghbin, recently spoke with Christopher Polk, a Professor of Finance at the London School of Economics, about factor-based strategies and the future of quantitative investing.

Q: There is debate in academic circles that factors are just capturing some non-transparent element of systematic risk. How can researchers and practitioners determine if strategies might be capturing a higher moment like jump risk?

A: Academics and practitioners should take that possibility seriously. The traditional measure of risk is market beta, how the return on an asset moves with the market return. I have a recent publication in the Journal of Financial Economics that shows that if markets are dynamic and investors have long horizons, other measures of risk become important both conceptually and empirically.1 In particular, my paper argues that sensitivity to news about aggregate volatility can explain the premia on some factors. For example, value stocks do relatively worse than growth stocks when there is news that aggregate volatility will be higher in the future. This fact makes value stocks riskier to a long-horizon investor as those stocks do worse when the attractiveness of the investing environment is deteriorating.


Q: How should investors think about whether a factor does or doesn’t work? Can you elaborate on some of the challenges with financial research (e.g., p-hacking, coincident vs. casual results)?

A: Good research is the same in both academia and practice. Such research should couple a careful use of the scientific method with a compelling explanation as to why patterns being explored should exist. Unfortunately, not all research is held to that high standard.


Q: Does factor investing work as well in developed markets outside the United States and in emerging markets?

A: Broadly speaking, factor investing works well in all markets. Indeed, the average performance of factors usually increases as markets become less developed. I have a recent work-in-progress measuring the performance of factor investing in the Chinese stock market. Most of the equity factors that have been verified as important in developed markets also work very well in China. Of course, transaction costs and investing constraints are significantly higher than in developed markets, and one might think those elements play a role in the stronger performance I find. My broader point, however, is that factors are pervasive.

I am also bullish on more sophisticated implementations of quantitative investing. New sources of data and the power of information technology will facilitate a new generation of quantitative strategies. Of course, the principles of good research must still apply.


Q: We frequently hear about the debate on tilting versus timing factors. Is there a difference?

A: My interpretation is that any distinction between the two is mostly one of semantics with tilting involving less aggressive portfolio dynamics. Factor timing suggests that there may be times when one bets against a factor which certainly involves stronger views.


Q: Systematic strategies have started to become somewhat mainstream. What are your thoughts about the future of quant investing?

A: Quant investing has a bright future. For one thing, quantitative strategies such as the ones found in multi-factor ETFs offer opportunities for better benchmarking of our asset managers in terms of identifying exactly what exposure they are providing and at what cost. Also, the specific implementation details of quantitative strategies matter; those institutions with better craft will excel. And as we have already touched upon, combining exposures when factor premia and return volatility vary through time requires significant skill.


Q: What are some of the newer signals that you find particularly useful? Any exciting new research you can share?

A: I am very excited about my research on connected stocks that was published in the Journal of Finance a while back.2 That piece argues that stocks are connected by institutional ownership, which causes them to move together for non-fundamental reasons. Consider two unrelated companies held by many common institutional owners. If bad news strikes one firm in the pair, the second firm will have its price temporarily pushed down as those common owners sell off stocks in their portfolio to meet redemptions. Of course, that temporary dislocation is an opportunity for a savvy, unconstrained investor to provide liquidity and profit. Indeed, a connected stock trading strategy based on my research generates significant abnormal returns of more than 9% per year, controlling for the standard factors studied in markets.

Christopher Polk is the Head of Department and Professor of Finance at the London School of Economics. Prior to the LSE, Polk taught at Northwestern University’s Kellogg School of Management; he has also been a visiting Professor of Economics at Harvard University and a visiting Professor of Finance at the MIT Sloan School of Management. Polk’s research interests are in asset pricing and include related topics in asset management, corporate finance, behavioral finance, and macroeconomics.

  1. ^Campbell, John, Stefano Giglio, Christopher Polk, and Robert Turley, 2018, “An Intertemporal CAPM with Stochastic Volatility,” Journal of Financial Economics 128, 207-233.
  2. ^Anton, Miguel and Christopher Polk, 2014, “Connected Stocks,” Journal of Finance 69, 1099-1127.