Position sizing: equal, conviction and volatility weights

Compare three ways of sizing stock positions and pick the one you will use.

Decide how much money each stock gets by a rule, and write that rule down before you find your next stock. Modules 8 and 9 tell you what a company might be worth. They don't tell you how much of your money it should get. This module covers that, starting with position sizing: the three common methods, the mistake each one guards against, and how to choose between them.

How Marcus's position sizes came from when he bought

Marcus adds up his four single stocks and finds S$51,000. The bank is S$16,000, the chip designer S$15,000, Larkspur S$8,000 and the S-REIT S$12,000 (16,000 + 15,000 + 8,000 + 12,000 = S$51,000). Looking at the split, he sees it is one he never chose.

He bought the bank because a colleague recommended it in the year he got a bonus. The chip designer is S$15,000 because its price rose. Larkspur is S$8,000 because that was the cash he had the week he heard about it. He cannot remember why he holds S$12,000 in the S-REIT. Each position's size came from when he happened to buy. His risk dashboard from Lesson 2.8, "Build a risk dashboard for your portfolio and two benchmarks", had already shown the result: more risk than his benchmark, without more return.

Sizing a S$50,000 portfolio three ways

To compare the methods, take a portfolio of ten stocks worth S$50,000 (an invented example). Their yearly volatilities, from lowest to highest, are 15%, 18%, 20%, 22%, 25%, 28%, 30%, 35%, 40% and 45% (also invented).

Equal weighting gives every position the same amount of money. Here each stock gets S$5,000 (S$50,000 divided by 10). It is the simplest method. Because it allows no favourites, it stops you putting a large share of your money, such as a fifth, into the idea you like best.

Conviction weighting gives more money to the ideas you rate most highly. Suppose you rank the ten and give them 18%, 15%, 12%, 10%, 10%, 8%, 8%, 7%, 6% and 6% (invented weights that add up to 100%). Your top idea gets S$9,000 (18% of S$50,000) and the bottom two get S$3,000 each (6% of S$50,000).

The case for this method is that there seems little reason to give your tenth-best idea as much money as your best. The case against is that confidence and accuracy are different things. Most investors overrate their own judgement, as Lesson 6.2, "Overconfidence and the cost of trading often", in the course "Behavioural Finance: why you make the money mistakes you make" shows. Conviction weighting makes overconfidence worse, because it puts the most money where you feel most sure, and that is also where you are most likely to be overconfident.

Volatility weighting gives less money to more volatile stocks, so that each position adds a similar amount of risk. The method has four steps:

Take each stock's yearly volatility. Calculate 1 divided by that volatility for each stock. Scale those weights so they add up to the portfolio total, S$50,000 here. Check that each position's amount multiplied by its volatility comes out about the same.

In the example, the 15% stock gets about S$8,280 and the 45% stock about S$2,760, with the rest falling in between. Each position's amount times its volatility comes to about S$1,240: for example, 8,280 x 15% and 2,760 x 45% both come to about 1,240. In a typical year, each position would move the portfolio by roughly the same amount. The weakness is that volatility is measured from past prices, and Lesson 2.1, "Volatility: what standard deviation captures and what it misses", showed that a calm period can understate it.

What each method protects you from

Compare the largest and smallest positions. Under equal weighting every stock gets S$5,000. Under conviction weighting the range is S$9,000 to S$3,000, a ratio of 3 to 1, and your opinions set it. Under volatility weighting the range is about S$8,280 to S$2,760, also 3 to 1, and measured price swings set it.

Equal weighting protects you from yourself. Volatility weighting stops a few wild stocks from dominating the portfolio's swings. Conviction weighting protects you from nothing and only pays off if your rankings are good. You can test your rankings over time by comparing the returns of your top-ranked ideas with those of your bottom-ranked ones. If the two groups do about equally well, your rankings are not earning the extra money conviction weighting gives them.

You can also combine the methods. A common approach is equal weighting with a cap linked to volatility, so the most volatile stocks get a smaller share. A volatility cap is a smaller position target for any stock above a set volatility level. In this approach, conviction decides only whether a stock is in the portfolio at all.

Marcus's rule: equal targets with a volatility cap

Marcus chooses equal weighting with a volatility cap for his single stocks. Every stock gets the same target, set as a share of the whole portfolio, and any stock above 30% yearly volatility gets a smaller target. His own estimates put the bank at about 20%, the S-REIT at about 18%, Larkspur at about 35% and the chip designer at about 45%.

The bank and the S-REIT get the full target. Larkspur and the chip designer get the smaller one. Larkspur is also thinly traded, meaning few of its shares change hands each day, and cyclical, meaning its profits rise and fall with the economy. Under this rule the chip designer, his second-largest stock, would shrink, and Larkspur would stay roughly where it is. Lesson 8.6, "Sensitivity tables and scenarios instead of one target price", found that a bad cycle could take Larkspur to about S$1.05 a share, a loss of about a third. The smaller target limits how much that loss would cost the whole portfolio. Lesson 11.3, "Position and sector limits, and concentration risk", puts actual numbers on these targets.

Marcus writes one line in his workbook: "Single stocks are equal weighted at a set share of the portfolio, with a smaller cap for any stock above 30% volatility. Conviction decides what I own, never how much."

That line is his sizing rule: a written commitment, made before he chooses new stocks, on how much each position gets. Write yours now, before you find your next idea. A rule written in the excitement of a new discovery tends to get shaped to fit that discovery. A rule written in advance works as a commitment and makes your next decision smaller, because you only decide whether a stock goes in, and the amount is already set.

Take a made-up ten-stock portfolio of S$50,000, size it with each of the three methods, and compare the largest and smallest positions each method gives you. Then write your own sizing rule and note which mistake it guards against.

Course

Junxiong-WFG Organisation is an authorised representative of AIA Financial Advisers Private Limited (Reg. No. 201715016G).