My Approach To Position Sizing

by: Bram de Haas

Summary

Asking readers for investment related questions they struggle with I already received a couple of great ones.

I'm publishing these and answering them as best as I can.

If I fail, you might still learn something useful through one of the many brilliant readers on SA.

The first one is from Michael about position sizing.

May 12 at 10:51am

Position sizing and position building (when and how to average down or use momentum to leverage up?). Surely depends on type of stock and style of investing, but what are time-tested winning coherent strategies?

Lots of questions here but in the end it comes down to what's a time-tested winning strategy with position sizing and position building. I'll go into position sizing and save the other half for later. It's a very interesting problem that I'm sure lots of readers, as I do, struggle with. The first problem I encounter trying to answer it is we didn't define winning. It kind of depends what winning means to you: preservation of capital, minimizing career risk, maximizing annualized return or something else.

Because everyone will have another definition of winning, I'll discuss the different approaches taken by famous investors and their pros and cons.

You can size positions evenly, size them according to the benchmark you are measured by or size them in order of attractiveness on some risk/reward scale.

Sizing them evenly works best if you are able to generate alpha with your stock picks but are somehow not able to differentiate between one of your terrific ideas and just solid ideas. This is quite unlikely because it involves mostly the same skills. I'd suggest working harder on your ability to do so because it will be much easier to boost your returns adjusted for risk through sifting the great from the good ideas as opposed to raising the average level of attractiveness of all of your stock picks.

Sizing positions based on a benchmark isn't something I want to be cheerleading so I'll skip over this one.

The most interesting approach, most common under famous investors, is to size positions by attractiveness of the investment.

It all starts with the premise you are able to:

  1. Generate alpha with your stock picking skills.
  2. Be able to determine the difference between an idea with a little bit of alpha and one with a ton of alpha.

If you can't do 1), active stock picking is just not such a good way to spend your time unless you do it to improve your skills and expect to be able to generate alpha in the future.

If you can't do 2), you are better off running a portfolio with equal sized positions.

However if you are able to do both 1) and 2), how big should you bet?

The Kelly Criterion is a piece of math that is extremely helpful to take notice of. I first learned about it as a professional poker player, when I had to determine how much of my bankroll to risk at any given game. Both in poker tournaments and when betting on sports it is very easy to determine, if you collect accurate stats on your bets, what the optimal bet size is to optimize the growth rate of your bankroll (read portfolio). In investing, it is immensely complicated because you run several bets at once. Bets take a long time to resolve (holding period) and therefore, it is difficult to collect reliable statistics to determine your expectation due to skill.

However, playing around with and understanding the Kelly Criterion will help you make better decisions.

There is a big problem with the Kelly Criterion however. The losses you will have to stomach in the short run in order to optimize growth rate of the portfolio for the long run will drive most people nuts. It can be extremely painful at times and at such times, you will question (perhaps for good reason and perhaps not) whether you have an edge at all or actually a loser destroying your portfolio.

Even if no amount of losses ever fazes you. You remain as calm as Yoda in the face of the destruction of the Republic. Your bosses may (if you are a professional investor), and you get fired for optimizing portfolio growth rate but going through a losing streak. If you don't have bosses and face no career risk, you may face divorce risk if your spouse doesn't like to see 70% of your portfolio evaporate because you are optimizing growth rate. These are real world additional risks that aren't represented by the formulae. Practitioners tend to adjust for these and move down from full Kelly to ½ or ¼ Kelly.

The guys with the most career risk and who generate the least Alpha will stay fairly close to the benchmark. From all perspectives, this makes sense. The most independently minded and guys with the most "permanent capital" (capital that can't get withdrawn) will bet closest to Kelly. A table I pulled from Gurufocus shows what % of capital several Guru investors put into their top 10 holdings:

Source: Gurufocus / portfolio value in million $ / number of stocks / allocation in %

A few guys on the list are diversified through other vehicles like Blum or are traders who also hold non-equity assets and shorts. They are in reality more diversified than the table would suggest. You will also notice a large number of investors who actively engage with their portfolio companies like Lampert, Winters, Ubben and Ackman. Because this takes time, energy and money but there is the potential for them to "create" Alpha it makes sense to load up.

The investors taking the exact opposite approach are not exactly slouches:

Source: Gurufocus / portfolio value in million $ / number of stocks / allocation in %

The traders with the lowest concentration of assets in their top 10 picks are Jim Simons and Joel Greenblatt who likely have a more statistical approach to investing as opposed to a deep research one. Yet, they are both famous for their incredible returns. Given they hold 3k and 800 stocks, they clearly also bet very heavily on their favorites. It should be clear from this you should never find comfort in Jim Simons also holding a stock you like.

Both holding lots of stocks and holding a few stocks can work but which approach you take should probably be related to how you achieve Alpha. If you do it by deep research or engagement that works better with a more concentrated portfolio. If you generate Alpha by finding pockets of inefficiency in the markets or use the Mauboussin approach of identifying the one thing the market doesn't know, you could be better served holding more stocks.

Something I don't love is when people hold eight stocks, weigh equally and it's a passive portfolio of their best ideas. If you have the skills to pick eight stocks that offer Alpha, you should be able to improve on that strategy. It's okay as a step 1 but now take step 2 and select over eight stocks that offer Alpha and differentiate between the amount of Alpha they offer.

My approach to position sizing

I bet heavily on a few of my best ideas (when I publish these on Off The Beaten Path, I call them Back Up The Truck Ideas, I published only one in the last year) but still hold quite a lot of different stocks at any given time. I practice at differentiating between the amount of Alpha ideas should generate. Although I never buy a stock I don't believe will offer a better risk/reward compared to the market, there is a much larger number of them where my conviction is modest.

I size these picks small and they are also much more often getting replaced as a better opportunity can easily emerge. I sacrifice a little bit of return, by diversifying outside of my high conviction picks, in order to more easily survive when I inevitably get hit by holding an Enron, Valeant (VRX), Horsehead Holdings (OTCPK:ZINCQ) type of pick. It's going to hurt but I will be able to make it back. This style of position sizing suits me very well as I have an appetite for illiquid issues and currently also high beta ones.

Disclosure: I am/we are long ZINCQ.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.