By: Lee Thomas, Allianz Dresdner Asset Management
Published: Winter 2005, Journal of Investing

It seems the money management industry is coming around to understanding something venture capitalists figured out many years ago: that it’s better to make a large number of bets on separate businesses than it is to put your eggs in one basket – even if you believe your ability to identify winners is without compare.

Lee Thomas proposes what amounts to a technology incubator as the ideal model for the asset management business. Remember the incubators? Part VC, part landlord, part mini-conglomerate. Thomas calls his investment management incarnation an alpha engine.

Firstly, he explains why alpha matters again:

During the boom we could all but ignore alpha, those extra few percentage points above the market return that active managers claimed they could deliver. When market indexes were producing double-digit returns, alpha seemed irrelevant. Now, with market returns to passive stock and bond investments reverting to their norms, alpha matters again.

The start-ups in Thomas’ incubator are actually small groups of alpha-producing managers. While many of these managers may not be technically market-neutral, their alpha can be delineated from the active bets they make for and against positions in an index.

“If we list the securities in the benchmark portfolio we are trying to beat, and compare them with the securities in our actively managed portfolio, each difference represents an alpha bet. I will call the collection of all these differences, or alpha bets, the investor’s ‘alpha portfolio.’”

To govern the relationship between the alpha and beta portfolios, Thomas proposes the “Fundamental Law of Alpha Portability”:

“The Fundamental Law of Alpha Portability: The composition of the optimal alpha portfolio is independent of the benchmark.

“In other words, when you design a portfolio of alpha bets to beat a benchmark, you can ignore the contents of the benchmark. When you have finished constructing the alpha portfolio, you simply attach it to the benchmark portfolio. Another way to think about it is this: when we build our portfolio, we can use modular assembly techniques. We can construct a benchmark-replicating portfolio as one module. And we can independently construct an alpha-seeking portfolio as a second module. To assemble the portfolio, just combine the two modules; the aggressiveness of the portfolio, its tracking error, will be determined by how much of the alpha-seeking portfolio you add to the benchmark-replicating portfolio.”

So an alpha portfolio can be constructed by engineering-out the alpha from within any active (long-only) portfolio. This is a process described in several postings on this blog back in the summer. Those of you who were on the beach, at the cottage or touring wine country around then, may be interested in this post on the topic of stripping out the embedded market neutral hedge fund from within what you might call the host mutual fund.

Adding several of these non-correlated alpha portfolios together produces a fund whose expected return might be the same as the average of each portfolio, but whose volatility is dramatically lower – especially if the alpha portfolios have low cross-correlations.

Thomas’ diversification 101 discussion gives way to a number of practical conclusions about the ideal model for a money manager:

Suppose you manage funds for 100 different clients, and each client has a different benchmark. What an expensive proposition funds management would be if a manager had to construct 100 different alpha portfolios for 100 benchmarks, requiring costly handcrafting of every portfolio for every client. But because the composition of the optimal alpha portfolio is independent of the composition of the benchmark, you need only find a single, optimal alpha portfolio and then attach it to each benchmark. The result is handcrafted quality at mass-production prices.

Now we’re talking!  Money managers will almost certainly read the last line here as handcrafted, (and high-priced), quality at mass-production costs (leading to huge margins).

Thomas suggests a meta-boutique approach to organizing the money management firm:

What I propose is…a firm that resembles a collection of boutiques—teams employing different styles, each of which contributes a handful of alpha bets to the overall portfolio. These different team bets are produced independently, based on different approaches to identifying potential alpha, thus generating diversification.

This article may have been a little ahead of its time back when beta produced double digits.  But its relevance is greater than ever now that asset managers are searching for ways to respond to falling returns and an onslaught from the burgeoning hedge fund sector.

Read Full Paper (subscription required)
Read Earlier Version of Paper (under the PIMCO name)