Climate Negotiation Experiments at Cologne

Experiments on negotiating strategies were started in early 2013, at the Cologne Laboratory of Economic Research.* (return to overview)

The experiments compare negotiations over quantity targets (often called caps) with negotiations over carbon prices.  Experiment 1 also compares common commitments (commitments to a single global price or quantity) with individual (differing) country commitments. In each treatment, there are 10 groups of three students who represent three countries. Each group plays the same negotiating game 10 times in a row, and the reported abatements are those of the final round.

Price is better than quantity. The central question addressed by these experiments is whether a price or a quantity commitment would produce the stronger treaty. In practice, price commitments have been shunned, but in these experiments they perform significantly better than caps. Comparing treatments 2 and 3, we see that price performs almost perfectly, while quantity does less than half as well. This is due to the riskiness of caps. But caps also prevent a common commitment, as Kyoto proved and as all now agree. Comparing treatments 3 and 8 shows that caps fail just as badly, and likely worse, without any risk, simply because they prevent a common commitment.

Treatment 1. This is basically individual quantity pledge and review—the plan for Paris and the approach used in Kyoto. In these experiments there are 9 rounds of pledge and review before the final round, which is reported. In T1, there is only a tiny indication of altruism (“ambition”).

Treatment 2. The same as T1, but the common commitment is enforced. It is determined by the lowest pledge made by any one of the three players. Because it’s “legally binding” players know that a high recommendation will help them as long as it is not higher than optimal. So the optimal outcome (15) is a Nash equilibrium.

However the optimum is calculated on the basis of the “expected” outcome, and players know there will be a “shock” such as Fukushima, Canadian tar sands or a recession, which will make compliance unexpected cheap or expensive. If they ignore this risk they should pledge to abate 15. But quantity pledges are risky, so one of the players always chose a weak pledge.

Treatment 3. This is the same as T2 except that the pledge and commitment are to a price. Price is far less risky. If your emissions go up 1Gt you only have to tax them and keep the revenues. With a cap you must buy 1Gt of abatement, usually from a foreign country, and they keep the much-greater revenues. Because of the greately reduced risk with price pledges, the pledges were almost optimally ambitious.

Treatment 8 and The Price Result. Comparing T3 and T2, we see that price beats quantity due to reduced risk under a price commitment. But that is when the quantity approach uses a common commitment. In fact Kyoto tried to find such a commitment for more than a year and failed. All now agree that a common quantity agreement is impossible.

Consequently, the relevant comparison is between T3 and T8. Treatment 8 is like T2, but with no shock and no common commitment. So there is no risk, and this time the failure of quantity pledges is due entirely to the lack of a common commitment.

Treatment 4. The countries are asymmetric, they pledge prices, and the lowest pledge is enforced. The outcome appears to be better than the Nash equilibrium, but it is far from optimal because it is held back by the country that benefits the least from a climate policy (hence asymmetry is the problem).

Treatment 5. A Green Fund is provided to help overcome the asymmetry of T4. The country that benefits the most pays the one that benefits least. That payment is just enough to make the Nash equilibrium the optimal outcome. Compared to T4, this improves the outcome dramatically.

Treatment 6. This is the same as T4 but with Q (quantity) pledges instead of P (price) pledges. There is a statistically weak indication that Q pledges to better. But note that this assumes a common quantity commitment, which has been acknowledged after Kyoto to be impossible.

Treatment 7. This adds a Green Fund to Treatment 6 to overcome the reticence of the low-benefit country. This seems to help. Again the assumption of a quantity commitment is of theoretical but not practical interest. And even in this case, the outcome is much further from optimal than what can be achieved with a price commitment (see T5).

Why a minimum agreement works. It may surprise some that an agreement to implement the lowest price pledge favored by any nation is stronger than an agreement that allows each to implement its favored price. But this makes perfect sense because a commitment to the minimum pledge encourages everyone to express their true opinion. This is because they know they will not be punished for it (as happens without a commitment). If their pledge requires them to abate more, it is because it raises the minimum and then all must abate more—so they are rewarded by more effort from others. With out a common commitment, a higher pledge just means they will have to do more and obligates no one else.

A common commitment. What these experiments indicate is that a common commitment is necessary to encourage countries to express their true ambitions for the world as a whole, and that price makes such a common commitment easier to implement. Hence a price commitment gives a stronger outcome.

*  The experiments are under the direction of Axel Ockenfels. Andreas Pollak is designing and implementing them, and Peter Cramton and Steven Stoft are also assisting.