I'm going to use the familiar Prisoner's deilemma game. For those of you who don't know about it, the story goes like this:
Two prisoners accused of the same crime find themselves in separate cells, unable to communicate. Their jailers try to persuade them to implicate one another. If neither goes along with the guards, they will both receive a sentence of just one year. If one accepts the deal and the other keeps quiet, then the turncoat goes free while the patsy gets ten years. And if they both denounce one another, they both get five years. The best strategy for each of the prisoners would be to rat on each other so that they both get five years in prison. The problem is that if they cooperated with each other, they could have gotten away with just one year.
Now, we apply this game to the situation of the Kyoto protocol. You've got several countries, but for simplicity we restrict our attention to just two, say America and India. If America and India cooperate, the world becomes a better place and everybody is made better off. However if any one of them did not cooperate the other should do likewise for the best gains. So again we face the same dilemma. America and India will choose to not cooperate leading to a less-than-perfect outcome. If it is suspected that any country won't sign the deal to reduce emissions, then it only makes sense for all the other countries to back out.
So does this mean that there is no way out? Is there no possible way to ensure cooperation? As matter of fact, a closer look at the situation shows us that there is a slight problem with the game we chose to describe the situation. The problem is that the game doesn't actually end after the meeting. All the countries will still be around and this will continue to be an issue where past actions play a role.
So a more realistic approach would be to consider an infinitely long game with each stage game corresponding to the one described above. A well known game-theoretic results shows that in such a repeated game, cooperation can be achieved if each player follows the appropriate punishment strategy. If there was an agreement between the countries that everybody would reduce their emisssions and anyone who broke away would be fined heavily, then the Nash equilibrium would be to cooperate!
Two prisoners accused of the same crime find themselves in separate cells, unable to communicate. Their jailers try to persuade them to implicate one another. If neither goes along with the guards, they will both receive a sentence of just one year. If one accepts the deal and the other keeps quiet, then the turncoat goes free while the patsy gets ten years. And if they both denounce one another, they both get five years. The best strategy for each of the prisoners would be to rat on each other so that they both get five years in prison. The problem is that if they cooperated with each other, they could have gotten away with just one year.
Now, we apply this game to the situation of the Kyoto protocol. You've got several countries, but for simplicity we restrict our attention to just two, say America and India. If America and India cooperate, the world becomes a better place and everybody is made better off. However if any one of them did not cooperate the other should do likewise for the best gains. So again we face the same dilemma. America and India will choose to not cooperate leading to a less-than-perfect outcome. If it is suspected that any country won't sign the deal to reduce emissions, then it only makes sense for all the other countries to back out.
So does this mean that there is no way out? Is there no possible way to ensure cooperation? As matter of fact, a closer look at the situation shows us that there is a slight problem with the game we chose to describe the situation. The problem is that the game doesn't actually end after the meeting. All the countries will still be around and this will continue to be an issue where past actions play a role.
So a more realistic approach would be to consider an infinitely long game with each stage game corresponding to the one described above. A well known game-theoretic results shows that in such a repeated game, cooperation can be achieved if each player follows the appropriate punishment strategy. If there was an agreement between the countries that everybody would reduce their emisssions and anyone who broke away would be fined heavily, then the Nash equilibrium would be to cooperate!
2 comments:
I heard a superb seminar this weekend and because I can't do a full post on it, I am relegating it to the comments. It was about jump and volatility modelling in high frequency financial data. The presenter was Jiangqing Fan from Princeton and what was most remarkable about the presentationwas the clarity with which he put the whole thing. I can describe it wihout looking at the paper becasue of the clarity. Volatility as a Weiner diffusion process and jumps as Poisson processes. Remove the jumps, and model the diffusion as you would ordinarily do, essentially wavelet based methods but using a basis which makes better use of the high frequency data. Brilliant. The mechanism does not matter, it needs but a night's reading and you are ready to apply.
However, the problem with making things simple is that people get the idea faster and then quibble about alternative ways of modelling (whichever ones you happen to champion). Another issue raised was the problem with 'removing' jumps is that jumps often are not entirely exogenous shocks and could form an important part of the volatility process and throwing away this information might not be the best thing to do.
Also adjusted for market microstructure, as a minor extension.
Tirthankar
the problem with international agreements is the fact that no one exists to impose punishments
Post a Comment