Unknown unknowns

Reports that say that something hasn't happened are always interesting to me, because as we know, there are known knowns; there are things we know we know. We also know there are known unknowns; that is to say we know there are some things we do not know. But there are also unknown unknowns — the ones we don't know we don't know. And if one looks throughout the history of our country and other free countries, it is the latter category that tend to be the difficult ones.
Former United States Secretary of Defense Donald Rumsfeld

The biggest change in my investment thinking this year has been regarding appropriate levels of concentration vs diversification.

Previously, I've agreed with the school of thought that says it is better to "keep all of your eggs in one basket and watch that basket carefully." The investment analogy is to own only a handful of stocks (5-10), but to intimately know each of your businesses. This approach makes intuitive sense and is perhaps most closely identified with Warren Buffett (who at one time had 40% of his partnership's assets in American Express) and Charlie Munger (who typically holds only 2-3 stocks in his portfolio). It is also a technique emulated by many value funds and investors, who regularly quote Buffett and Munger. But I'm beginning to think that it may not be the best approach for me.

The root of my evolving skepticism is the belief that anyone can know, with high probability, the future prospects of any particular company. If you do enough research you might be able to get that probability slightly above 50%. But to assume that you can do much better (without inside information), is probably a fallacy. The simple reason for this is that I believe the "unknown unknowns" - things you don't know you don't know - account for a large chunk of uncertainty, and are unresolvable no matter "how closely" you watch the basket.

I don't have a mathematical proof for this, but I do have some anecdotal thoughts and personal experiences that have shifted my thinking towards this direction:

  • Buffett success at concentration may simply be an example of survivorship bias. It's important to note that I don't think the value approach is a fluke - Buffett takes on this argument in his paper on The Superinvestors of Graham-and-Doddsville; but even in the examples he cites, Walter Schloss owns "well over 100 stocks", Bill Ruane avoids "excessive concentration", and Tweedy, Browne built their record with "very wide diversification."
  • When I read "professional" analyses of companies in industries I know very well (online travel, online advertising), they are often embarrassingly off the mark in terms of assumptions, risks and opportunities. It is unlikely that my analyses are much better in other industries.
  • When running my own businesses, in industries I know better than most, and with the benefit of "inside information", I haven't been able to predict the future with much accuracy. While I can quantify specific opportunities, and outline what I think are the most salient risks at the time, it is almost always the "unknown unknowns" (i.e. things I haven't even thought of) that determine the fate of the company, for better or for worse.
  • In my limited experience with angel investing, I've had essentially "unlimited", personal, face-to-face access to management of companies. Yet, although I pride myself on my personal judgement and have had this privileged access, there are entrepreneurs who have turned out to be complete, unexpected frauds. And I am not the only one they've fooled. The idea that an investor can reliably assess the "honesty" of management by listening to earnings calls, reading proxy statements, or even participating in limited face-to-face meetings, is probably an illusion.
  • Despite intense public scrutiny and supposed checks-and-balances, accounting fraud still exists and could have permanent, catastrophic consequences on a concentrated position; not to mention lesser, yet still material, accounting shenanigans that are more widespread. The idea that I'd be able to reliably detect these is probably a fallacy.
  • Many people, much smarter than I am, have spectacularly failed due to unanticipated consequences.

If one accepts that there is a large amount of irreducible uncertainty - unknown unknowns - what is the "right" approach? The simplest defense seems to be adequate diversification. What is "adequate" probably varies for each investor, but I've settled on a target of 15-20 total stocks in my portfolio. This seems to capture most of the benefits of diversification, while still being small enough for me to keep up to date with each position, and to have each company be a meaningful stake.

To give myself added flexibility, and allow myself some room for subjective judgement, I've set a size range of 3-7% for each holding, with the max position (7%) reserved for "top-20 lifetime" investment type ideas. The important part will be in adhering to these guidelines even when a stock appears to be "on sale"' - i.e. resisting the urge to "double down" after a stock has declined precipitously, if I'm already at my max position size limit - precisely to protect against these unknown unknowns.

(As a side note, I've been surprised how many fund managers / investors justify an unnaturally large position size, in violation of their own previously established guidelines, simply because the "opportunity is too good", the "circumstances are unique", or the "situation is temporary"; this how most stories of catastrophic loss begin, and precisely the type of overconfident situations guidelines are supposed to protect against)

The intent of this post isn't necessarily to argue that extreme concentration is the wrong approach - it's obviously worked for some people, has made most entrepreneurs their fortunes, and I was a great beneficiary of it last year - but the more I've thought about it, and reflected upon my own personal experiences, the less sense it makes to me as a long term, sustainable method for investment success.


If so much is unknowable, and the chances of an individual stock's success are little better than chance, how does one make any money? I think the answer lies in combining adequate diversification with the concept of margin of safety. The net effect (ideally) is that even if you are only right 50% of the time, as long as your downside is less than your upside, and you can still make money.

Show Comments