cAI23 - Causal Factor Investing

The Causal AI Conference 2023:
Causal Factor Investing
Marcos López de Prado
Visit www.causalaico... to learn more.
Virtually all journal articles in the factor investing literature make associational claims, instead of causal claims. The authors do not identify the causal graph consistent with the observed phenomenon, they justify their chosen model specification in terms of correlations, and they do not propose experiments for falsifying causal mechanisms. Absent a causal theory, their findings are likely false, due to rampant backtest overfitting and incorrect specification choices. I differentiate between type-A and type-B spurious claims and explain how both types prevent factor investing from advancing beyond its current phenomenological stage. This monograph analyzes the current state of causal confusion in the factor investing literature and proposes solutions with the potential to transform factor investing into a truly scientific discipline.

[@pedromacedo6202]

Amazing talk Señor Marcos López de Prado

[@JaazFelicio]

Tnks for sharing

[@Jay-eh2ch]

Great talk

[@antonisdee]

I just checked to see what the performance of a developed markets multi-factor UCITS ETF (JPGL) was for 2007-2022, and the result was 8.72% per year. The benchmark (IWDA) returned an average of 6.87% over the same time period. Just saying 🤷

[@Affepaul]

The fund was released in 2019, though. So most of the data you used is back-tested. Since inception JPGL underperformed IWDA. It might not be the counter argument it was intended to be.

[@antonisdee]

@@Affepaul All factor funds underperformed the last few years because factors other than the market one haven't done well during this period. The same thing happened in the late 90s, just before a decade of underperformance by cap weighted market funds.
Also, my initial comment wasn't meant to say that "factor funds performed better during x period so they will continue to do so". It was simply a response to the first graph in the presentation.

[@jonnyh.2167]

@@antonisdee having a great backtest and rolling “out-of-sample” performance while underperforming or not providing consistent alpha once after inception is exactly why he says many of the factor investing strategies aren’t scientific

[@jonnyh.2167]

⁠@@antonisdeeanyone can data mine a few factors that provide superior performance vs some benchmark over a certain period of history. Anyone can say a strategy is cyclical-- sometimes they work, sometimes they don’t. But is it scientific? And are the investors dumb enough to buy it?

[@q44444q]

​​@@jonnyh.2167 Neither is what he proposed, though. I mean, to be clear, he didn't actually propose anything, he just introduced some basic causal inference concepts, and complained about the state of econometrics 10+ years ago. But in finance, the causal graph changes so quickly and there are so many potential variable that the graph is almost pointless to estimate (if you even can estimate it--good luck! Causal Discovery algorithms aren't even good on simulated data!). Moreover, Marcos clearly doesn't understand what reflexivity means, even though he tried to cover that up by making some rough and incorrect analogy with science. Most frameworks for causal inference don't allow for reflexivity (technically, this corresponds to mixed direction causal graphs which may contain cycles and bidirective edges) which is another main problem in their use in investing... And finally with causal inference you also have a massive multiple testing problem, because as Marcos says at the end, you may have hundreds of graphs which are reasonable causal assumptions, then you need to test them all. Marcos is a smart guy but this causal factor investing stuff is mostly charlatanism in my opinion.

[@rafaelmastroberardino]

wait... this is a long short strategy that you talk about at the beginning. someone can select is factors preference and just go long and could have better results (still factor investing)

[@ДмитрийВербицкий-у7д]

Wilson Mary Thompson John Lewis Cynthia

[@annawilson3824]

why it was a physics lecture??

[@axe863]

Bro.... most quants know physics 😅

[@slapjuice]

This guy is so useless its a joke! He has no fund that shows performance, he states the most obvious thing that everyone knows. Then he thinks this is all science when he has no science background. He is a finance guy peddling more nonsense with zero performance. Simply put, if hes right, where is his Hedge Fund? He doesnt have one... Buy a passive tracker and you will out perform all of these people.

[@tuvshoo9992]

Simple google search is gonna tell you more than enough