Hacker News Reader

Comment by p0ckets | Hacker News Reader

From a link in the article: "

By Matt Levine

Here is a simple model for hedge fund fees:

1. There are some people who can reliably generate alpha -- returns in excess of the market return -- but those people are rare.

2. It is somewhat difficult to tell who those people are; in particular, at any given time, there are more people who look like they can generate alpha than who actually can.

3. If you are one of the people who can generate alpha, you should charge a fee for your services that is equal to the alpha that you generate.

4. If you are not one of those people, you should charge a fee equal to the alpha that those people generate, because then investors might think that you're one of them. "

View on HN

Ok, so I'll be the guy on the other side of this hypothetical deal. In the scenario he's described, either I get market returns by paying the elite guy to keep his alpha, or I get sub-market returns by paying too much to a guy who isn't actually elite.

So, best case scenario I get market returns? Then why not just invest in an index fund?

A)Because with 160bln you're the market;

B) Change the point of view from the retail investor to the pension fund, funds have a very narrow and constrained mandate, they need to hedge risks and deal with the increasing negative cash-flows. They can't merely park their money somewhere and hope for the best.

> So, best case scenario I get market returns?

Yes. Note: This is a surprisingly accurate description of reality. Aggregate hedge fund returns are goddamn terrible.

> Then why not just invest in an index fund?

Well, you should, if you're trying to maximise your expected outcome. Of course, not everyone is trying to do that.

In particular, what if you run a pension fund which is currently underfunded, but for political reasons is claiming to be adequately funded through the expedient of assuming that future returns will exceed any reasonable expectation of market returns?

If you invest in an index fund you'll get market returns; since that's not enough this means you will miss your targets, and be unable to pay promised pensions, at which point you'll be fired. But if you give all the funds cash to a hedge fund, or engage in some crazy snowball derivative[1], then you'll probably do even worse than an index fund, be able to pay an even lower percentage of the promised pensions, and you'll be fired. Which is actually no worse for you. But you MIGHT do really well, and actually be able to pay the promised pensions. Not likely, but if it works you avoid you getting fired, and if it doesn't you were going to be fired anyhow, so why not give it a shot?

All of why is completely hypothetical. The fact that many US pensions funds are horribly underfunded using any plausible actuarial projects and are simultaneously investing heavily in exotic asset classes is just a funny coincidence.

[1]: http://www.bloombergview.com/articles/2014-05-02/portuguese-...

If the guy generating alpha keeps most/all of his above-market alpha, then that means there's none for me to get or keep, right? I can't see how that's attractive at all.

That's the point. An ability to reliably beat the market is like having a goose that lays golden eggs.

If the ability exists at all, it's certainly very rare. And if you are one of the people who has it, why would you hire that ability out to other people? The guy who owns the golden goose is going to sell the eggs, not rent the goose out. Or if he did rent the goose, it would be for an amount no lower than the expected value of the eggs because otherwise he's losing money.

But given these dynamics, that means that there's no reason to think that if you see some guy offering to rent a golden goose it would be a great deal for you. I mean, he's actively trying to price it so it's not, and it's his goose, so he's probably right.

The other thing hedge funds offer is alternative markets to track. So while some may track the U.S. stock market (which you can get by just buying SPY), some funds track other markets like commodities, foreign equities, housing, bonds, etc. Almost anything that you can put a price on you can find a hedge fund willing to trade. For these "nontraditional" investments it can be rational to pay someone even if they cannot beat the benchmark, simply because you couldn't do it yourself.

>Then why not just invest in an index fund?

This is what most level headed people who study the stock market suggest you do!

I'm not sure I understand - what's the benefit to the customer? If you gain alpha from hiring (3), you gain proportion of alpha + alpha. If you hire (4), you're paying a fee of alpha for a gain < alpha. The customer assumes a lot of risk, especially the more common (4) gets. It seems the customer is most likely to lose in this scenario.

The model is from the perspective of fund managers. If you believe his model, then from the other side of the board, as an investor, your conclusion has to be that managed funds are a bad deal, or a break even at best if you find a manager with alpha.

And the simple model is wrong. Here is why.

If you are a customer for hedge funds, you should be willing to pay anything up to the expected future alpha based on past correlations between previous and future alpha.

That correlation is always affected by regression to the mean. Therefore someone with a good track record will not be able to charge on the expectation of future performance matching past performance. But only on the expectation of future performance being somewhat better than the market mean.

The ideal if everyone does everything perfectly is that, on average, fund managers that can generate excess alpha will, again on average, capture their entire outperformance as fees.

Reply on news.ycombinator.com