I admit that there is no type of result I enjoy encountering more than one that seems so counterintuitive, so against accepted wisdom, so surprising that I didn’t see it coming at all. Even more enjoyable is the rare occasion when that epiphany is my own. Alas, my subject today only fulfills the first criterion, as sadly, this particular insight wasn’t mine…
The related topics of, “How much active management is necessary?” (and, conversely, “How much indexing would start to be a problem for market efficiency in both pricing accuracy and liquidity?”) and the more speculative, “What would happen if everyone indexed?” are perennials that have fascinated me and many others for many years. But all such discussion always runs into the problem of Bill Sharpe. Well, not Bill in general, but specifically his observation that, properly defined (no easy task itself since this involves proper definition of the all-encompassing capitalization weighted investable market portfolio), all active management must net to zero (before fees and trading costs; and thus lower than passive returns after these subtractions).
We run smack into Bill’s math all the time. For instance, it seems obvious that when most of the world indexes, prices would become less efficient and active management easier. But there’s still Bill with which to deal! As Jack Bogle has admonished me many times, even in that world the average can’t beat the average. Our intuition is that active management would be easier in such a world, and more would rush to do it, restoring some kind of reasonable equilibrium; but running face first into Bill’s math makes it hard to stay confident in that intuition. Indeed, if active keeps losing after fees no matter how inefficient the market becomes, Bill’s logic seems to imply that the active management industry eventually should disappear. Horror for capital markets may ensue (maybe even worse than Marxism).
Maverick Capital's flagship hedge fund lost -22.9% in the first quarter of 2022 according to a copy of the firm's quarterly update, which ValueWalk has been able to review. The firm's flagship fund, Maverick Fundamental Hedge, accounts for $3.7 billion of the group's $8.1 billion of assets under management. Even after losses in the first Read More
So, along comes one of our partners, Lasse Pedersen. In many ways Bill’s original observation was obvious (one of those “obvious to me after someone smarter than me told me about it” type things). In a similar sense, so is Lasse’s. He points out that Bill’s identity is true (it’s an identity after all) but it’s a special case. It’s the case where managing a real world index portfolio includes no turnover; that is, the set of securities never changes. Lasse shows that once you add necessary turnover into real world indexing – which can come from additions and deletions to the index or from corporate actions – the ironclad tyranny of Bill’s math is broken. And once you add information asymmetry (not a big assumption as indexers, by definition, claim and have no information) you restore an important role for active management. Lasse also presents initial investigations of these more realistic cases, showing that the magnitude of the effects are not trivial (My first thought in reading Lasse’s draft was, “Sure, but this will be trivial.” Lasse says no!). Anyway, this is supposed to be an introduction to, not a full exposition of, Lasse’s work, which I obviously encourage you to read. I think it’s the beginning of an important, underappreciated line of thinking in a long, fascinating and important topic; and I think it achieves this by taking on something we all assumed was obviously true. How many of us often just cite “the average can’t beat the average” and move on assuming we’ve smashed whoever we were debating? I know I have done it many times. Sharpe’s insight is, in my opinion, likely still mostly true, perhaps entirely true. But Lasse has, at the very least, created some doubt for me in something I was pretty sure about.
Finally, some caveats. I think Lasse would be the first to agree that indexing is still the primary investor friendly financial innovation of the last 50 years (I certainly believe this – Jack Bogle is not going to be dethroned as my investing and business hero by Lasse any time soon!). Nothing Lasse has written says that indexing isn’t a gigantic positive for investor welfare. Nothing Lasse has written says that active management is generally a good deal for investors as it stands now (it can still be pursued by too many for society’s sake and be too expensive for what it provides). In my view Lasse’s current work does not mean it’s easy for you to find, ex ante, a good active manager net of fees. And, most importantly, Lasse doesn’t repeal Sharpe’s arithmetic when there isn’t the needed turnover, which may be much of the time. That Lasse finds a role for active management in the presence of turnover still doesn’t jibe with how most active managers describe their attempted value add. Their self-description, and I think how most still think of active stock picking, sounds much more like taking on Sharpe’s ironclad no turnover case (i.e., they’re picking better stocks than the static index rather than Lasse’s version where they’re accurately pricing and providing liquidity for turnover). Furthermore, we still have to deal with the vast size of the active management industry. Is Lasse’s finding a role for them enough to justify such a big industry? Lasse makes no such claim, but proposes that a sizeable active management sector may be necessary as the industry must exist and thrive beyond the times it is most needed (pricing and providing liquidity for turnover). This seems to me like one of the tougher parts of his argument to swallow (though I don’t have a better one). There’s a lot more to do here.
But Lasse has re-opened a case that I thought closed and it happens to be one of the vital issues in investing. That’s pretty cool and more than enough for today.
 Notice I’ve switched from the theoretical true market capitalization weighted portfolio to a real world index portfolio. I’m not sure such a switch is even necessary, but I’m being careful as I think this may matter to parts of the argument.