[ikeboy]

>frankly the point of this is it could be socially efficient for two of your clients to bet against each other

So I did think of your scenario, and actually one worse (directly moving money from clients who don't value the next marginal dollar as much as clients who do). A couple of points I thought of in response:

1. Require clients funds to be transparent, so no moving money around after the allocation decisions have been made (this protects more against my problem than yours)

2. Assume the manager's utility is linear in money, in which case hedging their own fees doesn't make sense. If it's not linear in money, then this all becomes more complicated. I got to the point of saying something like "get a model for client utility and for manager utility of various amounts of money, then set the fee to align incentives". If the manager's utility is linear, then fees should be proportional to utility the client gained. If not, it becomes more complicated and perhaps gaming like you describe is inevitable, this is ultimately a math question for which I haven't modelled everything out and so don't know. You may be able to still define a fee structure that gives a proportional increase of utility to the manager for an increase of utility to the client, but I don't know for sure how that would work for multiple clients.

3. If fees represent utility, and utility is maximized, then wouldn't it be good anyway? (Kind of the point you made in the part I quoted.)

4. Another point to keep in mind is how various kinds of returns are expected to attract or repel customers, which factors into manager utility in a sort of "side-channel". If that's significant enough, it might make the whole incentives not work.

There are a lot of interesting math questions that go into this. I don't have too many answers, but I feel like I have a good grasp of the questions.

Some other points I see in my notes on this:

1. it could be that various kinds of existing funds already cover most of what investors really want, and this won't be enough of an improvement to justify the overhead. Investors who want a small chance of a large win go to hedge funds, investors who want the whole market buy index funds, etc, could be there's no room for innovation there

2. incentive structures need to be very specified, won't just be a fee for various amounts but need to take time dimension into account

3. For that matter, utility is time-sensitive, so we need a complex model just to determine utility, and maybe quizzing investors on utility won't actually get the "truth"

I like to think of this as a combination of the insights behind insurance and hedge funds, basically skewed utility curves.

I'd love to talk further if you're interested. I ended up starting a retail startup instead of working on this idea because it's easier to break into, but I'd be thrilled if someone turned it into a real company or incorporated it into an existing one, and you guys seem to be capable of doing something like that.

[kaistinchcombe]

Love to help you out on this if I can. The other thing to note is that because of the fear of double dealing there's a lot of stuff on the regulatory side about what kind of fees you can charge – my sense is, even saying, "any quarter in which you lose money, I'll just waive my fee" is illegal (perhaps because it would push the advisor to focus on preventing loss rather than tracking the market). But it's something that to a typical retail investor seems incredibly obvious – why do you get paid for losing me money?? – and I'm kind of sympathetic to, like I'm happy to share the pain in a down market (and to retain the customer). And saying "I actually have to charge a fee in a down market because it's in your best interest for me not to try too hard to avoid bad quarters…" well… there's only so much game theory your typical retail client wants to be doing sitting there in your office.

My gmail address is the same as my HN name. :)