How To Exploit The Passive Herd With One Simple Strategy

If you frequent these pages, you’re well acquainted with the concept of passive investing and its rise to prominence in a world dominated by the proverbial “rising tide that lifts all boats.”

Well it stands to reason that if passive models are forced by … well… by the fact that they’re model-driven, to rebalance at predictable times, there will be opportunities to trade off that rebalancing.

If that sounds too good to be true, that’s because it’s… not?

Whether or not you have the capacity to actually trade on the following or not, it’s still an interesting read and says something about the extent to which the epochal shift towards passive creates opportunities for those with the wherewithal to exploit the predictable (note the bit about “…the strength and persistence of the signals suggesting there’s real information content here”).

Via Bloomberg contributor Cameron Crise

Given the ascendancy of passive investing over active management these days, there’s nothing I love more than a good fixing model.

  • NOTE: Fixing models seek to extract alpha from the predictable herding behavior of passive investors, typically around fixed calendar events like month-end rebalancing. While currency hedge adjustments are a popular class of fixing model, probably the most notable is the monthly equity/bond re-allocation from pension funds.
  • JPMorgan’s Marko Kolanovic and Bram Kaplan wrote a note on this a few years ago. I updated their work to see if there was still value in trading month-end flows.
  • The idea behind the model is a simple one. If stocks outperform bonds, by the end of the month pension funds will be overweight versus benchmarks and need to sell stocks and buy bonds to return to the target asset allocation. If bonds outperform, the same principle applies in the opposite direction. 
  • Using the S&P 500 Total Return Index and the Bloomberg Barclays Treasury Total Return Index, I modeled month-to-date returns since 1994. With five business days to go in each month, I determined which asset (stocks or bonds) had performed better on a month-to-date basis and assumed that pension funds would sell that asset class to buy the underperformer. I then tracked the performance of going long the underperformer and short the outperformer until the end of the month, at which point the trade was closed. 
  • It turns out there is quite a bit of alpha in these signals. While the performance of the strategy dipped soon after JPMorgan published the note, over the past couple of years the returns have been solid, particularly in bond markets.
  • The performance has been pretty good for some time now. (It yielded negligible returns in the 1990s.) The model also works over different time frames; I tested trading seven days and three days before month end. Each were profitable, though not as good as trading five days before.
  • Note that over the past five years, the information ratio of the bond model has exceeded 1.0; that’s not too shabby for such a simple signal. These results may not reflect real-world performance (there are no transaction costs included), but the strength and persistence of the signals suggests that there is real information content there. 
  • Although I cannot provide investment advice, I observe that we are five business days before month-end. As of last night’s close, bonds have outperformed stocks this month according to my input indices. Do with that information what you will.
  • And if you are an investor that engages in periodic asset-allocation rebalancing, consider an alternative to doing so at month-end.



Speak On It