Endowment effect

The tickets that went up in value

Months ago, someone bought two tickets to a sold-out comedy show for $40 each. Resale offers are now coming in at $150 a ticket. They say they would never pay $150 to see this comedian, but they turn down every offer: “No, I’m keeping them.”

Holding a ticket worth $150 on resale means passing up $150 to go to the show, which is the same trade as buying a ticket at that price. By their own account, they wouldn’t make that trade as a buyer. The only difference between the two positions is that one starts with the ticket in hand. This is the pattern Kahneman, Knetsch and Thaler described with a wine collector who would neither sell a bottle at the auction price nor buy another one at that price.

The pen you were handed

At the start of a workshop, organizers hand each attendee either a notebook or a pen, chosen at random, then later say anyone may swap for the other item. Very few swap, in either direction.

No money changes hands, and no one chose their item. If preferences didn’t depend on what you were given, the item people end up with would follow their tastes, not the draw: whatever share of attendees prefers notebooks, about that share of pen holders would swap, and roughly half of all attendees would trade. When most people keep whichever item landed in front of them, the random assignment is doing the choosing. This is the exchange form of the effect, which leaves prices, and the question of whether participants understood a pricing procedure, out of it entirely.

The auction you were winning

A collector has been the top bidder on an online auction for a vintage camera for five days, and has set a limit of $200. In the final minute someone outbids them, and they raise their bid to $260: “I can’t lose it now.”

The collector never owned the camera. Researchers have proposed that expecting to own something can shift the reference point the same way ownership does, so being outbid feels like losing something already held. If $200 was a considered limit, nothing about the camera changed in the last minute; only the sense that it was already theirs. This version is less clear-cut, and it shows the effect’s edge: the “endowment” may be a matter of expectation rather than possession.

• When ownership has created real value. Habits, sentimental attachment, skill in using the thing, or plans built around it can make an item worth more to its owner than to anyone else.
• When you know more about what you own. Your own car’s maintenance history is known to you; an identical-looking car for sale isn’t. Asking more than you’d pay for an unknown one reflects information.
• When trading has real costs. The time, risk and hassle of selling, replacing or waiting for an item can justify keeping it.
• When the item is a large share of what you have. Economic theory expects some gap between buying and selling prices when the amount involved is large relative to your wealth.
• When it’s a bargaining position. Asking high and offering low in a negotiation is a strategy, not a statement of what the thing is worth to you.

The test: if you’d been given its market value in cash instead of the item, would you spend the money to buy it back?

Grandfather’s watch

Someone inherits a watch their grandfather wore every day. A dealer offers $300, about what a similar watch sells for. They refuse, although they’d never pay $300 for a watch they didn’t know.

The asking price is higher than they’d pay for an identical watch, but it isn’t ownership alone at work. The watch has a history that no identical watch has, so to this owner the two aren’t the same item. That’s ownership having created real value, and it’s a legitimate reason to keep it.

The car you’ve maintained

A driver is selling a ten-year-old car they’ve owned since new, serviced on schedule and never crashed. They set the price above what they’d pay for the same model from a stranger’s listing.

It resembles the gap between selling and buying prices. But the seller is comparing a car whose history they know with cars whose history they don’t, and a buyer who could verify the records might reasonably pay more for it too. That’s knowing more about what you own.

Richard Thaler named the effect in 1980 as an example of loss aversion outside risky choice, and Kahneman, Knetsch and Thaler (1991) called it “a manifestation” of Loss aversion: giving an object up counts as a loss, getting it counts as a gain, and losses weigh more. In a 2014 review, Keith Ericson and Andreas Fuster concluded that loss aversion is still the leading explanation of the effect, while arguing that reference points are set by more than ownership, including expectations.

Other explanations compete with or refine it:

• Misunderstood procedures. Charles Plott and Kathryn Zeiler argued that the gaps in some classic experiments came from participants misunderstanding the pricing mechanism or reading signals into how the item was presented (see Evidence).
• Ownership itself. Some psychologists argue that owning something makes it seem better, which would raise an owner’s value for a second, identical item too. Ericson and Fuster review a study by Carey Morewedge and colleagues (2009) that reported this pattern, and note that procedural concerns may apply to it as well.
• What people think about first. On query theory (Eric Johnson and colleagues), sellers first consider reasons to keep the item and buyers first consider reasons to keep their money. Ericson and Fuster report that the gap disappeared in that work when participants were told to reverse the order.
• Inertia. David Gal (2006) argued that a general tendency to stick with the status quo is enough to explain the effect without loss aversion.

Is it a kind of loss aversion? Its originators classify it that way, but critics treat loss aversion as one explanation among several, and the empirical link is itself disputed (see Evidence). Because the sources disagree, this site doesn’t list the endowment effect as a kind of loss aversion. It is also closely related to the Status quo bias: Kahneman, Knetsch and Thaler used the same wine collector, unwilling to buy or sell, to illustrate both.

The endowment effect is sometimes confused with the Sunk cost fallacy. The sunk cost fallacy is about what you’ve already spent driving whether to continue; the endowment effect is about having something raising its value to you, even if it cost nothing. The mugs in the classic studies were free.

• Price it as a buyer. Ask what you’d pay for the item if you didn’t own it, or whether you’d buy it back with its market value in cash. This follows from the definition and hasn’t been tested as a remedy in the studies cited here.
• Experience may help. In John List’s field studies at sports card shows, experienced traders showed little or no reluctance to trade, and people who traded more over the following year became more willing to trade. The evidence that experience causes the change is suggestive rather than conclusive (see Evidence).

Status: contested. The gap between selling and buying prices has been reproduced many times and meta- analyzed. But credible studies have made it shrink or disappear with changes to procedure and with market experience, and whether it comes from loss aversion or from something else is seriously disputed. That meets this site’s criteria for “contested” on both counts.

The foundational work. In Kahneman, Knetsch and Thaler (1990), coffee mugs and other goods were given at random to half the participants, and markets were run between owners and non-owners. Far fewer mugs traded than the roughly half that standard theory predicts, and, as Ericson and Fuster summarize it, median selling prices were more than twice buying prices. When the same markets traded tokens redeemable for a fixed amount of cash, trading volume was as predicted, which the authors took as evidence that transaction costs couldn’t explain the result. An earlier exchange study by Jack Knetsch (1989), described in the same review, found that 90% of people given a chocolate bar kept it rather than swap for a mug, while only 11% of people given the mug chose the chocolate.

Reviews and meta-analyses. Tunçel and Hammitt (2014) meta-analyzed studies that measured both buying and selling prices for the same good. The gap was smaller for ordinary private goods than for public or non-market goods, smaller when participants had market experience or repeated trials, and smaller when the price-elicitation method gave people an incentive to answer truthfully. More recent studies found smaller gaps. An earlier review by Horowitz and McConnell (2002) is described by later studies as finding the same contrast, with much larger gaps for public and non-market goods than for ordinary ones.

Challenges from procedure.

• Plott and Zeiler (2005) gave participants extensive training and paid practice with the pricing mechanism, and kept their decisions anonymous. Under these procedures they observed no gap for mugs, and concluded that the results call into question reading the gap as evidence of loss aversion.
• Isoni, Loomes and Sugden (2011) reported that in Plott and Zeiler’s own unpublished data, and in new experiments, the same procedures produced a significant and persistent gap for lotteries. According to Ericson and Fuster, their new experiments also found no significant gap for mugs under either the original or the Plott and Zeiler procedures, which Isoni and colleagues took to mean that something other than those controls, possibly how participants were paid, explained the missing gap. Plott and Zeiler disputed the lottery results in a 2011 reply.

Challenges from the field. List (2003) ran trading experiments at sports card and collector pin shows. Among inexperienced non-dealers, only 3 of 44 traded the item they had been given; among experienced non-dealers, 14 of 30 did, close to the half expected with no endowment effect. A year later, 13 of the 21 earlier non-traders whose trading activity had increased chose to trade. List concluded that market experience plays a significant role in eliminating the effect; Ericson and Fuster note that self-selection into trading could also contribute.

Is loss aversion the cause?

• Chapman, Dean, Ortoleva, Snowberg and Camerer (2023), in a working paper using four incentivized representative surveys of 4,000 U.S. adults, replicated the gap for lotteries but found “little evidence” that it is related to loss aversion in risky choices. Buying and selling prices were at best weakly correlated with each other.
• Gächter, Johnson and Herrmann (2022), studying 660 customers of a car manufacturer, found loss aversion measured with endowment-effect tasks strongly correlated with loss aversion measured with lotteries.
• Simonson and Kivetz (2018) argued that the endowment effect and status quo bias, the most prominent support for loss aversion, are open to multiple alternative explanations.

What remains uncertain. The gap is real in many settings, especially for public and non-market goods and for inexperienced people. How large it is in functioning markets, which procedures remove it and why, and whether loss aversion, expectations, ownership or some mix explains it are all still open.