Position Sizing Matters More Than Stock Selection

The Four Parts of the Job

Selection It's deciding what to own and it's the part that everyone imagines

Sizing is deciding how much which determines how much a given decision actually contributes

Risk management It is understanding what the portfolio as a whole is exposed to which is often different from the sum of the individual theses

communication with client You are explaining all this to the people whose money it is and it consumes much more time than outsiders expect

Why Sizing Dominates

A correct view held in a small position contributes little. An incorrect view held in a large position can undo a year of good decisions

Size has to reflect conviction but conviction is exactly what is most difficult to honestly evaluate in your own analysis. It also has to reflect liquidity since a position that cannot be exited at a reasonable cost is more risky in terms of what its dollar value suggests

And it has to reflect correlation with everything else being held. Ten positions that all express the same underlying view are one position regardless of how it looks on a stock list

The liquidity constraint is most often left as a vague preference when it can be simply calculated. Express the position as a number of days of the security's normal trading volume assuming that you are willing to be a modest fraction of that volume without moving the price. A position that takes two days to exit is a position. A position that takes thirty days to exit is a commitment and it will take those thirty days precisely when everyone else has decided to exit as well. Therefore the same dollar amount in two different securities can completely pose risks.different and only one of the two appears in a report ordered by weight

The Arithmetic That Makes Losses Expensive

The reason size trumps selection is not a matter of emphasis. It follows from the fact that profits and losses do not accumulate symmetrically

A position that falls 10 percent needs to rise 11.1 percent to get back to where it started. Going down 20 percent requires 25 percent. Going down 50 percent requires 100 percent and going down 75 percent requires 300 percent. The requirement accelerates so each additional unit of loss costs more to undo than the previous one

That curve is what makes a large position in a wrong idea much more damaging than a small position in a right idea. The advantage of being correct is that the position size is approximately linear. The cost of being wrong is not because capital destroyed is capital that is no longer available to capitalize on each subsequent decision

It also explains why the manager who avoids the catastrophic sizing error often outperforms the best analyst. Selection determines how often you get it right. Size determines whether you can survive being wrong and survival is a precondition for the rest of the record

Sizing From Risk Rather Than From Conviction

The practical implementation that most managers converge on reverses the intuitive order. Instead of deciding on a position size and then asking what could go wrong decide how much you're willing to lose on the idea and get the size from that

The arithmetic is simple. Risking 1 percent of the portfolio on an idea whose thesis would be refuted by a 20 percent adverse move implies a 5 percent position of the portfolio. The same 1 percent risk on an idea with a 40 percent tolerance implies 2.5 percent. On a tighter idea with a 10 percent tolerance it implies 10 percent

What this does is separate the two judgments. How much I believe this and how far I can go before I'm proven wrong are different questions and combining them into a single idea of ​​correct sizing is where most sizing errors come from

The formal version of the same idea comes from the mathematics of betting where the optimal fraction of capital to commit rises with the edge and falls with the odds against. Professionals who use it almost universally use a fraction of what the formula suggests because the formula assumes that the edge is known and in the markets estimated and overestimating the edge produces sizes that are much too large rather than slightly too large

What the Portfolio Is Actually Exposed To

The recurring finding is that a portfolio constructed entirely from company-specific analysis turns out to have a large systematic exposure that no one chose. A manager who selected fifteen companies based on their individual merits may have made a leveraged bet on interest rates

Risk systems exist to bring to light exactly this and interpreting their output is a central part of their function

Interpret is the key word because the result is an attribution rather than a verdict. A report that shows a large loading on a particular factor is not automatically a problem. A value-oriented manager must have a bias toward value and a report that indicates this confirms that the strategy is doing what it says. The finding that matters is the exposure that no one intended and no one would defend if asked which is why the useful discipline is to write down the expected exposures before reading the report and treat each difference as something that requires aexplanation

Diversification That Only Exists on the Holdings List

The correlation point deserves its own arithmetic because it is very easy to miss the error when looking at a list of names

Let's consider ten positions of 5 percent each all selected independently and all based on the same underlying assumption about the direction of rates or the durability of a particular demand cycle. The holdings page shows ten lines and a maximum position of 5 percent. The actual exposure is 50 percent of the portfolio in a proposal

There was nothing wrong with individual decisions about size. Each of them was made carefully in isolation and the concentration was brought together by accumulation rather than by a single choice. This is why aggregate exposure must be measured directly rather than inferred from the size of the largest holding and why the number of names in a portfolio is almost useless as a measure of how diversified it is

Crowding is the same problem that extends beyond the fund. A position held by many similar managers is one that will be sold by many similar managers under the same conditions so its liquidity in a stressed situation is much worse than its normal trading volume suggests. This is an exposure to other people's redemptions and does not appear anywhere in the analysis of the underlying business

The Discipline That Is Hardest

Selling is more difficult than buying. The disposition effect operates on professionals as well as individuals and the position that does not work carries the accumulated reasoning that justifies it

Practices that help are theses written with falsifying conditions set scheduled reviews of each position regardless of performance and treating each participation as a new decision about whether to own it today

Managers who articulate a clear sales discipline tend to have thought about it. Managers who only describe their buying process usually don't

Adding to a Loser Is a Sizing Decision

The framework's toughest test is what happens when a position moves against you because two defensible responses point in opposite directions

Averaging down is correct when the thesis is intact and the price is better which is the whole logic of buying something you consider undervalued. It is also the mechanism by which a manageable error becomes a fatal one since the position grows exactly as the evidence against it accumulates

The distinction that separates the two cases is whether the falsification condition written in the entry has been met. If the thesis specified what would prove it wrong and that has not happened a lower price is actually more attractive. If it has happened adding is a decision to override the analysis with the discomfort of realizing a loss

This is the real reason why a written thesis is important and has nothing to do with record keeping. The judgment about what would constitute disconfirming evidence must be made before there is money involved in the answer because afterwards the reasoning is done by someone who is already depressed

Sizing Happens Again Every Day Whether You Decide It Or Not

A correctly sized position at entry does not stay correctly sized. Prices move and the weight of each holding moves with them without anyone making a decision

The direction of drift is systematic. Positions that have worked grow as a proportion of the portfolio and positions that haven't shrunk so a portfolio left alone concentrates on what has already appreciated. That's the reward for being right or an accumulation of unmanaged risk and which depends entirely on whether the original thesis still supports the broader position

The two disciplines addressing this issue are at odds with each other and both have serious advocates. Rebalancing to target weights trims winners and adds laggards strengthening the sizing framework and automatically selling things that have risen. Letting winners ride accepts concentration on the argument that the biggest contributors to long-term returns are a small number of positions that have been allowed to grow and that trimming them is the most expensive habit in business

The compromise that most managers come to is to distinguish between a position that has grown because the thesis is developing and one that has grown because the price increased faster than the thesis was advancing. The first has gained its weight and the second has not. That distinction again requires the written thesis since without one there is no way to know which of the two happened and the default is to hold on to what emerged and call it conviction

The Constraints That Shape Everything

Most managers operate within a mandate: a benchmark a tracking error budget position limits sector limits and liquidity requirements

These limit the strategy more than the manager's opinions. A high-conviction idea that would violate a position limit is sized to the limit not the conviction

The client base matters similarly. Capital to be withdrawn during a period of poor performance forces a shorter horizon regardless of what the manager believes about the long-term value

How Performance Is Judged

Against a benchmark over periods that are statistically too short to be meaningful by clients who are subject to recency bias

Managers know that evaluation is loud and yet they are evaluated by it. This is the source of the professional pressure that produces reference hugs and it is a rational response to how the industry measures people

The pressure is on size rather than selection which is the part worth highlighting. A manager worried about deviating from a benchmark usually does not stop having opinions. They express the same opinions in smaller increments so the portfolio is filled with positions sized too small to matter and tracking error is still comfortable. The result reads like a diversified active portfolio and behaves like an expensive index fund and no individual decision was obviously wrong

The Bottom Line

Portfolio management is selection sizing risk management and communication and the last three get less attention than they deserve. Size determines how much each decision contributes aggregate exposures often differ from the sum of the individual theses and sales discipline is the hardest part to maintain. Mandate constraints shape strategy more than management opinions. The arithmetic behind all of this is that losses stack up against you faster than gains which is why the size of the error matters.more than the frequency of perception