Jbt7493

Levine today:

Arguably the way to do that is to build a general-purpose factory that uses people (portfolio managers) as inputs and turns them into outputs (portfolios). The inputs, in this model, get used up to make the outputs. The inputs are people. “Citadel consumes new traders as voraciously as Starbucks consumes beans,” writes Sernovitz, and:A former Citadel portfolio manager said of such overhauls, “The next incarnation of what gets built is better than what was there before, and that’s always the case.” The downside: “a highway wreck of human bodies.” Almost every former employee I spoke to had stories about turnover. One had five bosses in five years; another kept a “Book of Souls” listing the fifty colleagues who had left his small trading unit in six years. Yet another moved to Chicago only to see his boss’s boss be immediately fired, and then his boss fired. ... (A Citadel spokesperson said of its firings, “We have deliberately built a high-performance culture.”) “It was a gift to be there and a gift to leave,” one former employee told Sernovitz, which is perhaps how Citadel feels too.

So this is much less harsh than it sounds, and that itself shows the benefits of arriving at a different equilibrium. It's true Citadel goes through workers like Skittles. Some vice-president might have a random idea, like a new method for combining HFT and NLP. They're hire one senior expert, 5-10 mid-level finance people with experience in one of those fields, and tell them to work on it for six or seven figures. Then a year and a half later, the direction of the company might change, or they might just see the experiment is bringing in mediocre yields and isn't about to be a new "Indian Options Market." They immediately fire all of those people.

That sounds unfair, but the unfairness is the point.

Because they are, well, nice about it. There's the famous finance garden leave - a no compete contract for 1 year that pays you your regular salary. And they are very generous with healthcare over that time. There's no hard feelings from the company, they're not treating you like a potential disgruntled employee who might sabotage the company so we have to block your access ASAP. *It's not your fault and it's not about you being bad and everyone knows that.* More like the last day of a freelancer's contract - the mission is simply over, but you might get picked up again.

And outside, the register holds. No one sees "1 year at Citadel" on your resume and calls to ask what the problem was. It's the mark of a veteran, you went through a high tier qualification sieve, got experience at the very top, and through no fault of your own are on the market again. Your resume might as well say "took a break to climb Mt. Denali."

This pays dividends for the company. When a project looks like it has plateaued - either failed, or crested with no unicorn returns - they can cut it loose. You don't have to find new spots in the org for those employees. You don't have an entrenched director fighting to keep his department from being cut. It's much less friction. AND because you know you can end trial projects painlessly, you can also START a lot more projects on a whim. The company catch more of-the-moment opportunities, and is not carrying the luggage of a thousand past projects with salaries dependent on pretending this project works.

萩原 卓哉／Hagihara Takuya

a girlfriend is an emergent property of complex systems