Even Fund Managers Do It: The V-Shaped Disposition Effect

A portfolio manager, two positions, one afternoon

Picture a mutual fund’s morning meeting. On the screen are two positions from the same portfolio. One has appreciated sharply since entry — a standout position, the kind that gets mentioned in the quarterly letter. The other has declined sharply — a position the manager has been carrying for months, the kind that gets asked about on every call with the investment committee.

Neither position has an obvious reason to close today. No earnings print, no analyst downgrade, no change to the original thesis. And yet both are more likely to be sold this week than the handful of positions sitting quietly in the middle — up a little, down a little, unremarkable.

That instinct — extremity itself pulling the trigger, in either direction — turns out to have a name and a paper behind it.

The disposition effect, and the twist in the tail

The disposition effect is one of the oldest documented patterns in trading behavior. Terrance Odean’s 1998 work described it simply: traders tend to sell winning positions earlier than losing ones, closing out favorable outcomes quickly while holding declining positions longer, often much longer than the original thesis would justify. The behavioral logic underneath it traces back to Kahneman and Tversky’s prospect theory — losses loom larger than equivalent upside moves relative to a reference point, so realizing a loss feels different from realizing a favorable outcome of the same size, and traders manage that asymmetry by avoiding the loss realization altogether.

For twenty-plus years, that was the shape of the story: a slope, not a curve. Sell winners fast, hold losers slow.

An and Argyle’s 2021 study in the Journal of Financial Markets looked at this pattern in a population you might expect to be immune to it — US equity mutual fund managers, professionals managing institutional capital, with research staff, risk committees, and mandates that explicitly discourage letting a losing position run unchecked. What they found wasn’t the simple slope. It was a V.

Fund managers, the study found, were more likely to sell both their big winners and their big losers — not just winners, and not just losers, but positions at either extreme of performance. The tendency to exit wasn’t a straight line from “appreciated a lot, sell” to “declined a lot, hold.” It bent back upward at both ends. Extremity, in either direction, was associated with a higher probability of the position being closed.

Why the shape matters more than the slope

It would be easy to read this as “professionals don’t have the disposition effect” and move on. That’s not what the finding says. It says the effect is there, but it isn’t operating exactly the way the classic retail-trader story describes it.

A pure loss-aversion story predicts one thing: reluctance to close losers, full stop, regardless of how large the loss has grown. A V-shaped pattern predicts something more specific — that as a losing position gets more extreme, the pull to close it can start to compete with, and eventually overtake, the pull to hold it. Something about a position becoming a large loser rather than a small one changes its treatment, and the change doesn’t move in a straight line.

There are a few candidate mechanisms in the surrounding literature, and it’s worth being honest that a single paper doesn’t settle which one is doing the work. Portfolio-level risk management could be part of it — an institutional overlay that forces an exit once a position’s decline crosses a threshold, independent of what the manager personally believes about the thesis. Mental accounting could be part of it too — closing out a position once it reaches an extreme, in either direction, can function as closing a “chapter,” resolving the ongoing tension of tracking it rather than reflecting a change in view. None of these explanations requires the manager to be acting on new information. That’s the interesting part: the V-shape shows up in the trading record even when nothing about the underlying company has necessarily changed.

The humanizing part

What makes this finding worth sitting with isn’t the shape of the curve on its own — it’s who was holding the account. The disposition effect is often discussed as a retail problem: individual traders without formal training, without a risk desk, without anyone reviewing their book. An and Argyle’s sample removes all of those excuses. These are professionals whose job is explicitly to manage a reference-point bias out of their process, evaluated by committees, benchmarked constantly, trained to treat a position’s cost basis as irrelevant to whether it should be held today.

The pattern showed up anyway.

That’s not an indictment of fund managers. It’s closer to the opposite — evidence that the disposition effect, in whichever shape it takes, isn’t a failure of discipline or sophistication. It’s a documented tendency in how people relate to a reference point once a position has moved a long way from it, and training or credentials don’t appear to make someone exempt from it. The research describes the behavior; it doesn’t describe a personal shortcoming.

What the footprint looks like

If you were looking for this pattern in a trade history — retail or institutional — you wouldn’t look for a single number. You’d look at exit frequency plotted against how far a position had moved from its entry, in both directions, and check whether the relationship is a slope or a curve. A pure disposition effect shows up as a downward-sloping line: exits cluster among winners, losers linger. A V-shaped version shows up as exits clustering at both tails, with the quiet middle of the return distribution holding the longest.

It’s a distinction that’s hard to see from a single trade — a distribution question, visible only once a body of trades is laid out against how far each one moved before it closed. Which is, in the end, the more interesting question raised by a study like this: not whether the disposition effect exists, but what shape it takes in a given set of decisions, and what that shape implies about the underlying process — professional or otherwise.