Do The Republicans Have An Advantage in the Electoral College?

There is an outstanding fear that, as in 2000, Barack Obama could win the national popular vote while losing narrowly in the electoral college. It has been asserted that the Republicans have a natural advantage in the electoral college due to two factors. First, less-populated rural states are over-represented. In 2004, each electoral vote in Wyoming represented around 80,000 voters while each elector in California represented roughly 226,000 voters. I will return to this point later. Right now I want to focus on the second argument. Charlie Cook has asserted that Republican voters are distributed more efficiently than Democratic voters. To quote:

Add in that Democrats need to win the popular vote nationwide by somewhere between a half point and 1.5 points to overcome the more efficient allocation of Republican electoral votes (Republicans only "waste" a lot of extra votes in Texas; Democrats do it in California, Illinois and New York) and that points to a particularly close fall general election.

Charlie Cook is a very astute observer of politics, but in this instance he is just plain wrong. In fact, in 2004 it was the Republicans who “wasted” more votes than the Democrats. Bush won his states by an average margin of 18%, while Kerry won his by an average margin of 13% (10% if you exclude D.C). Furthermore, if you define “wasted votes” as any vote margin greater than one in a state a candidate carried, then Bush has a LOT more “wasted” votes than Kerry. Bush "wasted" just under 8.8 million votes (14% of his total), while Kerry “wasted” about 5.8 million votes (10% of his total).

To break it down further, you can divide the states into landslide states (won by 20 points or more), base state (10-20 pts), solid states (5-10 points), leaning states (2-5 points), and swing states (less than 2 point margins). In 2004, here is how they fall out, in order of size of margin for each candidate:

Chart #1: Distribution of States by Popular Vote Margin
Bush states:
Landslide: Utah, Wyoming, Idaho, Nebraska, Oklahoma, North Dakota, Alabama, Alaska, Kansas, Texas, South Dakota, Indiana, and Montana
Base: Kentucky, Mississippi, South Carolina, Georgia, Louisiana, Tennessee, West Virginia, North Carolina, Arizona
Solid: Arkansas, Virginia, Missouri, Florida
Leaners: Colorado, Nevada, Ohio
Swing: New Mexico, Iowa

Kerry states:
Landslide: D.C., Massachusetts, Rhode Island, Vermont
Base: New York, Maryland, Connecticut, Illinois
Solid: California, Maine, Hawaii, Delaware, Washington, New Jersey
Leaners: Oregon, Minnesota, Michigan, Pennsylvania
Swing: New Hampshire, Wisconsin

The electoral vote totals in each category makes the gap pretty stark:

Chart #2: Distribution of Electoral College Votes by Popular Vote Margin
Bush:
Landslide: 96 EV (33.6%)
Base: 87 EV (30.4%)
Solid: 57 EV (19.9%)
Leaners: 34 EV (11.9%)
Swing: 12 EV (4.2%)

Kerry:
Landslide: 22 EV (8.7%)
Base: 69 EV (27.4%)
Solid: 92 EV (36.5%)
Leaners: 55 EV (21.8%)
Swing: 14 EV (5.6%)

In other words, Bush ran up gigantic (and hence inefficient) margins in the majority of electoral college votes he won. 64% of his electoral college votes were from states he won by more than 10 points, while only 36% of Kerry’s were. Yes, the Democrats waste a lot of votes in New York and California, but all those 60-40 wins by Bush in the South, Rocky Mountains, and Farmbelt really start to add up.

Now one could look at this from a different perspective, namely that the Republicans have a larger base: they have more electoral college votes in their pocket that no Democrat has a chance to win. But this does NOT mean they have an electoral college advantage, per se, since it means that they are more likely to win the popular vote while narrowly losing the electoral college, assuming the national popular vote total is close. Yes, Democrats have a smaller core and thus have to win nearly ALL of the competitive states, but it also means that they have the potential to do so while losing the popular vote. Remember, heading into election night 2000, everyone was concerned that Bush would win the popular vote but lose the electoral college – not the other way around.

If you look at the historical record, the Bush victory in 2000 looks pretty unimpressive in electoral college terms. Rather than a lock, it looks like squeaking by. Normally the electoral college vote margin is substantially larger than the popular vote margin. John Kennedy won with less than .2% of the national popular vote but won over 300 electoral college votes. I have developed a “magnification rate,” which is the proportional increase in the electoral vote margin with respect to the popular vote margin. It’s determined by dividing the electoral vote margin by the popular vote margin. The magnification rate should give a reasonable indication of which candidates were particularly efficient in accumulating electoral college votes in any given election year. The chart below only includes those who won the popular vote. Every election since 1828 is included, except for 1860, 1912, and 1968 because those really weren’t 2-party elections. The chart below is ranked by the magnification rate.

Chart #3: The Magnification Rate of Popular Vote Winners



As you can see, George Bush’s 2004 magnification rate is one of the worst in the history of electoral politics at 2.64. His electoral vote margin (6.5 pts) was 6th worst all time. And let’s not forget that his popular vote margin was the lowest for an incumbent who won re-election since 1916, and the 2nd worst ever (so much for the “Bush Mandate”). Compare this with Bill Clinton’s 1996 magnification rate of 4.79, which is just under the median of 4.86. The most impressive electoral college performance of all time was James Garfield’s in 1880. He won by less than a tenth of a percent, but crushed Winfield Hancock in the electoral college. (By the way, here’s a nice bit of trivia: the biggest popular vote margin ever was Warren Harding’s in 1920 – 26.2 pts, and the biggest electoral vote margin was Franklin Roosevelt’s in 1936 – 97 pts)

I think part of the reason people like Charlie Cook may think the Democrats are at a disadvantage is 2000. Hey, you can see it right there at the very bottom. Gore indded lost even though he won national popular vote, but this is partly a mirage. Florida is a very large state and was the closest election that year, so a swing one way or the other is going to cause a big swing in the electoral college. If Gore had won by 537 votes rather than lost, he would have won the electoral college vote by an 8-pt margin, for a magnification rate of 16.41 – which would have been the 4th highest in history.

So much for the efficient allocation of popular votes.To return to the first point, that the overrepresentation of small states helps the Republicans, it should be clear from Chart #1 that Bush won his small states by ridiculous margins. Now, voters who “pile on” in a given state are in a strange sense disenfranchising themselves – they’d be better off if they were voting in another state. Bush’s huge margins in small states means that Republican voters are packed together in those states, which dilutes their over-representation. In addition, the Democrats win small states too (D.C., VT, ME, RI, DE, HA, NH), so in the end it pretty much cancels out.

So I posit something pretty controversial: I think that if anything the Democrats may have a slight advantage in the electoral college, because they tend to be very competitive in lots of big states. They can take fewer electoral college votes for granted, but they have an edge in enough of them to give them a realistic chance of winning the electoral college even if they lose the national popular vote.

One final note: the probability that there will be a divergence between electoral college and popular vote winners is most likely at the end of re-districting cycle, since the distribution of electoral votes will then be furthest from census totals. The one time the electoral college really “misfired” was in 1888 – 2 years before the census. The 2000 election (which wasn’t really a misfire because the Republicans probably stole Florida), was itself a re-districting year. Charlie Cook’s estimate that Obama needs to win by 1-1.5 points is really the opposite of the case. In fact, because Republican states are growing faster than Democratic ones, the electoral college is if anything more biased to the Democrats.

If the 2008 election is close, there is a greater chance this year than in 2004 that the popular vote winner will lose simply because it’s at the end of the decade. But I would hazard a guess that if it does happen, the odds are pretty good it will hurt the Republicans.