Politimetrics

So The Lindsay and i were discussing the Supreme Court on Friday night, and we analyzed a little of the game theory behind Bush's strategy in choosing O'Connor's replacement. (Given the developments over the last 24 hours, i wish i had posted this on Saturday; i would have looked like a prophet.) Consider:

Bush is even more of a lame duck than most second-term presidents, if Cheney indeed decides not to run. (If any Cabinet members or senior staffers run, they may be able to distance themselves from the President if it is politically expedient - even going so far as resigning from their post if necessary, which is decidedly not an option for the VP as a political consideration) As such, we need to consider what action is most closely aligned with his own self-interest in terms of ensuring his legacy. Why play it safe in order to salvage an executive administration that will end in 2009 when you can squander what remaining political capital you have to put your favorite pick on the Supreme Court for a couple of decades? Given that, I'm not the least bit surprised that Bush picked a crony. I am a little surprised that Reid appears to like her, but then pissing off the Democrats was not per se an element of Bush's game-theoretic strategy..

Let's look at Bush's expenditure of political capital (whatever that is) from an economic perspective. First, we assume a use-it-or-lose-it rule, because Presidents usually don't follow up their stint in the White House with a political encore (the experience of Mr. Chief Justice Taft notwithstanding). Therefore, Bush has to use up his political capital somehow. He can invest it in sustaining the administration's reputation and thereby helping his successor to win in 2008 (but see the analysis above regarding why in the present case he has nothing to gain from this). Or he can squander it on a highly unorthodox Supreme Court pick (or something else). Let's compare these two options on their own merit..

To begin, we will measure political power in what i will term "justice-years". We say that in any given year a Supreme Court justice wields one justice-year of power. Thus the entire Supreme Court wields nine justice-years of power in any given year. Assuming the three branches of government have equal power, the executive branch wields nine justice-years of power per year as well. (For this analysis it will not be necessary for me to allocate these nine justice-years between the President and Vice President.) It could also be noted that any given Senator in one year wields .045 justice-years and a Representative just over .01 justice-years..

Ignoring as a first order of approximation the fact that a 100% stake in an enterprise is worth more than nine times a 1/9 stake (this approximation might not be too inaccurate in our system, seeing how even a controlling interest in the Executive Branch does not imply total control of the government), we now have an adequate yardstick for measuring the net effects of an expenditure of political capital. We have already analyzed Bush's incentives, so let's look at the problem in the abstract. For a randomly-chosen President faced with a choice between investing her political capital in her successor and spending it on pushing through a Supreme Court nominee, which should she choose in order to maximize her legacy? Let's measure a legacy in terms of raw justice-years (no adjustment for present value of future justice-years), and for now consider only the tradeoff between an Executive successor and a Supreme Court nominee (neglecting the effect of Congressional legacy or the possibility that the President won't use up all of her political capital)..

i'll define a President's Executive legacy to include any unbroken line of essentially handpicked successors (eg. President's VP gets elected, her VP gets elected, and so on). This is probably not the best way of measuring a legacy, but let's work with it. Just keep in mind our expected value will probably be high based on this methodology. Let's assume that on average, the President's successor will be elected with 60% probability (this reflects an incumbency advantage attributable to the President's expenditure of her political capital to ensure the election). An investment of political capital in this endeavor yields a four-year term (36 justice-years) with probability 60%, or an expected value of 21.6 justice-years. Assuming a similar strategy is applied four years hence, there is a 36% probability of an additional 36 justice-years, or an expected value of 12.96 justice-years. Carrying this process on indefinitely yields a total expected value of 54 justice-years for this investment..

What if, on the other hand, the President were to squander her political capital on a 90% chance to push through a Supreme Court nominee that reduced to 30% the chance that her successor would retain the Executive branch? Let's say that nominee is likely to spend 30 years on the court (hey, if you're gonna go all out, might as well make it someone who'll be around for a while). This gives us an expected value of 27 justice-years (neglecting, of course, any feedback legacy effects derived from such fortuitous anomalies as your nominee casting the deciding vote that puts your OWN SON in the White House ten years later--oops, sorry, this was supposed to be ABSTRACT). Add to this an expected value of 27 justice-years of Executive legacy (exactly half of what it was in the previous example, as we assume that after the initial 30% drop successive administrations will have the usual incumbency advantage), and you can see that the numbers i have proposed have us poised at a point of indifference. Any tweaking of the variables and the balance would be tipped in favor of one approach or the other..

Of course, there are a lot of loose ends to tie together here. For instance, what is political capital exactly and how is it measured? To what extent can a President allocate her political capital toward various outcomes and to what extent do these expenditures increase the probability of each outcome? We might be able to provide this technique with some empirical teeth by substituting approval rating as a proxy for political capital and using some hard data regarding average Supreme Court tenure and confirmation rates. But at least this is a hopeful start toward developing a theory of what i might call Politimetrics..