A Blog by Jonathan Low

 

Feb 25, 2019

The Reason the Age Of the Financial Quant Is Coming To An End

Quants may have programmed themselves out of a career. Financial institutions have learned from quants' analyses that committing their own capital is less advantageous than deploying other people's money.

And the advent of machine learning and automation, whose use they pioneered, may have now rendered them obsolete. Which is ironic given how many other careers their insights have torpedoed over the past generation. " The captains and the kings depart..."


Joshua Gibbons reports in efinancial careers:

Banks' use of quants has changed. Ten years ago, quants needed knowledge of mathematical techniques. Today, quants have been dumbed down.Their role is to put a seal of approval on the numbers. Most quants are no more than a hygiene function. Nor will things improve with the spread of machine learning. People say that artificial intelligence will herald a new era of quant power. But machine learning is little more than linear algebra and you don't need a quant to do that. The real threat to the quant is the trader of tomorrow (who) will be more quantitative and will know how to code. There will be less math, more programming, and less pay.
I've worked as a quant in finance for over two decades, on both sides of the Atlantic. And I can smell the rot setting in.
In the past decade, banks' use of quants has changed. Ten years ago, when we were still building structured products, quants in banks needed deep knowledge of mathematical techniques to do the job. Today it's much less groundbreaking - quants have been dumbed down.
If you speak to the average quant in finance today, he or she will be working on model validation relating to the fundamental review of the trading book. Their role is to put a seal of approval on the numbers - no more, no less. Most quants are no more than a hygiene function.
Nor will things improve with the spread of machine learning. People say that artificial intelligence will herald a new era of quant power. But machine learning is little more than linear algebra and you don't need a quant to do that.
This comes as a shock to a lot of young people entering the industry, as does the persistently low pay of the more traditional quant's unfortunate sidekick - the quant developer (quant dev). Quant developers don't have structuring ideas and don't know finance like actual quants, but they know how to code the models. Quants and quant developers work together, but while quants can still get bonuses that are 100% of salaries, quant developers are on a fraction of their pay. Never go down the quant dev route.
I can't see things getting better. The trend is still for banks to move away from structured products and towards simpler products that require less capital. As this happens, the demand for really good quants will go down. Excellent people, will still be needed, but in smaller quantities.
The rest will be stuck doing low-end quant validation jobs. And for these, you won't really need a PhD. Who needs stochastic calculus now?
Nor are things likely to improve in the long term. The real threat to the quant is the trader of tomorrow. In future, traders will think differently. Traders themselves will be more quantitative and will know how to code in Python. However, traders won't want to spend their time coding in Python and so this will become the task of a new sort of quant-quant developer hybrid. There will be less maths, and more programming, and less pay.

1 comments:

harshak said...

apkeditor is a useful asset that can alter/hack apk documents to do bunches of things for the sake of entertainment.

Post a Comment