We Had The Same Crazy Idea

“Not everyone can become a great artist, but a great artist can come from anywhere.”
- Anton Ego, from Ratatouille (2007)

“One look and I knew we had the same crazy idea.”
- Remy, from Ratatouille (2007)

No narrative talk today. I want to talk about something else.

There is a moment – a very special moment – in the life of every asset allocator or financial advisor. Maybe it happens at a dinner with a fund manager, and that pre-dinner cocktail and the second glass of wine are starting to make some magic happen. Maybe it happens in a planning session for the next year, or in a marathon strategic asset allocation roundtable, looking for ideas to fill the gap on expected returns. Maybe it’s after hearing a brilliant peer speak about it at a conference at the Breakers. But sooner or later, if you stick around this industry long enough, it will happen to you, too.

You’ll convince yourself that it’s time to start a best ideas portfolio.

During my time at Texas Teachers, one of the most common topics of conversation was this holy grail: how do we get more out of our big third-party relationships? Sure, like you’d probably guess, ‘more’ often meant better aligned fee structures. We were pretty good at getting those. Sometimes it meant access. That was easy, too. But the real holy grail, that elusive thing you never quite achieve that’s really about the friends you made along the way? Information.

The standard thinking goes like this: we are paying millions in fees to the smartest, most successful investment firms in the world. Why can’t we leverage the information we get from them and how they manage our money to improve our internal strategic and tactical asset allocation, and maybe even some of our internally managed security selection portfolios? There are parts of this thesis that are, in fact, feasible - and a pretty effective use of resources. Most of those good parts are on the asset allocation side. But the really sexy, interesting notion to most investors isn’t about improving asset allocation methodologies. It’s the idea that we as asset owners – or financial advisers - can comb through the positioning and trades of our external managers, decompose and map out what some of them are good and bad at, and build directly implemented portfolios of best ideas.

It almost never works.

I’m sorry. Leaving aside even the most basic criticisms of the difficulty of identifying sources of alpha, the problems native to this kind of analysis are many.

The first thing you are likely to find if (and when) this particular bug bites you is that neither absolute nor benchmark-relative position weightings are a persistent indicator of their contributions to portfolio alpha. In other words, you’re going to look how you would have done if you just bought all of your best managers’ biggest positions, and wonder why it looks so lousy. I have developed hypotheses on manager sizing skill across hundreds of long-only and long/short managers, and haven’t been able to prove out those hypotheses out-of-sample in more than a handful of cases. This is not always a bug. Both absolute and relative position sizes are often a reflection of risk management or confidence in a non-negative outcome in a position, rather than an expression of a higher expected return. In my experience, big positions are usually portfolio ballast, and the names that drive whatever alpha might exist are unexpected positions of small-to-medium size.

The second thing you are likely to find is that sector-related security selection skill is painfully inconsistent. When we discover that we can’t build a portfolio out of our managers’ biggest absolute or active positions, the next step is usually to try to identify what they’re good at. We know that Manager X has done quite well historically – and has a good analyst or PM – in this sector, and that Manager Y has done quite well in this sector. What if we simply built a Frankenstein’s monster out of the positions from our best funds in each sector? There are invariably two problems with this: liquid markets manager skill within sectors is more inconsistent than most investors believe, but more importantly, positioning almost always incorporates portfolio construction necessities of the rest of the portfolio you’re ignoring. Except for dedicated sector funds, you can rarely know that the positions a manager has are an unbiased expression of their conviction since both holdings and size will often reflect the risk profile of names in other sectors within that portfolio.

The third thing you are likely to find – and this is the thing which usually blows up the beautiful dream for most investors - is that fund managers themselves rarely really know what their best ideas are. It is a truth, I think, that not every trade can be a great one, but that a great trade can come from anywhere in a well-managed portfolio. If you haven’t been going through the process of developing a sense for this among your portfolio managers over long periods of time, consider doing so.

It’s also useful for reminding yourself of the very frustrating truth that good investors don’t always know why they’re good.

This last point is of particular importance for the most common implementer of “Best Ideas Portfolios” – the financial advisors and individuals who meet with PMs or wholesalers with an aim to find 2 or 3 individual stocks they might put into a client’s portfolio because they are that manager’s best ideas. No matter what they tell you, I’m telling you: they don’t know.

Now, maybe you’ve had some historical success doing this. Even if that’s the case, I’d ask you to take a really long look at the evidence behind what you’re doing, and your confidence in the analysis of what your managers are good and bad at. And yes, with any such rule, there are going to be exceptions. As it happens, I have one in mind, too.

But that’s fodder for an upcoming longer form note, and unlike this one, it will have everything to do with narrative.