Discussion about this post

User's avatar
AqpW's avatar

An underrated value drain of HFT on the wider world is the black hole of incredibly talented people it sucks in. But there's no good answer to "How do you fairly compensate these people for their skills in other industry?", because the honest answer is "They (almost) always don't".

People who want to have an impact on the world and have the skillset to be a candidate in these types of firms "should" work for these firms, because the expected value of doing anything else is so paltry in comparison. Maybe some of these people go on to realize their original dreams; many do not. This is a well appreciated problem in tech these days, where Google and its peers were hiring so many of the best and brightest who would be a 10x at any small-mid size company and putting them to work at, like, Google Toolbar extension maintenance.

Of course at HFT firms these people are going to be more intellectually (and financially) stimulated than they would be elsewhere, but the inherently secretive nature of these firms mean that the techniques, knowledge, and skills developed at these firms almost never see the light of day, beyond a few select choices (Jane St's annual 'What the interns have wrought' gives a tiny slither of the type of things they develop). I have no doubts that there are secret forks of FOSS projects which have efficiency and performance savings at these firms which simply can never see the light of day because companies (correctly) acting in line with the incentives to not give up any edge they have, no matter how small.

You only need to look at the type of job listings posted for CitSec or whoever to realise that it's not just math Olympiads who these firms suck in: software engineers, highly specialized FPGA engineers, weather analysts, and only ever the best of the best.

What could these people achieve if they weren't just trying to save a few μs on trade execution?

Expand full comment
Bernke's avatar

What about rent-seeking:

Suppose there is news which means that Apple(~$3 trillion market cap) is under(or over) valued by 1%.

And let's say that one could capture just 0.1% of this price difference by being the first to get ahold of these news and trading on it.

That implies you could make a profit of $3 trillion*1%*0.1%=$30 million just by being something like 0.000001 seconds faster or whatever than the competition. So you'd be willing to spend up to $30 million to be first, the more competition there is the worse it gets.

Meanwhile the social benefit of updating the price to it's correct value 0.000001 seconds earlier is obviously(how could it not be?) ~$0.

Private benefit=$30 million

Private cost=social cost=up to $30 million

Social benefit: ~$0

So DWL is $0-30 million.

Now it's certainly true that there are positive externalities that one has to take into account but I see no theoretical reason to believe they happen to always exactly happen to balance out the external costs, nor does it seem likely that in this example it would amount to a huge number like $30 million.

See also: https://www.thebigquestions.com/2014/04/21/high-frequency-rentseeking/

Expand full comment
9 more comments...

No posts