Typically, the disposal of clinical trial mice isn’t something an informed economist is usually preoccupied with. Yet it seems to be very much an economic question when set in the context of bargaining, admittedly under different circumstances as demonstrated last year at Bonn University’s Econ Lab. Armin Falk and Nora Szech had invited students at Bonn University to make an unpleasant choice: saving the life of a mouse or receiving money.

“What drives human behavior, induces us to wear yellow rather than green jumpers or to drink beer while wine is also available on the menu?” Armin Falk’s often unorthodox answers to these questions have earned him the position of visiting professor at Harvard as well as a frequent visitor to the Toulouse School of Economics. Recently, I’ve had the pleasure of sitting with him at Bar Basque in St. Pierre. His latest project focuses on a long-neglected aspect of decision theory as well as a branch of the science of philosophy: morality.

The mice experiment was famously published in Science and drew attention from a lot of major newspapers. Why is it important? The key hypothesis behind the experiment is that “markets display a tendency to erode moral standards.” In order to verify this hypothesis, one needs to find consensus on two things: the definition of a market and the idea of morality. Arguably there”exists a basic consensus that harming others in an unjustified and intentional way is considered as immoral.” An example of killing animals fits this notion well. In the experiment Armin Falk and Nora Szech imitated different institutional setups and randomly assigned participants to three different trials.

In the first setup individual faced a choice between 10 € and sparing a mouse’s life. In the second (bilateral) and third (multilateral) treatment participants were divided into groups of one buyer and one seller and seven buyers and nine sellers respectively. The life of the mouse was “entrusted” to the care of the sellers. Consequently sellers and buyers within their respective groups bargained over a mouse’s life for a total gain of 20 euros that the parties could split up between themselves.

If trade was to occur on a morally neutral good, say a university gift voucher, one would not expect the results to differ in different institutional frameworks. And in fact, they do not! Yet this is exactly what happened upon trading the life of a mouse. The percentage of traders willing to kill the mouse for an amount equal to or less than ten Euros rose from just below 50 % in the individual choice treatment to exceed more than 75 % in both bilateral and multilateral trading. What happened here? Armin Falk believes that in part the finding can be reconciled in diffusion of being pivotal. Diffusion of being pivotal implies that “actors may perceive themselves as irresponsible for the outcome”, or, bluntly said, “if I don’t buy, someone else will. If I don’t kill the mouse, someone else will.” When thinking about institutions, we should perhaps favour those which do not allow actors to run away from their responsibility. In addition Armin Falk argues that in markets it takes two people who agree on trading to complete a trade, implying that responsibility and feelings of guilt may be shared and thus diminished. Moreover, market interaction reveals social information about prevailing norms. Observing others trading and ignoring moral standards may make the pursuit of self interest ethically permissible, leading further individuals to engage in trade.  On the other hand, the experiment hinges on the foundations of decision theory. A fraction of participants within the bilateral trading setup believed that the gassing of the seven mice would have happened in any case. At the same time, Armin Falk and Nora Szech verified a share of 10-15 % of participants who rejected payments well exceeding 100 Euros in order to save a mouse. These participants were not known for being animal rights activists, who would be willing to spend a large share of their income just to save a doomed laboratory mouse. If their true valuation for the life of a mouse indeed exceeded 100 Euros, transitivity would have induced them to just do that – a paradox.

Indeed Economics does not know “reasonable and mathematically tractable concept of morals”. Until now, the discipline is deeply rooted within a utilitarian paradigm. If a significant fraction of people don’t follow its prescriptions, we might ask ourselves if we are not missing out on an important part of human behavior. The killing of mice might be a stylized setup – but what if non-utilitarian, say Kantian moral values also prevail in a more realistic context. What if moral values bias parent schooling decisions, green energy investments or property rights? Then axiomatically defined rationality falls short of explaining human behavior on goods, which are perceived as morally loaded.

  1. Does an economic notion of morals exist?

In my view there does exist a notion of prosocial behavior we sometimes conceive as a moral notion. This includes cooperation, altruism or the supply of positive externalities. But we do not have a reasonable and mathematically tractable concept of morals in economics, which is the reason why we [Jean Tirole and Armin Falk] are currently working on introducing such a concept into Economics.

  1. Will such a decision theory rely on a utility function?

Without doubt economic thought is deeply rooted within the utilitarian concept. But clearly there exist other ideas, take Kantian principles for instance. Indeed we would like to consider decision criteria that are in some dimension orthogonal to the cost-benefit analysis. Our [Armin Falk and Nora Szech] results do emphasize that the majority of people follow utilitarian principles. Yet we can identify a fraction of participants whose decisions in experiments contradict utilitarian criteria. In the mouse experiment a fraction (10- 15 %) of participants had a firm belief that almost certainly all mice would be killed, but still refused to trade the life of a mouse for an amount exceeding 100 Euros themselves. One could perhaps argue that these participants acted on Kantian grounds, perhaps in just following some moral heuristic.

  1. Could you match moral beliefs to other personal determinants?

Based on my experiments with Nora we consistently find that the likelihood to trade a mouse’s life against 10 Euros is lower for female, more intelligent and left-wing participants than it is for their respective counterparts.

  1. Can we really infer moral beliefs from chosen actions?

It is in fact not straightforward to infer values and beliefs from actions, in particular when it comes to moral decision making. What I find fascinating in this respect is our capability of reinterpretation of beliefs and actions. For example, if my environment allows me to construct a narrative, that will make my action appear in a favorable light, such stories will be told and the demand for them will be high.

  1. Can Economics construct such a narrative?

Absolutely. I believe that we often impose normative system without reflecting upon them. We are used to interpreting the world along utilitarian concepts, or Pareto efficiency but often claim to be a non-normative science. Our concepts are productive, and can be justified, of course. But a critical reflection about what they imply in terms of our normative conclusions is important.