Quant’s winter tail

[Maverick2608]

I'll repeat one more time, there are factor ETFs tracking all of these factors and their returns are pretty fucking transparent. All you need is a Yahoo Finance search.

Where do I ask for that? Read what I am asking for. Hint: Look for the question mark.

Yes, there are real world factor ETFs, but very few with long term track records and no one that I can find with long term outperformance. But again. That is not my question.

If you just like to insult people, then do it somewhere else. It is boring.

 

[Maverick2608]

I would add, however, and I cover quant ans factor investing professionally, a lot of the most popular factor funds are very water down versions of what someone like AQR would do.

For example, look at MTUM, that is by far the biggest momentum fund, but it only rebalances its portfolio twice per year. This goes against all of the academic research that suggests that the momentum factor in particular requires more recent rebalances to capture the premium.

Point is, Just be aware that these funds are more for the quant aware masses than real quant Alpha you would get from someone like AQR.

The more you rebalance, the higher the transaction costs.

Since you cover factor investing professionally, can you identify a public factor fund with a long term track record of outperformance?

 

[Slow Learning Elf]

If you just like to insult people, then do it somewhere else. It is boring.

Oh, insulting people on ET is buckets of fun. Arguably, that’s one of the very few valid reasons to be here for people like myself or @newwurldmn . It certainly not the intellectual thrill that brings us here

But seriously, if you want to see real life performance you have to pay for access. If you are simply looking for a proof that factor investing is “a thing”, there are several ways to do it.
first, you can either take a portfolio of these ETFs and track it since inception. IIRC all of these top factor ETFs have been around at least since 2013, giving you 10 years of historical returns. As @GlobalMacro90 mentioned above, the results are gonna be of questionable quality, but probably a good place to start.

A more grownup way would be to bite the bullet and backtest these yourself. I’ve only look at factor data in a very different context, but I can tell you it’s gonna be a fair amount of work. However, you can tweak which factors you like, play with weights to match your personal liquidity considerations and throttle the signal for best quality/TC trade off.

 

[vanzandt]

It certainly not the intellectual thrill that brings us here

What do you call it when a big group of crows schedule a meeting?
A premeditated murder.

 

[metalztrader]

Give me a break. You are obviously an insecure moron. I hate people like you. Always such a fucking pussy in person too. Thankfully people like you on here make yourself so easy to block and don't have to read

Oh, insulting people on ET is buckets of fun. Arguably, that’s one of the very few valid reasons to be here for people like myself or @newwurldmn . It certainly not the intellectual thrill that brings us here

 

[newwurldmn]

Give me a break. You are obviously an insecure moron. I hate people like you. Always such a fucking pussy in person too. Thankfully people like you on here make yourself so easy to block and don't have to read

slow learning elf is the second most respected member on this forum (as his prior nicknames). He left the site because he was tired of the trolls here.

He has a lot of experience as a buy and sellside trader in exotic and vanilla derivatives in both fixed income and equities. He is also one of the most giving.


The loser in this transaction, if you block him, is you, especially as you are running options strategies

 

[Slow Learning Elf]

He has a lot of experience as a buy and sellside trader in exotic and vanilla derivatives in both fixed income and equities. He is also one of the most giving.

Doesn't make me less of an insecure moron

BTW, for all the haters out there. A friend who is a fund allocator is gonna get me long-term historical performance for AQR’s Absolute Return fund. I remember seeing their tearsheet at some point so now I am curious if I remember their performance correctly.

 

[Kevin Schmit]

TLDR, just skimmed over: nothing of value in this extra long useless advertisement article.

Congratulations, you may be the single most stubbornly ignorant poster on this site! From your inception as botpro, through TheCoder and EarthEmperator, and now Quanto, you have shown yourself impervious to good advice and incapable of learning anything.

AQR has, over time, been extremely generous in its publications. Options, Factors, Trading... covering a broad range of market related subjects. Papers by Bryan Kelly (mentioned in the article) are particularly insightful. I've cited his work previously here on ET and here is another extremely useful quote from a recent working paper:

"Contrary to conventional wisdom, low in-sample variance principal
components (PCs) are key to out-of-sample model performance."

In fact, most of the non-predictive noise in your model design matrix reside in the larger (higher variance) components. And those noise factors are much more temporally stable than factors with high predictivity. Constant noise-beta models have smaller forecast errors than dynamic noise-beta models across nearly all model structures, which is in contrast to predictive-beta models, where dynamic beta models dominate.

By filtering out even some of the noise in the larger components, you can improve your forecast models significantly. And all from reading just a few AQR sponsored papers!

 

[Slow Learning Elf]

Congratulations, you may be the single most stubbornly ignorant poster on this site!

Now that I am back, he’s got competition!

As a side note (not for you, but for the public) even if factor investing is not your glass of hooch, it is so pervasive by now that you have to know and understand it. For example, I’d venture that a good vol trader (and I aspire to be one) has to pay attention to the recent performances in the factor universe

 

[Atikon]

Congratulations, you may be the single most stubbornly ignorant poster on this site! From your inception as botpro, through TheCoder and EarthEmperator, and now Quanto, you have shown yourself impervious to good advice and incapable of learning anything.

AQR has, over time, been extremely generous in its publications. Options, Factors, Trading... covering a broad range of market related subjects. Papers by Bryan Kelly (mentioned in the article) are particularly insightful. I've cited his work previously here on ET and here is another extremely useful quote from a recent working paper:

"Contrary to conventional wisdom, low in-sample variance principal
components (PCs) are key to out-of-sample model performance."

In fact, most of the non-predictive noise in your model design matrix reside in the larger (higher variance) components. And those noise factors are much more temporally stable than factors with high predictivity. Constant noise-beta models have smaller forecast errors than dynamic noise-beta models across nearly all model structures, which is in contrast to predictive-beta models, where dynamic beta models dominate.

By filtering out even some of the noise in the larger components, you can improve your forecast models significantly. And all from reading just a few AQR sponsored papers!

Interesting read.

I've been following the discussions on factor zoo papers, and the most recent one ('factor.zip') consolidated the over 500 'Factors' into approximately 14 clusters. This resonates with what other quantitative forums have been saying about the proliferation of Factors, which can essentially be distilled down to about 8-12 stable ones.

I've always been somewhat puzzled by sell-side research creating a plethora of baskets that seem like derivatives of the big five factors.

What are your thoughts on a model architecture that merges stable, constant beta factors with a narrative Factor (for situations like KRE/regional Banking this year), incorporating shorter term /regularization for the latter?