Position Sizing
A universal problem faced by portfolio managers of public market securities is the question of portfolio position sizing. In reality, 80% of the discussion is based on what level of volatility clients can handle and what the portfolio manager can handle psychologically. For the duration of this note I will make the supremely unrealistic assumption that portfolio managers have the steel moxie of the T-1000 and all client/LP money is locked up for 100 years.
Below is a quote from Chase Sheridan of Ruane, Cunniff, Goldfarb on their position sizing excerpted from the transcript of the 2016 Sequoia Fund annual general meeting:
… We do not size positions according to expected IRR because as you correctly note, you have to factor risk into the equation. There are certain investments in our portfolio that are real anchors, incredibly steady, defensive businesses that we know will do well in good times and bad. If the IRRs are the same, the steadier, more durable business deserves a higher weighting. We recently put something into the portfolio that we think has very good prospects. But we limited the size of this particular investment because there is some element of tail risk on the downside. We want to be aware of that. It is a not a formula. I do not think given the way we invest, for the durations we invest, that formulas work. It is a subjective judgment. But certainly, we factor risk into our decision-making process when constructing the portfolio.
For reference, Sheridan is alluding to the portfolio of long-only public market equities that are owned by the Sequoia fund. Sheridan is answering a question from an owner of the fund and is explaining in quite plain English the parameters by which one of the best long-only mutual fund managers has historically sized positions in their book. Unfortunately, most of what Sheridan discusses is tautological at best, and confusing at worst. Here, one should look to bifurcate defensiveness and durability of cash flows into:
- durability of cash flows as it relates to business intrinsic value- what discount rate is used to value the business, what visibility you have in looking out into the future with what steady-state mature growth rates (t+10, t+30 or maybe t+6months in the case of a net-net)
- durability of cash flows as it relates to downside risk and asset protection in the event of a downside event. You can think of this as loss given default or severity in the language of fixed income investors.
One of the best examples of bifurcating these risks is the case study of VeriSign. VeriSign provides critical internet infrastructure that is both irreplaceable and consistently demanded, providing a growing, recurring, revenue stream (that falls to the bottom line for the most part). This gives great visibility into how many domain names VeriSign will have registered in the future, and the capital profile necessary to maintain support of those domain names (basically $0). You could probably make a reasonable band of estimates around how many .com and .net domain names will be registered with VeriSign 30 years out. However, VeriSign faces one large existential threat: the risk that they lose their government contract or they somehow lose favor with ICANN and lose their monopoly. However remote this risk might be, it exists and whatever assumptions you’re making about VeriSign into the future you’re assuming they still have their ICANN contract.
(John Huber at Saber Capital has a great summary of the nature of VeriSign’s business (linked here) if you’d like to know more.)
Position sizing should be the mechanic by which you’re protecting the bankroll from what seemed like ex-ante durability and defensiveness that really ends up not materializing the way envisioned by the portfolio manager. VeriSign is the reductio ad absurdum comic book version of this- if your first principles are incorrect, the company is worth $0 (virtually no tangible asset protection in liquidation), and worth multiples of what it’s trading at if first principles are correct (contract is maintained, domain name unit demand is resilient and unit pricing can go higher).
So do we account for VeriSign’s perceived defensiveness in our forward IRR calculation in our valuation, or do we account for it in the size of our position? The right answer is, to steal from Sheridan, “a subjective judgement”, and based off of holding period. What are you going to get and when? How sure are you? What if you’re wrong?
A quick note on the Kelly criterion: as I understand the math, this formula is based off of knowable, a priori probabilities with the intent to maximize the longevity of a bankroll such that it exists long enough to see the favorable outcomes in a given set of probabilities eventually benefit the bankroll. I have no idea how to apply this to public market securities. I also don’t know how to size short positions after thinking about this for a great deal of time.
