osewalrus (osewalrus) wrote,

More implications from D Game Theory

Actually, as I've been noodling with this, a number of things become clearer.

1. Why is it that as Ds "play to the center," they end up losing the center over time? Yet those who care most passionately about the issues and are most well informed appear to fall into the same trap, as evidenced by civil rights orgs, unions, and now progressive orgs such as MoveOn?

Ans: The model demonstrates that, over time, results for all leg/reg fights will converge on Republican outcomes. This also explains why the most progressive legislation and regulatory initiatives occurred earliest in the Presidency, but D advantage gradually wore away. As a result, the marginal difference between Democrats and Republicans becomes increasingly small.

This is problematic for Ds because their chief campaign plank tends to be "we're not Republicans, and Republicans are worse." If we assume that supporting political candidates/parties costs effort (and, potentially, money), the benefit to a voter to support a party must be greater than or equal to the cost of support. As D outcomes converge to R outcomes, the value of "not being Republican" diminishes.

For those most passionate about the issues, the value of being different from Rs, even marginally different, is very high. Thus, dedicated civil rights orgs, reproductive rights orgs, and now economic progressives (as witness the MoveOn "get out the vote" campaign despite frustration with Obama) fall into the trap of supporting Ds no matter what they do, because for them the "not R outcome," even if it is only marginally different, has significant value. By contrast, and counter-intuitively, those who care less about the issues (the "center"), find voting for Democrats less valuable and therefore -- rationally -- stop. This is why people who hate the DLC and "centrist" Democrats still vote for them, but people who merely "lean" Democrat gradually wander off. Rationally, for those who care less, the marginal difference between "R Outcome" and "Almost R Outcome" is not worth the cost of participation.

2. Why don't Ds perceive the actual nature of the game?

This is what is so surprising. Ds have been playing the same game and using the same strategy repeatedly since The DLC takeover of Congress and the election of Bill Clinton in 1993. They have essentially remained unshakably committed to the same losing strategy ever since -- treat each round as a unique, independent game with no relation to any other game and where any outcome greater than zero is considered a "win." Yet these are long term players, both on the decisionmaker side and on the electorate/participant side. Why do core Democrats, as well as Democratic decisionmakers, consistently talk themselves into the same "well, something is better than nothing, don't let the best be the enemy of the god, etc., etc." argument every time, despite constant reiterations of the same game and the same outcomes? We will note that the only times Ds achieved noteworthy electoral success with this strategy was in 2006 and 2008 when, for various reasons, the value of "not being Republican" was extremely high.

I have developed three possible theories. I suspect a combination of them is correct.

a. The Corruption theory: Ds are corrupt/decisions are made endogenous to the game. This theory was advanced by fatlefty in the comments to the previous post and has the advantage of simplicity. If we assume that you are intelligent enough to get elected, you are actually trained in game theory (which Tim Geithner and Larry Summers certainly are), and you keep failing to recognize the true nature of the game, the most logical answer is that your are subject to other factors that require you to make losing decisions. Here, we assume that Summers, Geithner, and others are constrained in their ability to make decisions for the overall game because they need to produce certain outcomes for their Wall St. masters. The only way to maintain the necessary boundaries for their decisions is to require that the game be treated as a one-time game so that these less desirable outcomes can be portrayed as "victories."

The problem I have with this theory is that it is not limited to a handful of decisionmakers. I can list half-a-dozen people on my f'list who make the same "reasonable" assessment of every political contest (health care, tax cuts, DADT repeal) as if it were a one-time game, and can list another half-dozen who will vote for centrist Ds they hate despite hating them entirely because they assign such a high value to "not R" that even an extremely marginal difference between R and "not R" is considered worth the investment.

So what other theories are there? I will invoke two concepts of the great Akerlof -- information assymetry and cognitive dissonance -- for the next two theories. These are interlinked.

b. The "Factions" Theory: As Will Rogers once remarked, "I do not belong to any organized political party, I am a Democrat." Democrats, more so than Republicans, start with a base that is much more invested in "not R" and has much less agreement about long-term outcomes. Large factions of the Democratic Party may prove indifferent or even hostile to the interests of other factions. These factions regard the political fights on their core issue as unique and place a high value on achieving a non-zero outcome. Further, because these factions are indifferent to the party over all, they are less likely to perceive the game as an iterative game.

Even if the faction perceives its own issue struggle as a repeated game, the faction fails to perceive the linkage with the other rounds overall and how they impact their specific sub-game. Finally, even if the faction recognizes these factors, the weak identification of the faction with other factions makes the faction interested in any specific round unwilling to sacrifice its round for the benefit of the game as a whole, since it is unclear that this would translate into an advantage for the specific faction in future rounds.

In other words, the "single issue voter" who is a Democrat because s/he values "a woman's right to choose" but is indifferent on other issues may regard the game as around reproductive rights as both unique and the only game, wholly unrelated to the other games -- such as economic equity for example -- because s/he lacks the necessary information and engagement to perceive the pattern. For this voter, any result in the one game in which the voter is interested is better than a zero result, and the voter does not perceive the consequences to his/her single issue game of failing to engage in other issue games. Nor is this voter going to be persuaded to sacrifice his/her round to "punish" Rs for the benefit of future rounds that may involve other issues. It is impossible for the "reproductive rights" voter to gauge accurately what advantage accrues in future rounds of the "reproductive rights" subgame, and the reproductive rights voter is sufficiently indifferent to other factions that the trade off of a catastrophic outcome in a reproductive rights round for advantage in an economic equity round in the future (especially where the linkage is not explicit) is not worth it, and the faction will urge his/her representative to accept the merely "bad outcome" rather than the "catastrophic outcome" for the specific round -- thus failing to take the result that would "punish" Rs, which may be the most effective result for the iterative Prisoner's Dilemma game overall.

c. The Cognitive Dissonance Theory: Cognitive dissonance occurs when one is forced into a situation that requires acknowledgment of facts that are painful -- for example, because they violate core ideas of our self-conception or because we are helpless to change them. Akerlof demonstrated this in his study of workers in dangerous jobs who, after the creation of OSHA, refused to obey safety regulations. By contrast, new workers (unless socialized to the contrary) would obey the procedures. Akerlof concluded that workers confronted with unsafe job conditions had undergone cognitive dissonance to manage their stress. i.e., they persuaded themselves that job conditions were safer than they were, or that accidents would not happen to them individually because of some other reason ("I'm careful," "I'm lucky," "accidents only happen if you aren't paying attention"). When safety equipment was made available, using the equipment would require acknowledgment that the work environment was less safe than they had persuaded themselves it was. Rather than admit that, they declined to use the equipment.

Recognizing policy fights as an iterative version of the Prisoner's Dilemma (as suggested by fatlefty) requires strategies that are extremely painful for many Democrats for a number of reasons. First, many Democrats/Liberals are motivated by sympathy for others. Consider the primary justification for health care reform for Liberals/Progressives was that it would benefit other people -- notably those with no access to healthcare. Yes, these same people/groups often made arguments based on self-benefit and economic gain. But these arguments tended to be secondary and primarily framed in rebuttal. By contrast, Republican/conservative arguments were much more focused on the cost to "me/us" to benefit "them." (I am not making a moral judgment. Self-interest is equally alive and well in Ds/Liberals/Progressives and one finds altruistic Rs/conservatives. But the self-conception for political purposes frames virtue on the Liberal side as concern for others, and on the conservative side as preservation of self from threat/attack.)

A willingness to accept nothing as an outcome in the healthcare game would require deliberately sacrificing those who benefit from even the sub-optimal response. Recognizing such an outcome is necessary for any long-term change that would produce much more optimal results is extremely painful. Similarly, recognizing that accepting the sub-optimal outcome in the healthcare round would of necessity require accepting sub-optimal outcomes in subsequent rounds on other issues (e.g., financial reform, tax cuts) because it is an iterative game is likewise painful. It is much easier for Liberals/Progressives to conceive that the healthcare fight was a unique, non-iterative game with no consequences outside that game than to accept either side of the trade off by accepting that policy is an iterative game.

Similarly, treating the process as an iterative repeating game requires Dems/Libs/Progressives to accept strategies they find morally repugnant. Consider the following:

"We don't 'punish' Republicans because we're better than that."

"This shouldn't be about politics, it should be about the merits/facts/best policy/best interests of the American people -- not about winning."

All of these ideas about self reverberate soundly in every Obama speech scolding the "far left" or whatever other euphemism we use to separate "people who should like and agree with me but who now no longer like and agree with me."

By contrast, the concept of this as an iterative game where sacrificing a more optimal outcome in any specific round resonates with core conceptualizations of conservatives about themselves. Conservatives believe in sacrifice. Indeed, they like to conceive of themselves as tough warriors, willing to do what is necessary to win in the long run, and utterly uncompromising in the face of an enemy that threatens their core values and way of life. Any set back is temporary. The only way to lose is to compromise, because an essential part of the self-conceptualization if the utter rightness of cause. As a consequence, we have seen the exodus of moderate Rs and of R/conservative policy wonks who place high enough value on specific outcomes that they are willing to engage in cooperative strategies that provide benefit to Ds, or who are unwilling to accept sub-optimal outcomes in high value rounds (Frum after the healthcare debate, for example).

Conclusion: Ds/Liberals are stuck in a losing strategy that must end badly over time, punctuated occasionally by brief periods of success when the marginal difference of "not R" is highly valued by the general electorate. The only solution is for Ds to recognize that it is an iterative game rather than a series of unrelated unique games, and alter their strategies accordingly. Otherwise, rational players should simply quit -- unless they attach such a high value to "not R" that they consider it worth supporting Ds as the "not R" choice even when the actual difference in outcomes is extremely marginal.

  • Post a new comment


    Anonymous comments are disabled in this journal

    default userpic

    Your IP address will be recorded