If factors can explain, almost everything and then flip, or stop to work and alpha that was there yesterday suddenly vanishes, there is only up to some time you can play this magic trick and keep laughing. Eventually, the reality that Alpha is a tragedy of commons, and just because it’s not my job or your job, doesn’t mean it will solve itself.
6-10 basis points fee for Smart Beta, 100-150 basis points fee for Active Management, factor timing, quantamental, or sentimental, if you can’t give an absolute return, all the remaining relative fees are going to zero. Charging for relative performance is old tech. The individual investor has access to every information. He/She/Her the investor, can do the fundamental debate with you, are capable to read academic papers, and are initiated. They can discern your capability. They already know, human alpha is in its last mile, they are the ones demanding zero fees, they are aware of the spectrum that stretches from “When Genius Failed” to “The Big Short” and they have been at value investing longer than you can imagine.
How do you get the new investor’s confidence back? How do you create the wow factor? So that He/She/Her can say, “I like this machine. I like this bot.” The future is of machine beta and it’s here to give relative outperformance, the holy grail of beating S&P500 to He/Him/His for free. All it has to do to be relevant is first do what less than 1% of global asset managers can do over 5 years, beat the S&P500 on a risk-weighted basis by 100 bps. Once it takes the margin to 300 basis points, it will compete with the 0.1% elite managers of the world, and then eventually, Gary Kasparov and Deep Blue will happen all over again. Machine dominance in asset management space.
Let’s imagine this machine. How will it look like? Will it be a supercomputer? I think it will be an app, on your phone, a regular app, somewhere in the cloud. Data and computation light, focussing on weights, thinking about the various permutation and combinations of weight schemes across 500 components, classifying them, anticipating them, and hence reaching a super optimal set of combinations to beat the S&P500. And since it is smart, it won’t be crunching all these combinations, it will be screening, eliminating combinations based on design, dynamics, structures, and substructures that will have nothing to do with informational content but just context, working under constraints, making sure the new S&P 500 has a similar risk, low tracking error, low turnover like the S&P500 and it's passive at heart.
We don’t need a sufficiently advance Science to do this magic.