Census Impact on 2012 Presidential Race is Limited

Census Impact on 2012 Presidential Race is Limited
By NATE SILVER

Under the new allocation of electoral votes that will go into effect in 2012 as a result of this year’s Census, President Obama would have won 6 fewer electoral votes than he did in 2008, gaining 2 in Florida and 1 in each of Washington and Nevada, but losing 2 in both Ohio and New York, and 1 in New Jersey, Pennsylvania, Illinois, Iowa, Michigan and Massachusetts.

Mr. Obama would nevertheless have won his election against John McCain; his revised total of 359 electoral votes would have been well in excess of the 270 required. But it is not that hard to come up with scenarios in which the new electoral math might have altered the outcome of the election. If, for instance, Mr. Obama had won each of the states that John Kerry did in 2004, as well as Ohio, his total would be 264 electoral votes under the new allocation, leaving him 6 shy. By contrast, those same states would have been worth 272 electoral votes in 2008, giving him a narrow victory.

Over all, however, it would be easy to exaggerate the difference that these shifts in the electoral college might make.

Consider this: under the new electoral college allocation that will go into effect in 2012, the outcome of every presidential election in the past century would have been the same had the new numbers been used.

Jimmy Carter would have won 288 electoral votes in 1976 rather than 297, for instance — but still a sufficient number for victory. John F. Kennedy’s victory over Richard Nixon in 1960 would have been more significantly impacted: instead of winning 303 electoral votes, he would have won just 273 (assuming one gives him no credit for the District of Columbia, which did not vote for President until 1964). Still, that would have been just enough for him to beat Nixon.

Other candidates who won in close elections would have benefited. George W. Bush would have taken 285 electoral votes in 2000 rather than 271 — although Al Gore would still have prevailed had he won Florida. Woodrow Wilson’s victory in 1916 was close: he took 277 electoral votes, barely more than the 266 that were then required. Under the 2012 rules, however, he would have won at least 345 electoral votes, giving him a robust victory. (Wilson’s weakness in 1916 was in the Northeast, which has steadily lost electoral votes and which was then heavily Republican.)

And over the broader course of American history, years like 1916, 1960, 2000 and 2004 are really the exceptions. The Electoral College tends to exaggerate even relatively small margins in the popular vote, and most of the time, it isn’t especially close at all. In the 25 elections since 1912, the winning candidate has finished with at least 350 electoral votes on 19 occasions — more than three-quarters of the time. The new allocation will only matter if Mr. Obama had been destined to win in the first place, and then only by a very small margin.

That does not mean that their impact will be zero: there is certainly a theoretical chance that they could change the identity of the next President. But the probability is likely not much more than about 2 percent. (If, for instance, Mr. Obama’s chances of re-election had been 56 percent under the 2008 electoral vote rules, perhaps they are 54 or 55 percent now.)

One simple way to look at the problem is as follows: there are 538 electoral votes, of which 6 have now shifted from states that Mr. Obama won to ones that John McCain did. Six divided into 538 is around 1.1 percent.

That fraction — a little over 1 percent — is probably a good proxy for the likelihood that the new allocation will change the winner of the election. Over the course of time, the number of electoral votes received by the Democratic candidate is quite evenly distributed over the possible range from 0 votes to 538:

What this means is that if you wanted to to estimate the probability distribution for the number of electoral votes won by the Democrat, you couldn’t do much better — in the absence of other, more specific information — than simply picking a random number between 0 and 538. If the Democrat were then given a penalty of 6 electoral votes, the outcome of the election would only be altered if the random number had been between 270 and 275 votes originally, of which there would be a 1.1 percent chance.

Of course, we do have other information about the 2012 election, and some of it suggests that it is at least somewhat more likely than usual to be close. For one thing, electoral vote (and popular vote) margins have become slightly closer on average over the course of the past century, although the trend is not statistically significant. Perhaps more salient is that Mr. Obama’s approval rating has hovered in the mid-40s for the past several months, which is right about at the threshold where an incumbent President is about as likely as not to win re-election. If an election were held tomorrow, it would almost certainly be quite close, and the the vagaries of the electoral math might matter.

So perhaps the impact is more like 2 or 3 percent than 1 or 2 percent. But as much fun as we expect to have in gaming out Electoral College scenarios over the next two years, it is probably not much more than that. A president’s current approval rating is at best a very weak predictor of what it will be two years hence, and the election could easily turn into a landslide either for Mr. Obama or against him. And even if the election is close, it is nevertheless a long shot to be quite close enough that 6 electoral votes would make the difference.

Finally, Democrats would do well to keep in mind that the population shifts that have triggered the changes in the Electoral College are not unfair to them. In fact, the opposite is true. Our simulations throughout 2008 showed that Mr. Obama, not Mr. McCain, was more likely to win the Electoral College despite losing the popular vote. The changes, by bringing the electoral vote into better alignment with the population, will reduce that disparity.

Nor have the population changes worked uniformly to Republicans’ benefit. States like Virginia, North Carolina, Colorado and Nevada — all of which Mr. Obama won in 2008 — have become more competitive in large part because people moving to those states have tended to be more liberal than those who were there previously. So if winning Mr. Kerry’s states, plus Ohio, will no longer suffice to re-elect Mr. Obama, he will also have more opportunities than Mr. Kerry did to make up the difference.