D.D. 10/21/05

Game theory, something that has been in the news recently because two of its academics were the recipients of the Nobel Prize in Economics has some interesting applications in the social sciences. I've been reading several books on the subject recently and I'm particularly interested in its applications in business, governance, and political campaigns. Most of the widely used applications to political science have been in international relations, particularly in studies of the Cuban Missile Crisis and other Cold War era conflicts. However, I believe, that game theory can offer significant advantages in developing strategies for the operations of political campaigns and that's what I'll explore in this post. If you're interested, it wouldn't hurt to read this introduction, which deals with game theory in business.

I offer here a simplistic example. The reason that we are using game theory is because the ramifications of each campaign's decision depend on the decisions of the other campaign. The layman's way of displaying game scenarios is through a bimatrix, such as the one below.

*	****	Ed Bryant
*	****	go negative	quiet
B o b C o r k e r	go negative	(-2, -2)	(-1, -2)
	quiet	(-2, -1)	(0, -1)

How did we get these numbers you might be asking? Let's take a look- consider that going negative in a campaign has some negative effect on your opponent- that much is obvious. But going negative can also negative affect your own campaign- because some people might perceive you as being spiteful or just mean. However, we can assume that any negative effects you would get by "going negative," would be some degree less than the negative you have inflicted upon your opponent. I use -1 for the effect on your own campaign and -2 for the effect on your opponent's campaign. So for instance, if through one run of this game, Ed Bryant decides to "go negative" and Bob Corker decides to stay quiet. Corker's "score" would be -2 and Bryant's score would be -1 (see the lower left-hand quadrant- Corker's score followed by Bryant's score). This would be a completely equal planned game except for the fact that Bryant has significantly fewer resources for getting message to voters through traditional media (i.e., he has less money). Therefore, Bryant is actually at a slight disadvantage in not going negative, since the longer he stays quiet, the less time he has to deliver his message, whereas Corker already has the money to pay for TV ad buys, etc. If this seems like a rationalization and just making a simple problem complex, just put that fear aside for a moment.

Let's look at the chart and consider next what the dominant strategy would be- that is, based on the available options what both campaigns, being rational, and are to do. Look down each column and consider Bryant's position- if he stays's quiet; the possible loss based on what Corker does is -3. (-2 if Corker goes negative plus -1 if Corker stay's quiet). Now consider going negative, it's the same result of -3. Since both of the options produce the same potential result for Bryant's own position, he will likely be interested in determining Corker's potential positions. So, looking at the rows, we see that the best Bryant can hope for not going negative is for Corker to go negative and thus be diminished to -1, but this still leaves Bryant at -2, so he might as well risk going negative, because -2 is the worst possible result- and by going negative, he will ensure that Corker is also at -2. If Corker stays quiet, Bryant will have gained position because his score will be -1 and Corker's will be -2. Thus for Ed, the decision will be to go negative.

Now consider Corker. The sum of possible loss of going negative is -3, but the sum of possible losses of staying quiet is -2. Thus, the choice is obvious- Corker's analysis doesn't even have to look at what Ed will do; he will stay quiet. -2 is greater than -3 after all. What that means is that the lower left quadrant is called the dominant strategy. That is the strategy that if you repeated this exercise every day during the course of the campaign would be picked the most- because it is the most rational for both players. However, it is not the only strategy. If this exercise were carried out through 100 iterations of the game, and the dominant strategy was picked each time, Corker would lose, because he is continually being hit with -2's, while Bryant's decrease is the marginal -1. That's why Corker will eventually have to mix it up and go negative against Bryant, depending on the actual value of the loss (in polling numbers) in each iteration (or day). But Corker is in the dominant position because he can change his strategy, whereas Bryant is more or less locked into going negative.

Of course this is a very cursory and basic analysis- and when you add Van Hilleary to the race it becomes complex fast, especially considering that a 3rd person's numbers may be affected even in games they aren't involved in. For instances, consider the added exposure that "going negative" gives you. So when Ed Bryant decides to "go negative" against Bob Corker, he is contributing an added advantage to his own score by reducing Hilleary's since he is the one in the limelight. I'm still studying this, but it is quite fascinating.