11/21/17

Wondering what is WSO's take on the coming automation of financial analysis. Will markets ever get so efficient from the algos that CFA's and research associates won't be needed?

Comments (10)

11/15/17

To be more specific, similar to how the old school stock brokers disappeared and were replaced by computers. Will CFA's and that sort be replaced by software?

Hedge Fund Interview Course
11/16/17

I do not think so. There's only so much that an algorithm can do. The parts of the industry that automation has replaced are not very complex. There is a lot of clamour about how Machine Learning is taking over and all.

But it's not that simple, as a computer at the end is an idiot box that needs to be told what needs to be done. So if you want to develop the rules that a machine will follow you need some one who is knowledgable enough about the rules.

Yes, may be there comes a time when CFA holders need to learn programming as well but I for one highly doubt it.

Also, Machines cannot reason or form their own opinion. If they could they would be sentient.

"The markets are always changing , and they are always the same."

Best Response
11/17/17

I think to play devil's advocate on this - the reason that machine learning space is growing in the analyst role is because many analysts themselves are "idiots." While I personally agree with you that there will always be a place for analysts - the sentiment is if I have to pay a team of analysts or I can buy an out of the box solution (as a quant) that can perform just as well as these analysts (and most do over longer periods of time)- I will always take the box because its cheaper, less emotional and consistent. The reason people like the quant products whether it be indexing, smart beta, stock pickers or whatever is because its time tested over a lengthy periods of time (think decades) vs what you read into a financial statement now and come up with a valuation. Machine Learning on top of that is supposed to be able to segment markets even further (ie what is important today and be more dynamic being able to identify what are the drivers in the market) which really very few people are able to do consistently.

edit: an example of a fund that does this today (stock picking - http://www.euclidean.com/)

11/20/17
Sasquatchosaros:

I think to play devil's advocate on this - the reason that machine learning space is growing in the analyst role is because many analysts themselves are "idiots." While I personally agree with you that there will always be a place for analysts - the sentiment is if I have to pay a team of analysts or I can buy an out of the box solution (as a quant) that can perform just as well as these analysts (and most do over longer periods of time)- I will always take the box because its cheaper, less emotional and consistent. The reason people like the quant products whether it be indexing, smart beta, stock pickers or whatever is because its time tested over a lengthy periods of time (think decades) vs what you read into a financial statement now and come up with a valuation. Machine Learning on top of that is supposed to be able to segment markets even further (ie what is important today and be more dynamic being able to identify what are the drivers in the market) which really very few people are able to do consistently.

edit: an example of a fund that does this today (stock picking - http://www.euclidean.com/)

This is correct imo

11/16/17

Thanks for the reply. Interesting take. As a follow up question: Wasn't the same thing being said about stock brokers and the exchanges? And then, one day, the floor traders and human brokers were gone. Can we possibly see that again with analysts due to AI tools that can provide more detailed data and can do better research than their human counterparts?

11/17/17

Yes, you are right. But they have not been completely replaced.

The floor traders who could not adapt fell by the way side and the ones who did moved upstairs and set up an office. The exchanges and brokerages have become electronic but they are still run by humans.

Like I said unless machines becomes senitient they will always follow rules. And they will go wrong every time they come across a situation that they have not faced before.

AI may be able to crunch numbers and give a detailed analysis but can they derive the inherent reason behind it? That is the question.

A very good example is the Flash Crash. One idiot box followed another idiot box and this continued to form a huge chain leading to the crash.

So you will require people. Another part of running a business is the "human touch" and that is something that I do not think can be codified in the next 100 years.

"The markets are always changing , and they are always the same."

11/18/17

No, because the valuation of stocks is ultimately based on human preferences (w/e investors collectively happen to prefer growth, dividend yield, etc at any given point in time). Its not some mathematical formula

11/18/17

2 things to add:

  1. Computers might be "idiot boxes" today that are governed by specific principles that allow you to extrapolate what is and isn't possible. But I don't think you can use these same rules to figure out what they will conquer next (or what they can't). The future will be vastly different than the present and it will not grow in a neat, linear fashion. I think the rules of the game could change entirely. What if computers can learn to reason like humans? There's plenty of philosophical ground to cover here, but I just think the future will be radically different than anyone here can guess and that we don't yet understand the full capability of artificial beings.
  2. When computers start replacing financial analysts, it's not going to be all of them (at first). They might replace 8 out of 10 and make the remaining 2 incredibly productive. It will be a symbiotic relationship between humans and machines.
11/19/17

Based on what I have seen when I see engineers work, it won't necessarily. You will create more issues by bringing in more "tools" so to speak, and at the end of the day, it is that simple. These are tools used for people to make business decisions. Imagine a security breach were to occur...

11/20/17

Unlock These Comments - Free

Join Us or Login

We are here to drink beer. We are here to kill war. We are here to laugh at the odds and live our lives so well that Death will tremble to take us. - Charles Bukowski

Add a Comment