Gideon's Blog

In direct contravention of my wife's explicit instructions, herewith I inaugurate my first blog. Long may it prosper.

For some reason, I think I have something to say to you. You think you have something to say to me? Email me at: gideonsblogger -at- yahoo -dot- com

Site Meter This page is powered by Blogger. Isn't yours?
Monday, August 23, 2004
So a few months ago, I wrote a piece that described how I understood our political system, what I saw as its structural defects today, and how I would amend the Constitution to address these defects. The piece is here. One of the amendments I proposed was to adopt nationally the Maine-Nebraska system for allocating electoral votes. (Maine and Nebraska allocate their electoral votes as follows: 2 votes go to the winner of the state-wide vote, while 1 vote goes to the winner of each electoral district in the state.) As I mention in the piece, such a reform would make far more sense if combined with an amendment to prohibit gerrymandering (e.g., by having every state adopt the Iowa districting system). But even without such a reform, the Maine-Nebraska system has the virtue of (a) reducing leverage in the system (fraud in one location could steal at most 3 electoral votes - one for the district and two for the state - as against the current potential to steal an entire, closely-contested state) and (b) pushing campaigns toward the political center (candidates could not win a diverse state like Pennsylvania by running up their base in the state; more would depend on winning voters that actually occupied the political center).

What I didn't have a view on at the time was: would such a reform help either political party? But now I have the data to opine on this important question. Here can be found the district-by-district electoral results for the past nine Presidential elections. (I got the link from this article in Slate about the electoral prospects this year in Maine, which may well split its electoral votes for the first time since the district-based allocation system was adopted in 1969.)

The data yield the following results, by year:

Year EV change Change Election Result?
2000 GOP + 16 NO
1996 GOP + 33 NO
1992 GOP + 45 NO
1988 DEM + 49 NO
1984 DEM + 57 NO
1980 DEM + 94 NO
1976 GOP + 27 NO (but almost; new result is 270 to 268!)
1972 DEM + 47 NO
1968 DEM + 11 NO (George Wallace would also have gained electoral votes)

Apart from the most recent election, the pattern is clear. In each case, had the Maine-Nebraska system obtained nationwide, the election would have been closer in terms of electoral vote count.

This makes sense. Even in a landslide election like 1984 or 1972, a big chunk of the country votes for the loser, and to some extent this portion of the country is concentrated in certain Congressional districts. So it makes sense that, if you allocate the vote by district, you'd give somewhat greater representation to the losing side in the election.

But the 2000 election results suggest that this might not be the case with the current electoral map. Indeed, that result suggests that under the current 50-50 red-blue division of the country, switching to a district-based system would slightly favor the Republicans.

Why? The simple reason is that the GOP dominates more states, and more low-electoral-vote states, while the Democrats dominate fewer, but large and diverse states.

The GOP controls more states. In the 2000 election, the GOP took 30 states; the Democrats took 20, plus the District of Columbia. The popular vote split almost 50-50 in the last election. Imagine if each state had similarly split nearly 50-50, with 50% of the districts going Democrat and 50% going Republican (if the number of districts is odd, give the extra district to the state-wide winner). That's what the "maximum splitting" result would be. (The "minimum splitting" result would be exactly the same result as under the existing syste, - i.e., the state-wide winner would win each district in the state). With "maximum splitting" under the Maine-Nebraska system, the GOP would have won 281 electoral votes - 11 more than they actually won in 2000 - which entirely reflects the fact that they won 9 more states than the Democrats did.

The GOP takes 60% of states in a 50-50 year, but even in a losing year they do better than you'd think. In 1996 - a solid Democratic victory year - the GOP took 19 states. In 1988, the Democrats took only 11 (including DC), even though Mike Dukakis got a greater percentage of the popular vote than Bob Dole. If we look at the number of states in each party's "base" - narrowly construed - I count 17 GOP base states (Alaska, Utah, Idaho, Montana, Wyoming, North Dakota, South Dakota, Nebraska, Kansas, Oklahoma, Texas, Alabama, Mississippi, Georgia, South Carolina, Kentucky and Indiana) versus only 10 Democrat base states (Hawaii, California, Illinois, Maryland, District of Columbia, New York, Connecticut, Rhode Island, Massachusetts and Vermont). So in any given election, not just 2000, the GOP should be favored to win more states than the Democrats which, all things being equal, should benefit them in a Maine-Nebraska system of splitting electoral votes.

Relatedly, the GOP controls more small states, and more states with uniform political complexions. Of the 8 states (including DC) with 3 electoral votes, 5 (Alaska, Montana, North Dakota, South Dakota and Wyoming) are solid GOP states. The Democrats control only 3 such states. None of these states' votes would be changed by moving to the Maine-Nebraska system, since they have only one district each. And a number of other key GOP states - Mississippi, Utah, Oklahoma, Nebraska (which splits its votes currently, not that you'd notice) - do not have much political variety internally from district to district. So for many core GOP states, there would be no change resultant from the switch to a Maine-Nebraska system.

By contrast, the Democrats' strength comes from clear dominance of three mega-states: California, New York and Illinois. But all three states show real internal diversity, and if their vote was split by Congressional district, the GOP would pick up a nice minority of votes (32 electoral votes between the three states in 2000, as many votes as New York has in the current scheme all by itself). Yes, the GOP controls one mega-state - Texas - which in 2000 would have given 10 electoral votes to the Democrats under the Maine-Nebraska system. And the GOP would have lost electoral votes in swing states that they won in 2000 (Ohio, Florida, Missouri). But these would have been offset by Democrat losses in states that they won (Michigan, Pennsylvania, Minnesota, Oregon). All-told, because there are fewer states that are wall-to-wall Democrat (e.g., Massachusetts, which would have lost no electoral votes under Maine-Nebraska) than there are states that are wall-to-wall Republican (e.g. Alabama, which actually would have lost 1 electoral vote under Maine-Nebraska, but close enough), the GOP would have a structural advantage today if the country shifted to a Maine-Nebraska system.

Of course, if the country did shift to such a system, campaigning would change and this could erase part of the structural advantage. But it does seem to me that at the present moment, my proposed reform of the electoral college would slightly tilt the playing field towards the GOP.

Just thought you might like to know.