The Prisoners' Dilemma Meets File Sharing

Hi,

If you’ve been reading the cmuinliberia2010 blog before now you may have noticed that my posts have been mostly travel stories on my search to understand Liberian culture through food. That particular series of writings on culture through food is ongoing, but what follows in this post is part of a far-different series of writings.

In this post I begin another series of writings. I intend to explain international development’s challenges through game-theory and also to explain game-theory through international development’s challenges. Don’t worry too much about knowing anything about either of these concepts now; much of what I will be writing will be explanatory in nature.

We begin with one of the most famous of game-theory concepts, The Prisoners’ Dilemma.

In this situation two people are caught by the (abstract) authorities for a (unspecified) crime. Each prisoner is taken to a separate room. Each prisoner is told that if they confess to the crime, and help the authorities convict the other prisoner, that they themselves will be let go with no jail time at all. Bluntly, if you snitch you walk out - a free person. Seems pretty simple, right? Ah, but there are several catches.

If neither of you snitch then both of you receive 2 years jail time, if you do not snitch but the other prisoner does snitch then they walk free and you receive 10 years of jail time. If you both snitch then you will both receive 5 years of jail time.

(also, snitches tend to get stabbed, but we’ll leave that part out)

To use the convention of game-theory, here is a decision matrix. In the matrix we have two players. Each player makes a choice. The result of those two choices is reflected in the boxes. Example: If Prisoner A snitches and Prisoner B does not, the outcome is the circled portion of the matrix. Because A snitched and B did not, A walks free and B is sentenced to 10 years in jail.



SO, we see that each prisoner has a choice to make. To snitch or not to snitch, that is the question. The trap is beautiful in its simplicity. This is a trust trap. The only way to beat the trap is to trust the other prisoner and have the other prisoner trust you.

Inevitably you find yourself in a room by yourself trying to decide, snitch or don’t snitch? It turns out you can do better by snitching in either case. Snitching when the other doesn’t, gives you 0 yrs of jail time versus 2 yrs, therefore snitching when the other doesn’t is better for you. Snitching when the other snitches, is ALSO better because it gives you only 5 yrs instead of 10 yrs. The rational choice then is to snitch, even though that leads to a sub-optimal solution.

Two very rational people under such a trap will end up with 5 years of jail time each, even though they could have ended up with only 2 years each if they had trusted each other and not snitched. In this way the Prisoners’ Dilemma shows us how smart people make dumb choices, but those dumb choices are completely rational.

To ‘solve’ The Prisoners’ Dilemma is something that an occasional philosopher will attempt. However, to ‘solve’ The Prisoners’ Dilemma would be a lot like trying to solve a measuring cup. The measuring cup only helps you measure other things, it holds no wisdom on its own. So too The Prisoners’ Dilemma is so vague and lacking in context that there is nothing to solve, only an instrument to help us with other problems.

And now we hit my problem, the problem I will be explaining through game theory, a scourge of international development work in Liberia and many other nations, FILE SHARING.

-

Imagine you are the head of a government ministry and you are given a choice. You could spend two days (two is an arbitrary number) to share all of your files with all of the other government Ministries but it is a voluntary process, once the data is uploaded everyone else in the other government ministries can see it. All of the other government ministries are also given the same choice. To share files or not to share files, that is the question. The cost is of course in the time you spend to collect and share your information. The optimal solution would be for everyone to spend the two days to share files, then for the cost of only two days of work you would have access to every work document you needed. Productivity rises, sunshine pours down on a parade of smiling government workers all rolling up their sleeves, and free flowing access to information lets each individual ministry meet its mandate quickly and efficiently. In theory…

But, you still have a choice to make, to share or not to share. Worse yet, it’s a trap, a trust trap.If you do not share files, but everyone else does, then you have access to all the files (because you automatically have access to your own files) and you will have spent 0 days. And, if you share your files and no one else does then you have spent 2 full days and received nothing in return for it.

Or, to put it into a game-theory decision matrix:



You can expect the other ministries to reach the same conclusion you do, or at least make their choice in a rational way.

The rational choice is for no ministry to share files. Each ministry would be rational in choosing not to share. That is at the macro-level. However, at the micro-level, for the individual working inside the ministry, something else is happening. That something else is the rise of the ‘Bureau Squirrels’.

For the individual working inside a government ministry there is something more at stake than just files. The coming re-organization of the government will threaten all non-necessary jobs. The government of Liberia is going to re-evaluate the ministries at some unknown point in the future, but it’s believed to be coming soon. If you are the only person with your information, then you can say,

I am needed,

My job is needed,

I am the only person with this information.

So as an individual within a ministry, you too would face a choice. When your boss asks you for all of your files to share in the great two-day file sharing you would also face the choice, to share or not to share. If everyone shared their files, then the playing field in the coming re-org would be even and merit based and you would be more productive in your current job until whenever the re-org happens. But it’s a trap, a trust trap. You could squirrel your work files away and wait for the re-org to happen, thereby gaining leverage. And if you share your files and no one else does then you would be at a disadvantage when the re-org happens. It is the rational choice to not share your files.

The individual would be rational by not file sharing. The ministry would be rational by not file sharing.

There are many transitional governments in the world and I believe this game is being played in many of them. The leadership is calling for transparency and change but it would only be rational for each individual to share information if everyone shared at the same time. Progress is therefore delayed, opaque governance that no one really wants continues, and no one seems to understand why. It seems so obvious; if everyone shared their information then everyone would be better off, right? So what’s the hold-up?