Here's a statement of the obvious: The opinions expressed here are those of the participants, not those of the Mutual Fund Observer. We cannot vouch for the accuracy or appropriateness of any of it, though we do encourage civility and good humor.
Support MFO
Donate through PayPal
Ben Carlson: Peer Preasure: Thoughts On David Swensen Interview
(This is a follow-up to the David Swensen's interview posted by bee below.) Yale Endowment CIO David Swensen doesn’t make too many public appearances so I was excited to run across a back and forth he had with Robert Rubin this past week at the Stephen C. Freidheim Symposium on Global Economics. Regards, Ted http://awealthofcommonsense.com/2017/11/swensen/
In this part of the Swenson interview, reprinted in the article above, I was a little surprised at these comments:
The rise of quantitative investing adds another wrinkle to this equation but many in the institutional world rely more on black box quant funds than the transparent quant approaches that are taking over the ETF and mutual fund world. Here’s Swensen’s take on the problem of dealing with black box quant funds as an investor:
"You know, I have never been a big fan of quantitative approaches to investment. And the fundamental reason is that I can’t understand what’s in the black box. And if I don’t know what’s in the black box, and there’s underperformance, I don’t know if the black box is broken or if it’s out of favor. And if it’s broken, you want to stop. And if it’s out of favor, you want to increase your exposure."
"And so I’m an old-fashioned guy that wants to sit across the table from somebody who’s done the analysis and understand why they own the position. And then if it goes against them, I can have another conversation and try and figure out whether the thesis was wrong and we should exit, or whether the thesis is intact and we should increase the position. And I don’t understand any other way to invest."
Swenson was around when LTCM blew up, and it seems his views may be colored by that experience (see his repeated use of the phrase "black box").
Algos are not complicated. They consist of four basic parts:
1. Define the universe from which you are selecting securities 2. Screen down the universe with "buy rules" 3. Rank the remaining securities and invest in the Top XYZ in the ranking list 4. Build some "sell rules" that tell you when to liquidate a position
With the rankings and buy rules (which anyone would share with Swensen to get a piece of the Yale endowment), it's pretty clear what factor(s) the algo is trying to exploit. And if you know that, then you know when it underperforms whether it's out of favor or broken.
Comments
In this part of the Swenson interview, reprinted in the article above, I was a little surprised at these comments:
The rise of quantitative investing adds another wrinkle to this equation but many in the institutional world rely more on black box quant funds than the transparent quant approaches that are taking over the ETF and mutual fund world. Here’s Swensen’s take on the problem of dealing with black box quant funds as an investor:
"You know, I have never been a big fan of quantitative approaches to investment. And the fundamental reason is that I can’t understand what’s in the black box. And if I don’t know what’s in the black box, and there’s underperformance, I don’t know if the black box is broken or if it’s out of favor. And if it’s broken, you want to stop. And if it’s out of favor, you want to increase your exposure."
"And so I’m an old-fashioned guy that wants to sit across the table from somebody who’s done the analysis and understand why they own the position. And then if it goes against them, I can have another conversation and try and figure out whether the thesis was wrong and we should exit, or whether the thesis is intact and we should increase the position. And I don’t understand any other way to invest."
Swenson was around when LTCM blew up, and it seems his views may be colored by that experience (see his repeated use of the phrase "black box").
Algos are not complicated. They consist of four basic parts:
1. Define the universe from which you are selecting securities
2. Screen down the universe with "buy rules"
3. Rank the remaining securities and invest in the Top XYZ in the ranking list
4. Build some "sell rules" that tell you when to liquidate a position
With the rankings and buy rules (which anyone would share with Swensen to get a piece of the Yale endowment), it's pretty clear what factor(s) the algo is trying to exploit. And if you know that, then you know when it underperforms whether it's out of favor or broken.
This is not rocket science.