Media Briefings

New Evidence Of The Impact Of Emotions On Economic Decisions

  • Published Date: October 2009

Emotions such as anger and guilt can have a powerful impact on the decisions we make
about whether to benefit ourselves at the expense of our clients, our colleagues or the
government. That is the central finding of research by Astrid Hopfensitz and Ernesto
, published in the October 2009 issue of the Economic Journal.
Their study, which uses an experimental two-player game, finds that participants are likely
to retaliate if they feel they have been treated unfairly, even if retaliation is costly. But even
more striking is the fact that if the ‘punished’ participant feels angry but not guilty, they too
will retaliate. This can lead to an escalation that is bad for everyone.
The researchers conclude that moral feelings are an important ingredient of ‘social control’
and need to be considered by economists and politicians. Understanding the role of
emotions can help to design policies that induce ‘socially desirable’ behaviours in markets
and societies.
Why should I report all my income to the tax authority – even though the chances of being
caught are very small? Why should I be honest to my clients – even though they might
have no way of verifying what I did?
For long, economists have been occupied studying how we act in social dilemmas such as
these. Understanding them is important for policies ranging from environmental regulations
to figuring out what went wrong with ‘immoral’ investors.
Social dilemmas are situations where everyone is better off if everyone cooperates, but at
the same time, each person has no (obvious) individual motivation to do so. Laws and
regulations are one way to enforce certain behaviours. But things get tricky for situations
where neither state nor police can observe everybody and enforce the rules.
One alternative is decentralised social enforcement. But under which conditions does such
enforcement work? Situations might escalate because the implied parties feel unjustly
punished by their peers. Do moral emotions, such as guilt, inhibit escalation and make
social control truly effective?
These are the questions approached in this study. It features an economic experiment in
which two players faced a social dilemma. Both players receive points that represent real
money. One player is asked to move first and can decide whether to ‘trust’ the second
player by transferring some of his points. If he does, these points gained in value and the
second player gets comparatively richer.
Now it is up to the second player to decide what should happen with this gain in wealth. He
can act fairly and split the gains equally, be selfish and keep everything for himself or take a
middle way by returning the investment but keeping a larger portion.
After this interaction, players can ‘penalise’ the other if they feel they have been treated
unjustly. But this is costly: players have to give up one point to be able to reduce the other’s
earnings by four points.
When a player decides to penalise his partner both therefore lose real money. In addition
after being penalised the punished party can retaliate and so on until one of the two stops.
The results show that social control ‘works’ and prompts around 60% of second players to
return high amounts in the first place. But it is especially interesting to look at those who
return less than the ‘fair’ amount. Most of these players receive punishment from their
The question is how do they react to it. The study finds that 40% of participants that get
punished decide to spend some of their money to retaliate. This is striking since these
players had initially breached a social norm and had higher earnings than their partner.
Instead of stopping the game by renouncing punishment, they decide to lose some of their
earnings and risk angering their partner even more. Indeed, in 55% of the cases they get
punished again. This leads in most cases to the complete destruction of both players’
resources. Escalation is thus highly inefficient.
The experiment shows that the experience of moral emotions reduces escalation. Of those
players that feel angry after being punished, 100% retaliate if they do not experience guilt.
None of those experiencing guilt decide to succumb to their anger.
Notes for editors: ‘The Importance of Emotions for the Effectiveness of Social
Punishment’ by Astrid Hopfensitz and Ernesto Reuben is published in the October 2009
issue of the Economic Journal.
Astrid Hopfensitz is at the Toulouse School of Economics. Ernesto Reuben is at Columbia
For further information: contact Astrid Hopfensitz on +33/(0)5 61 12 85 30 (mobile: +33/
(0)6 37 67 48 39; email:; Ernesto Reuben via email:; or Romesh Vaitilingam on 07768-661095 (email: