Sunday, February 12, 2012

Don't expect "expected" returns

Reading this article on PIMCO, I was struck by the following line:
Industry analysts also wonder whether PIMCO's $250 billion Total Return Fund, the world's largest bond fund, is such a behemoth that Gross sometimes has to swing for the fences to generate the kind of returns investors have come to expect.
Think about this for a minute. It may seem very natural to you that people who pay large fees to a company to manage their wealth should expect that company to earn them a higher return than they would earn if they just stuck their money in a low-fee index fund.

But does this expectation make sense in the long run? Suppose you gave $1M of your money to a hedge fund with an awesome brilliant manager who was incredibly talented at picking bonds that were going to go up in price. Soon your money doubles to $2M, then to $4M, even as the average bond price goes up only a little bit. You keep your money invested with the same manager (paying the same high fee), assuming that he will continue to double your money in the same amount of time, again and again. But as the manager accumulates a bigger and bigger share of the total bond market, it becomes harder for him to beat the market average.

To see this, just imagine if your hedge fund manager did so well that pretty soon he was managing the wealth of every bond investor. In that case, it will be impossible for him to beat the market, because he is the market, literally. So it makes no sense to expect the same excess return (i.e. the same percentage points of market-beating performance) year after year.

In your intro finance textbook, you will learn that various investments have various different expected rates of return. These expected rates of return may vary randomly, but they are not believed to decrease over time. Does that contradict what I just told you?

No, it doesn't. It's theoretically possible for stocks to go up by, say, 3% a year more than bonds forever and ever (or at least for a very very long time). Comparing one asset class to another is different than comparing one asset to the average performance of its asset class. To see this, note that it's possible for Apple stock to earn a 5% higher return than RIM stock forever, but it's not possible for Apple stock to earn a 5% higher return than the average of Apple and RIM stock forever, since eventually Apple will dominate the average.

Note that this is not Efficient Markets theory. This is just arithmetic. But the more "efficient" markets are, the quicker it will become impossible to beat the average as you grow in size, because the available mispricings that you can exploit to get excess returns will be more limited.

So here's the lesson: if investors "expect" an asset management company to beat a comparable index fund by the same number of percentage points year in and year out, they will be disappointed, because the company will soon get so big that it is almost indistinguishable from an index fund (except that it will have higher fees). Such fixed expectations of excess returns are not rational. But if investors do have such expectations, asset managers will have an incentive to accept greater and greater risk in order to have a chance of holding onto their clients. This is exactly what that article says is happening with PIMCO.


  1. It's a big problem for the fund business when the fund is the index, the index yield is minimal (< 1% on five-year Treasuries), and you charge more than half that in expenses.

    It also seems worth noting that at the current price, bonds mathematically cannot achieve the risk/return that would be accorded to them by historical estimates in a traditional mean-variance efficient frontier framework. (even disregarding inflation, tax disadvantage vs. equities)

    One shouldn't expect 'expected' returns in that sense either.

  2. PIMCO,which has many brilliant young minds working for them, also benefits from some survivor bias. They were started in the 70s. Since the early 80s though, treasuries have been on a run Did they really squeeze out a lot more than the expected returns during that time?

    It isn't that there aren't winners in the market game - it is a distribution after all. Most people being average looking isn't mutually exclusive to the existence of Tom Brady and Gisele Bundchen. (Can they please, please, breed like Duggars?!) Winning tends to be a temporary state in markets though.

    Take Paulson. Here was a "hedge" (what was he hedging exactly?) fund manager who came down on the winning side and had one of the biggest windfalls in history. Media, congress, investors declared him brilliant at picking the ponies. Now his fund, after increasing AUM to become one of the biggest in the world, lost 50% of its value. Those who thought they were getting a winner have to wait years in order to break even (graph is after inflation and management fees)

  3. @Chris:

    Want to do a guest post on failed hedge funds?

  4. Hey Noah, this topic reminded me a lot about Long-Term Capital Management, which suffered from the pressure to maintain its absurd returns even as it got bigger and the arbitrage opportunities it relied upon shrunk. I wrote a post about this if you're interested: Hope you enjoy it.

  5. Chis, yes they did. It's fair to say PIMCO revolutioized fixed income money management back when bond managers clipped coupons for a living. Of course they rode a tidal wave of credit, but that goes for everyone else too. And no one beat on the index like PIMCO did.

    Not following your arithmetic here Noah. There is no conceptual difference between an asset class and an asset, at least in this context, and there is no reason why an asset that is 99.9% of a portfolio couldn't outperform the .01%er by 5%. And on to infinitum.

    U r stuck with market efficiency to make this argument work, even st the limit, though an exceedingly week variety thereof which wouldn't even bother me- something like, 'you can't fool all the people all the time'.

    The problem PIMCO faces with it's size is a variant of what all company's face as they grow, as much as it has to do with the particulars of investments- namely, scaling the people and the resources without losing the product in the process. Money management in particular struggles with this, and not just because it typically overestimates strategy depth and its people ability to generate ideas, but also because it pretty much requires small, cohesive groups to work well, like software development.

    On top of all that, they have simply been wrong about a great many things in the last decade, and secular things where great sums can be made or lost and implementation is not an issue, even at their size. They are tremendously insightful across a broad range of capabilities, but they have not been able to put together the big picture very effectively at all. One wonders where this Reuters piece comes from in that context.

  6. "So it makes no sense to expect the same excess return (i.e. the same percentage points of market-beating performance) year after year."

    It may be able to happen for many years before he starts to get 10%, 20%+ of the market. People only live so long (at least currently). If this guy can give you a 30 year great run, you're very happy, and very rich, especially if we're talking about stocks. A lot of economists make the mistake of "eventually". Like eventually people will learn ridiculous amounts. Well, unfortunately, they'll die long before that happens, and new people will be born who know nothing. Nonetheless, what you say about getting too big does seem to hold to a large extent sometimes. Buffet seems to be an example.

    Expanding on Buffet, he has (or had) for decades generated very excessive risk-adjusted returns. But he was, to a large extent, buying companies, taking control, and making those companies a lot more productive, in real terms. You can keep earning, say, a 12% real return indefinitely, if you're actually increasing the global pie that much. Of course, if that keeps up, it then becomes the new normal, and you aren't above normal anymore.

    I actually have a hypothesis, that I think there's strong evidence for, that a big part of the explanation for the equity premium puzzle is just that money invested in stocks can generate a lot more real wealth than money invested in bonds. The reason is that stocks give a lot more flexibility and ability to think long-term, to invest in high expected return projects that are not going to pay off for a long time, and are highly uncertain, but are nonetheless very high expected return, even for their risk. I actually saw this a lot as a Ph.D. student. I wanted to invest my time in very long run intuition, hard things that weren't very narrow and cut-and-dried in a clear area. There was no pub on the near horizon, just a lot of learning and thinking in many areas, both broad and deep and technical (like advanced Bayesian econometrics and continuous time finance, very time consuming to learn). Clearly, they weren't pleased. They were like bankers, not stock holders. They wanted tangible returns in the near future, and clear metrics to show how hard you're working. Fine, but this excludes many projects that can be far higher return over the very long run, even adjusting for risk.

    For more on this supply side explanation of the equity premium puzzle, see:

  7. By the way, just to be clear, a big reason why debt financing discourages investing in long term and/or risky projects (even if those projects are nonetheless very positive NPV, especially when in a well diversified portfolio of thousands of stocks) is that interest payments won't wait long. With stock, the project can be taken and if returns don't happen for many years, or ever, the firm won't go bankrupt. There's just a lot more flexibility to take such projects the more the firm is stock financed.

    Other reasons are monitoring and restrictions from bankers and bondholders.

  8. So it makes no sense to expect the same excess return (i.e. the same percentage points of market-beating performance) year after year.

    Yet people continued to invest and expect these returns year after year.

    Blows yet another hole in Rational Expectations, no?