ADAM SMITH remarked that no dog ever made a free and fair exchange of one bone for another. Many pooch-owners will agree, having spent frustrating minutes trying to wrestle a stick or a ball off their pets.

But humans are supposed to be rational. They should not value things they own more highly just because they already possess them. Experiments in the classroom have shown, however, that people may be subject to this trait, dubbed the “endowment effect”. In a classic study, students were randomly given a mug. Those who received the mugs were asked what price they would sell them for; the mean was $5.78. Those who did not were asked what they would buy them for; the average was $2.21.

In a similar test, students were asked to state their preference between a mug and a chocolate bar; 56% favoured the former and 44% the latter. Two other groups of students were then randomly assigned the mug or the chocolate and then asked whether they were willing to swap; in each case, around nine in ten of the students were unwilling to do so. You might say they were “possessed”.

That is the kind of result that gets behavioural economists excited, since it suggests people are not the rational calculating machines (Homo economicus) that standard models assume them to be. Traditional economists, by contrast, are sniffy about such experiments, arguing that they bear little relation to the kind of decisions made in the real world.

But a new paper* claims to spot the endowment effect at work among a supposedly hard-headed group of individuals: stockmarket investors. It looks at the flotations—initial public offerings, or IPOs—of companies on the Indian market.

When Indian IPOs are oversubscribed, issuers often use a lottery to assign shares to eager investors. Winning investors are allocated shares at the flotation price; losing investors can buy only after the issue starts trading.

In theory, both groups should be equally keen to own the shares. After all, they applied on the same basis. But the authors found that 62.4% of winners were still holding the shares after a month, while just 1% of losers had bought them. Now, these groups will have paid different prices, since the losers have to buy in the secondary market; the average IPO increases in price by 52% on the first day of trading. On the surface, it makes perfect sense that the losers would not want to pay what they perceive to be an inflated price.

However, if this were the driving force, you would expect to find more losers buying IPOs when the initial gain was small or when prices fell. But the authors found no relationship between the endowment effect and early price movements.

Furthermore, in cash terms, IPO investors’ trading gains are limited. In order to apply for shares, investors have to put an average of around $1,750 into an escrow account, but are allotted shares worth only around one-eighth of that. That means the average first-day trading gain is just $62. Would this really drive such a big difference in behaviour?

Another puzzle is that, after its initial gain, the average IPO falls in price as time goes on; after 12 months, it is 54% below the issue price. So it would make sense for the lottery winners to offload their holdings after the first day’s trading and for the losers to try to buy the stock on the cheap later.

But that doesn’t happen: after a year 45.8% of the winners are still hanging on to their stock, whereas only 1.7% of the losers have piled in.

Perhaps inertia is the driving factor? Investors simply might not get around to trading. But the authors reject that explanation; the endowment effect is still present when they study investors who make more than 20 other stockmarket trades in the month of the IPO.

Another possibility is that those winners who hold onto their shares are naive investors; more experienced traders are more rational. So the authors looked at investors who had taken part in more than 30 IPOs, and found there was still a substantial endowment effect; winners were four times more likely than losers to hold shares at the end of the first month.

It looks, therefore, as if the act of winning the IPO lottery makes investors more willing to hold on to their shares, regardless of their profit or loss. So next time your dog refuses to give up his ball—the very ball he expects you to throw for him to fetch—show him some sympathy. He is only imitating his human owner.

* “Endowment Effects in the Field: Evidence from India’s IPO Lotteries”, by Santosh Anagol, Vimal Balasubramaniam and Tarun Ramodorai.