Click here for search results


Site Tools

Spite and Development: Altruistic Punishment vs. Spiteful Punishment

Research Brief
October 15, 2008, Karla Hoff

An experimental trust game conducted in one of the least developed states in India shows that spiteful punishment is surprisingly widespread, and that it could also be a considerable barrier to social and economic cooperation

Every society has mechanisms for restraining opportunism, and informal social sanctions are a critical source of such restraints. Experimental evidence suggests that the willingness to punish violators of fairness norms, even at a net cost to the punisher, substantially reduces opportunism and free riding (altruistic punishment). By contrast, the desire to increase one’s relative payoff drives spiteful punishment, which hurts everyone when it curtails pro-social behavior.1

The trust game with third-party punishment is designed to reveal the willingness of disinterested third parties to engage in pro- and anti-social punishment

The experimental trust game with third party punishment has three players, A, B, and C, who receive 50, 50, and 100 rupees, respectively, from the experimenter (50 rupees equals roughly one day’s unskilled wage). Players A and B participate in the trust game. Player C is the “enforcer.”

Player A can choose to keep his endowment (and then the game ends) or to invest it by sending it to B. If A sends his endowment to B, the experimenter triples the amount sent.

Spite and Development Figure At this point, player B has 200 rupees. He must decide whether to share the amount equally with A or keep all of it. If B shares with A, all 3 players have 100 rupees before the third player, C, makes a decision. If instead B keeps all the money, the distribution before C makes a decision is 0 for A, 200 for B, and 100 for C.

Player C, the “enforcer,” has more than two options. For every coin he spends on punishment, B loses 5 times as much. C can choose to spend nothing, spend enough so that B is empty-handed, or spend anything in between.

To summarize, if A keeps his endowment instead of investing it, then A and B each earn 50 rupees from the game. If A invests his endowment and B sends half the money back to A, then they both have 100 rupees in their pockets, so there is a large mutual gain from trust and cooperation. If B chooses to take advantage of A, then B would have the whole pot, but what he actually goes home with depends on C’s decision to punish him, and by how much. C can punish cooperation as well as opportunism.

The trust game captures two features of market exchanges and investment decisions

Trust and cooperation make the creation of an economic surplus for all players possible. But a trusting individual exposes himself to the risk of opportunism. Punishing opportunism mimics an individual’s reproaching the violator of a fairness norm (or ostracizing him through word of mouth). The action entails some cost to the punisher, but a much larger cost to the individual punished.

Not surprisingly, in the game, many, but not all, “enforcers” punished opportunism. The surprising finding was that “enforcers” also punished cooperators consistently across a wide spectrum of conditions, between 61 and 73 percent of the time.

On average, opportunists were punished more than cooperators, but anti-social punishment reduced the opportunism-deterring force of altruistic punishment by 50 percent. The incentive effect of pro-social punishment depends on the difference between punishment of an opportunist and punishment of a cooperator. When “enforcers” were asked why they punished cooperators, statements like this emerged: “I want B to lose.” “I am jealous of B. That is why it is important to impose a loss on him.” “I want to bring down B.” “Imposing a loss in the game gives enjoyment.” Other experiments in India reinforced the widespread presence of spiteful preferences.

High-caste participants exhibited spiteful preferences more often than low-caste participants

In a two-person game, when player 1 was given the opportunity to be nice—to raise the income of player 2 at a small cost—a large proportion of player 1’s chose to do so. However, high-caste 1’s were much less likely to undertake such altruistic acts if player 2’s gain was going to make player 2 better off than player 1.

Still another experiment showed that high-caste individuals were much less able to coordinate their actions in a joint investment game, compared to low-caste individuals. In the experimental game, it was in the interest of each individual to invest if his partner did, but each individual had to make his decision without yet knowing his partner’s decision. If one individual invested and his partner did not, then the investor would earn a low absolute and relative payoff. The lower ability of high-caste members to coordinate on the cooperative outcome may be due to concerns about status and a strong aversion to disadvantageous inequality.

How widespread is anti-social punishment and spiteful behavior across societies?

A study that implemented a public goods experiment in 16 comparable participant pools around the world found huge cross-societal variation.2 In some countries, players punished high contributors as much as they punished free riders, whereas in other countries, people punished only free riders.

The higher the anti-social punishment was in a participant pool, the lower was the average cooperation level in that participant pool. Societies with low anti-social punishment, and high cooperation, were societies that also had high scores on civic cooperation as measured by the World Values Survey and high scores on rule of law as measured by the World Bank.3

Both studies suggest that the best way to ensure social and economic cooperation is to limit reliance on informal norms as a substitute for formal enforcement mechanisms, especially where these are weak.

KARLA HOFF is a Senior Research Economist in the Development Research Group (Macroeconomics and GrowthTeam). Her research focuses on institutions and institutional change, particularly in the former Soviet Union and India. Email:


1. Fehr, Ernst, Karla Hoff, and Mayuresh Khsetramade. 2008. “Spite and Development.” American Economic Review, Papers and Proceedings 93(2): 494–99. (Based on World Bank Policy Research Working Paper 4619.)

2. Herrmann, Benedikt, Christian Thoni, and Simon Gachter. 2008. “Anti-social Punishment across Societies.” Science 319 (7, March): 1362–67.

3. Kaufmann, Daniel, A. Kraay, and Massimo Mastruzzi. 2007. “Governance Matters VI: Governance Indicators, 1996-2006.” World Bank Policy Research Working Paper 4280, July.

Further reading

Mullainathan, Sendhil. 2005. “Development Economics through the Lens of Psychology.”  Annual World Bank Conference on Development Economics 2005. New York and Washington, DC: A copublication of the World Bank and Oxford University. [with comments by Colin Camerer and Karla Hoff]

Permanent URL for this page:

© 2016 The World Bank Group, All Rights Reserved. Legal