We are pleased to welcome “Burt Monroe”:http://polisci.la.psu.edu/facultybios/monroe.html of Penn State University with the following guest post on the “proposed electoral college reform”:http://www.nytimes.com/2011/09/19/us/politics/pennsylvania-republicans-weigh-electoral-vote-changes.html in Pennsylvania.
In Pennsylvania, our legislature is considering a bill to distribute our Electoral College votes according to results in Congressional districts, as Maine and Nebraska do, rather than in a statewide bloc.
For the Republican legislature and governor, this has a partisan logic, shifting a risky 20-vote bloc (won by five consecutive Democratic candidates), to a more certain Republican-leaning split of 11-9 or so.
The statisan (word) logic is different. From the point of view of state interests, the conventional wisdom is that doing this unilaterally is stupid. (Thomas Jefferson wrote my late cousin Jim to say “folly and worse than folly,” but he wasn’t the wordsmith I am.) Commentators assert Pennsylvania will cease to be a battleground, its influence reduced to that of New Mexico, Delaware, North Dakota, or — as with anything that is either smaller or bigger than expectations — Rhode Island. Well, which is it? How much does this decrease the value of Pennsylvania’s electoral votes?
Think of the campaigns as using their resources — advertising dollars, campaign visits, steel tariffs, cabinet posts — to try to buy electoral votes. They’re engaged in 50+ simultaneous auctions (where the loser pays, too) and the prices vary wildly. The value of winning is larger the more votes are at stake, but the price goes up with how difficult it would be to shift the winner from one side to the other. So, bids get run up in close states with big prizes (that’s what makes Ohio, Florida, and Pennsylvania “battlegrounds”), but bids are lower where the prize is smaller (e.g., New Hampshire or Maine) and as low as zero where states are so uncompetitive that one side can’t pay that high a price (e.g., New York or Texas).
Theoretically, this leads us to a model something like the following: the total bid on a given state is a Gaussian function — a bell-shaped curve — of expected two-party vote fractions in the state, peaking at 50-50, with the height of the curve increasing with electoral vote. This approximates the logic given by Strömberg (2008).
I estimated the parameters of this model using observed campaign advertising as a proxy for the total resource bids of the campaigns and actual vote fractions as a proxy for expected fractions.^*^ The results are summarized in this figure:
In 2004, Pennsylvania received $38m in post-convention advertising, about 15% of the total. We have about 4% of the US population, so we punched almost 4 times our weight. Under the alternative plan, with 2004 Congressional districts, the model estimate for any one district (those red dots) is small, but they add up to $21m, a net loss of $17m. It’s hard to say how this will translate into 2012 campaign dollars (estimates range from 50-200% more, so maybe $25m-$70m), but this is in any case only a proxy for the value conveyed by overall “attention” which also includes visits, ground game, policy promises, and so on.
The bottom line is that this suggests Pennsylvania would be expected to receive about 8% of overall attention under the alternative plan, a reduction of about half. So folly, yes, but perhaps no worse than just regular folly. Not the 0% of Rhode Island or North Dakota, or even the 3% of New Mexico, but closer to the 8% of Wisconsin.
The reason for this counterintuitive result is the risky, but so far successful, Republican districting strategy of 2000, which made at least eight of Pennsylvania’s 19 Congressional districts highly competitive. This is not expected to change much when new districts are drawn. The model still sees about half of the electoral votes as still in play, mostly in suburban Pittsburgh to Erie in the west, and suburban Philadelphia to Allentown in the east, as well as the statewide two. Add in the role these districts will play as battlegrounds for control of Congress, and Pennsylvania will still get plenty of attention.
*******
^*^(The most recent data available to me at the moment are post-convention spending from 2004. They are from Shaw 2006, sourced to the Bush-Cheney campaign, and clearly have at least minor mistakes. For example, I’m almost certain Mississippi still exists. Caveat emptor.)
