Even the pros sell worse than a coin flip

2026-10-21 · behavioral-finance, investing · about the paper: Klakow Akepanidtaworn et al., 2023

Watch yourself the next time you prune your portfolio. You probably don't review every position. You sell the thing that's been annoying you, the big red loser glaring off the screen, or maybe the shiny winner you want to lock up before it escapes. It feels like a decision. A paper in the Journal of Finance suggests it's closer to a reflex, and that the reflex is worth real money.

(Akepanidtaworn, Di Mascio, Imas and Schmidt, "Selling Fast and Buying Slow", Journal of Finance 78(2), 2023)

TL;DR: Across 783 professional equity portfolios and 2 million sells, the pros' buys beat a random pick by about 100 bps a year per dollar traded. Their sells lost to a random pick from their own holdings by 80 bps a year. Buying and selling turn out to be different skills, and selling is the one even experts don't have.

The data

This isn't a survey of who thinks they're good at selling. The researchers got trading records from Inalytics, a firm that monitors institutional money managers: 783 long-only equity portfolios from January 2000 to March 2016, average portfolio around $573 million, covering 2.4 million buys and 2.0 million sells.

How you grade a pro against a coin flip

The clever part is the benchmark. For every trade, they ask: what if the manager had instead bought (or sold) a different stock they already owned that same day, one picked at random among holdings with similar size, value, volatility and momentum? Now every trade gets a fair counterfactual, and skill is the gap between what the manager did and what a coin flip would have done.

For buys, the pros cleared the bar. For sells, they tripped over it.

The scoreboard

Put the sell number in perspective. Mutual fund fee ratios run around 20 to 50 bps a year, and people lose their minds about fees. These professionals were burning 2 to 4 times the fee on the sell side, relative to just selling a randomly chosen position. Both measures predicted actual fund returns too, buys adding and sells subtracting, so this isn't a statistical ghost.

They sell whatever's loud

The mechanism is the best part. The researchers call it neglected ideas: the sells cluster on extreme positions, the very largest and the very smallest weights, the stocks that moved a lot recently. In other words, managers sell whatever is salient in their attention right now, not whatever their analysis says to exit.

The kicker: when a manager's sells avoided the extremes, the underperformance disappeared entirely. The skill gap isn't about knowing less, it's about attention getting captured by the loudest thing on the screen.

Taylor and Riley prune a portfolio

Taylor and Riley each hold a $100,000 portfolio. Taylor prunes by feel, twice a year, selling whatever has been bothering them. Riley reviews every position on a schedule, once a quarter, against criteria written down back when the purchase was calm and unemotional.

If Taylor's sells run anywhere near the professional gap in the study, that's roughly $800 a year burned versus a coin flip, on sells alone. Riley's whole review process, the calendar reminder and the hour with a spreadsheet, costs one evening per quarter. The expensive part of Taylor's approach isn't the fee, it's the intuition.

There's a registered-account bonus for Canadian readers. Inside a TFSA or RRSP, a sale has no tax bill, so a rule-based exit costs you nothing extra. In a taxable account, every sell can be a capital gains event, which is one more reason to decide the rule before the position is screaming at you.

The honest caveats

These are institutions that opted into performance monitoring, so the sample skews sophisticated. It's long-only, the sell decision is "which position to exit", not "go short", and the window ends in 2016 before the quant-est era. Whether retail investors selling inside a TFSA do exactly 80 bps worse is a guess. The direction, though, matches a mountain of behavioural work: salience drives the sell button.

What to actually do

• Write the exit rule at buy time, while you're calm: "I sell when X happens", not "I'll know it when I feel it."
• Review every position on a schedule, not the ones shouting for attention. The study's fix and the data agree: remove the salience, the gap goes with it.
• If writing rules for every holding sounds like homework, outsource the sell decision entirely. One diversified fund that rebalances inside itself means the only sell you ever execute is the one that pays for the cottage.
• Know what your funds charge you either way, because a 20-to-50 bps fee is the same order of magnitude as the entire skill gap being measured here.

How strong is the evidence?

Scale: Strong = replicated across multiple peer-reviewed studies · Moderate = one solid peer-reviewed study or consistent findings in limited settings · Suggestive = working paper, small samples, or not yet replicated · Thin = non-peer-reviewed source or conflicting evidence.

Next steps

• Your portfolio can tell when you're tired — attention strikes again, this time by sunset.
• The skim view: digest + full PDF — every claim above, one tap from the source.
• The paper itself — Akepanidtaworn et al, Journal of Finance 78(2), 2023.