From the January 2008 issue of Wealth Manager Web • Subscribe!

A Question of Supply

Chess is one example, poker is another. But is investment alpha a zero sum game, too? Do the winners in money management always triumph at the expense of the losers?

Many if not most investors accept the equilibrium of alpha as a zero sum game. And so it is for investment returns generated from a specific opportunity set, such as the stocks in the S&P 500. But what happens if we add risk into the equation? Does alpha still sum to zero in risk-adjusted terms? Not necessarily, at least not all the time.

That's the message from a small but assertive group of investment strategists including Max Darnell, a partner and chief investment officer at First Quadrant, an institutional money manager in Pasadena, Calif. that specializes in quantitative strategies including global tactical asset allocation. A sampling of Darnell's reasoning appears elsewhere in this month's issue (see Portfolio View on page64) in a story that explores the idea that the supply of alpha may not be as limited as convention suggests. After interviewing Darnell, we decided that his extended comments should receive a fuller airing on what is for some a controversial subject.

Readers may or may not agree with Darnell, but all strategic-minded investors are likely to benefit from listening in on the discussion. Alpha, after all, is at the center of all investing. Some seek it, others avoid it by design through indexing. But one way or another, alpha has everyone's attention. Investigating the nature of alpha's existence in the investment universe is likely to be a productive debate, if only to test and perhaps strengthen one's assumptions about how the capital markets operate.

What follows is a conversation on just that, courtesy of Darnell, who joined First Quadrant in 1991 as manager of derivatives research, a post he held until 2000, when he was named director of research and, two years later, CIO

You maintain that alpha's not a zero sum game. Is that an academic view, or does it also have relevance in the real world?

It's more than an academic exercise. Depending on how you answer the question, it should make a lot of difference in terms of how an investor sets his objective.

To make sure I don't run into any terminological issues, I'm using the term "alpha" loosely. What I really mean is "value added." When people talk about alpha as a zero sum game, they're trying to figure out if they can do something to add value. Technically speaking, alpha should be referred to as idiosyncratic risk.

Meaning something other than beta?

Yes, something other than beta. There are really two basic ways to add value. You can play the game of trying to improve your returns through idiosyncratic risk. Or, you can do it through the management of beta across time. That is, you can vary your exposure to beta over time, or you can arrange the betas in a way that you expect will lead to a superior outcome.

Real world examples include the endowment funds of the Harvards and the Yales of the world. They're choosing a configuration of betas and alphas that work to their advantage vis ? vis other market participants. Foundations and endowments can live with a longer time horizon. They also have a smaller asset size to manage than the biggest institutional pension funds. So endowments and foundations have flexibility and different objectives that distinguish them and allow them to take on risks that others might not.

The sell-off in August is a recent example. Consider two types of investors. One has a relatively short-term horizon and is unwilling to bear short-term liquidity issues. Examples include some hedge funds and individual investors who may react with shorter-term horizons even though they shouldn't.

On the other side are the foundations and endowments, which may see others selling because of liquidity issues and decide that it's just a blip, a momentary event. These buyers may have very different objectives, and so they can wait for a return of liquidity. Is a foundation or endowment generating alpha or adding value by buying assets that looked risky to one person but didn't look risky to another? Absolutely.

Because...

Because they recognize that they have a set of differences that vary with other investors and decide that they're going to take advantage of those differences. And so they recognize that they ought to be able to do better.

That view tends to favor investors with a longer-term view on risk. Examples include some individual investors, particularly those with a lot of wealth that they may be expecting to pass on to subsequent generations. They may be in a position to behave like longer-term investors. Very often that means doing the most uncomfortable things, like buying assets that have been beaten up.Meanwhile, there's a whole set of investors who don't want the ugly looking assets in their holdings list--mutual funds, for instance. Or, maybe it's a hedge fund with a very short-horizon objective; or an absolute return fund that's trying to hedge its downside risks and so it can't bear further declines.

The bottom line: There's the potential for trading between these two kinds of investors that's good for both.

Is that because one is laying off risk while the other is embracing it, and so that ends up being a net gain for both?From the perspective of their objectives, the answer is yes. It's not necessarily a gain in the return space for both, but in risk-adjusted return terms it very well may be [a gain for both].

Does thinking in risk-adjusted terms go to the heart of why alpha isn't necessarily a zero sum game?

Yes, and that's the way people ultimately have to think about their objectives. One simple example: If you're an individual without a lot of wealth, and dependent on that wealth for surviving one or two years out, you need to think about the risk side of the equation because if you hold more risk than you can financially bear, your life could change dramatically. You may not be able to pay the mortgage, etc. So you may be able to improve the risk-adjusted quality of your returns by selling assets that have recently become more risky.

So alpha sums to zero for straight returns relative to a benchmark, but the principle doesn't necessarily hold if you add risk to the equation.

That's right. It's on that risk side that we all distinguish ourselves. That's why the foundations and endowments are so different from individuals of modest means. What really distinguishes the institutions is the ability to bear investment risk over different types of horizons. And that's critical in the investment puzzle.

Here's one way to think about it. You always have a set of investment choices. Let's take two assets that you think have similar return opportunities and similar risks. You're invested in one, which starts going up. What should do you do? You can switch over to the one that has the risk/return profile that you can bear. It doesn't necessarily mean degrading your return, but it does mean adjusting the risk/return profile.

If you ignore the risk side of the equation, this whole argument goes away. If you're living in a return-only world and ignore the differences between investors, which are risk differences, then return is a zero sum game, if you will.

Assuming that alpha sums to zero seems to be the conventional wisdom.

From an academic perspective, a heroic assumption is made that risk is the same for every investor. We all value return in exactly the same way and so the return side of the issue is uninteresting when you say there's heterogeneity among investors. But when you move to differences in preferences, it's all about those differences in risk and objectives and constraints. And so the academic literature--and I think this is the big flaw in the zero sum game argument--generally starts with an explicit or implicit assumption that all investors look at risk in the same way. But they don't. Your readers are remarkably different from institutional investors, for example.

If someone accepts the idea that alpha's not a zero sum game, how does that inform investment strategy?

One example is thinking about how your risk profile and objectives differ from other investors. In the past, pension funds often had similar asset allocations. But that was a mistake. How could they make that mistake? They assumed that there were no differences in their risk profiles and so they assumed no advantage in exploiting the differences.

They assumed alpha was a zero sum game?

That's right, and so they behaved similarly to the funds considered to be their cohorts. A classic example was in the U.K., where there were dramatic differences in the duration of the liabilities of the pension funds. Yet they were all holding 70/30 mixes of stocks and bonds--or roughly so. That shouldn't have been the case. But they assumed their differences away and decided that there was no value in distinguishing their objectives, which allowed them to hold conventional allocations.

If alpha doesn't sum to zero, how does that view square with the principle that all investors can't be above average?

If we limit that comment to idiosyncratic risk, it's absolutely true. But idiosyncratic risk is so boxed in, if you will, because we accept a benchmark and it's valued in the same way by everyone. In that sense, generating return in excess of the benchmark means taking that return from someone else.

Meanwhile, compare that to a real- world case that's topical. Fundamental indices, for example. [Benchmarks that weight securities by "fundamental" factors, such as dividends and earnings, and are considered alternatives to traditional cap-weighted indices.] You can say that investors with formalized liabilities may place a higher value on more fundamentally oriented-type stocks in their portfolio. And while they may or may not get a return advantage from holding those stocks, they might get a greater utility advantage because they're holding something that is a match for their liabilities.

So investor utility helps explain why alpha isn't a zero sum game.

You just can't get away from it [investor utility]. If you say that not all managers can beat the benchmark, that's only true in a return-only space. I accept that argument all day long if we're talking about return only. In that case, alpha is a zero sum game.

The question is whether that's how to look at investing in the real world?No, not even close.

JAMES PICERNO (jpicerno@highlinemedia.com) is senior writer at Wealth Manager.

Reprints Discuss this story
This is where the comments go.