First, I'd spend a sizable amount attracting and testing nurses to find one with sufficient mathematical aptitude and motivation. We're not really talking about any random person from a population but whether a particular type of person exists in that group.
If I could catch a person who would become a nurse earlier (assuming some aptitude and motivation in the desired direction) and train them to be a trader or a developer it could be cheaper -- not so much wasted time learning skills that would not be applicable.
You seem to assume that people end up fulfilling their full earning aptitude and, if they are at the 'level' of a nurse, they are obviously not cut out for something more rarefied.
You might be right, but I don't think it is wise to assume the end result, simply because it is the result that occurred, is proof that is the optimal result.
If we assume that $20M/year traders are the ones skewing the data on the right side of the graph, the question is what is keeping others from entering that field and competing?
I really don't think 'raw aptitude' is the ultimate answer. I don't have an answer, and that is my point -- we need to explore why such a deep moat exists and what we can do to drain it. If there are implicit costs that reduce the economic profit so that it is so divorced from accounting profit, what are they and can they be reduced? I don't think we can assume that all is as it should be and move on.
I'm not assuming this at all. I'm just trying to illustrate that implicit costs are real, you just can't necessarily see what they are without delving into very concrete programs. E.g., running your gigantic screening program, the training program for folks with high aptitude who pass the screening, the ones who flunk out of the training program, etc.
If you recall, tptacek and patio11 recently started a startup (starfighter) trying to do something along those lines. They've now moved on to something else.
Now I don't think every single person with aptitude is already there - Zoho University demonstrates this pretty well. But I think exploiting the high hanging fruit is a lot harder than you think.
If we assume that $20M/year traders are the ones skewing the data on the right side of the graph, the question is what is keeping others from entering that field and competing?
Nothing. Trading is one of the most open, meritocratic and low barrier to entry fields out there. Just google how easy it is to get a Robin Hood or Interactive Brokers account [1]. In terms of barrier to entry, it's far lower than something like building a SAAS app.
The only thing that prevents more people from doing it is the fact that predicting the future is hard.
[1] I have an IB account, it took a couple of weeks to set everything up. I'm good at predicting the future, e.g. here's me publicly talking up my book (currently up 15%): https://twitter.com/stucchio/status/747493530277294082 But I'm not $20M/year good.
That gave me a chuckle (though I suspect you're trying to bait me -- you seem to have a background in finance and would know this all very well). The one obvious barrier to entry is access to capital. It is a lot easier (or at least possible) to play with different strategies with more than less. Even a brilliant trader would have a hard time playing different strategies with, say, $10K. Or even $100K. They'd quickly hit a bad play or a down trade and lose beyond the capacity to absorb with a chance of bouncing back. Game over.
But the capital is out there so that's not really the issue. And that capital is paying traders $20/mil per year. That is what I mean by competing...not competing on trading talent but competing for the jobs. Its not like hedge funds are scanning Robin Hood for diamond in the rough trading talent...if they did, it would be a mistake since they would be nearly guaranteed to find people who do very well up until they point they hire them then they crash back to earth through regression to the mean.
>The only thing that prevents more people from doing it is the fact that predicting the future is hard.
I doubt that is the reason. Predicting the future is hard for everyone, including top-flight traders, as has been shown year after year, study after study. What is easier, and a skill (though, granted, a challenging skill to master and one that is still, even after some level of mastery, subject to more luck than most people are comfortable with), is using all of the tools at one's disposal to win more than you lose. Or rely on luck. That works sometimes, too -- for a while.
I'm just saying -- if there's a great amount of money being reaped year after year, decade after decade, the system surrounding that activity is probably worth a careful look.
This is completely backwards. Every strategy has a capacity and most of them have a very low capacity.
I'm currently just a hobbyist trader. I've got a strategy or two, but I can't even put all of my (software engineer level) savings into it - additional marginal dollars will lose money. My current strategy never does more than pick off the top level of the order book - it can't, and if I try my additional dollars will lose money according to my simulations.
A bigger strategy might need to not just buy from deeper in the order book, but actually make repeated purchases and face market impact. That's way harder. The book "Flash Boys" is all about how market makers try to give you and me a better price than big players trying to do this.
The only strategy I run which has enough capacity for me to really put money in it is my "reason from first principles to figure out when everyone has gone insane" strategy. E.g., that's what made me a bunch of money from Brexit. But that's a pretty risky strategy involving only a few big bets/year. (Next one will be about $50-100k on "stop overreacting, Trump won't be that bad", conditional on him winning the election.)
Trading a big book is vastly harder, not easier, than trading a small one.
Its not like hedge funds are scanning Robin Hood for diamond in the rough trading talent...if they did, it would be a mistake since they would be nearly guaranteed to find people who do very well up until they point they hire them then they crash back to earth through regression to the mean.
I'm confused. I'm arguing that top talent is uncommon and hard to find. Your tone suggests you are disagreeing with me, yet you just agreed. Can you clarify?
That is not talent. That is skill. It is a skill that can be learned. A large part of the skill of poker is calculation. There are people who back into that skill by sheer volume of playing -- they develop an intuition based on facts repeated over and over more or less haphazardly. This important probabilistic intuition can be greatly expedited with some mathematical training. It is possible to learn this skill. It is also possible to learn some rudimentary information and game theory to help with those aspects of poker. All teachable skills.
There are soft skills that are also teachable like controlling one's emotions when dealing with large amounts of money.
If I wanted to go out and find a team of poker players to bankroll, I might go and find people who tend to win (longer time frame observation under similar high stakes conditions would be preferred) but I would add the additional filter of testing their skill in those areas.
This is opposed to observing a handful of low stakes games and picking those that do best and running with that group.
That would be a mistake. I might get some of the same people. But I would also sweep up people who merely got lucky. I'd also miss some people that got unlucky.
Picking stock traders from their performance on their Robin Hood accounts would be a similar mistake.
It is subtle, I know, but critical.
There is no economic law that the price of gold should fall to the same price as other metals. Rather, economics says that if gold mining is a competitive industry, gold production will increase until the economic profit on producing more drops to the same level as economic profits in all other productive directions.
Concretely, to understand why top earners outstrip others, it's important to figure out who they are. A lot of them are traders. A trader earning $20M/year makes that much because his trading desk produces $200M/year for the hedge fund he works for, and because no one else can.
Now, suppose you want to take a nurse making $40k/year and bring her up to the trader's level. Suppose I give you $500k. Do you think you can do it? Concretely speaking, and without handwaving, what would you spend the money on? A harvard education (using up $240k already) is far from sufficient.
Forget even bringing her into the 0.1%. How much money would you need to spend, and what would you spend it on, for her to reliably become a mid-level developer earning $200k/year?
The implicit costs are hard to pin down. But when you ask how to move a specific person into a specific 1% or even 10% job, it suddenly becomes a lot clearer why this is hard.