Friday, December 26, 2008

A Projected 2012 Electoral College Map (version 2.0)

For a look at the 2012-2020 electoral college map based on the 2010 Census click here. And for look at how those changes would have affected the 2008 presidential election click here.


On Monday, Election Data Services released the US population estimates for 2008 and a series of projections for 2010. So let's have a look at how the electoral college map would change given the new information we have.

The first set of numbers we have is the estimate of the population changes during 2008. This is more of an "if the Census was done today" scenario. It was this set (from 2007) that we used to project the the map for 2010 before. In other words, this gives us an idea of what the changes might be for 2010, but not the full picture. EDS even said in its report that they expected quite a bit of volatility over the next two years despite the economic downturn's impact on mobility.

[You'll notice that the map has taken on the red and black color scheme that denotes gains and losses in the financial world. Red states are states losing seats while black states are those likely to gain seats following the 2010 Census.]

[Click Map to Enlarge]

Despite that, we can set a baseline of potential changes based on the changes from 2008 (Well, from 2000-2008). The picture here isn't that much different from what we witnessed a little more than a month ago. Basically, Texas would gain an additional seat, Missouri wouldn't lose the seat it was projected to have lost and Michigan and New Jersey would lose a seat apiece. That would have fueled an additional two electoral vote swing toward John McCain in November's electoral college. Taken with the information we had before, though, that still would have put the Arizona senator at a 360-178 electoral vote disadvantage. Still, we're talking about a population shift toward redder states. Of course the traditional question is, "Who are those people?" Are they more Democratic or Republican? Are these states becoming redder, like, say, Georgia, or purple like Virginia (though the latter is not a state projected to gain or lose any congressional seats)?

But let's move from an estimate of the map based on the population changes we have seen during 2008 and focus on the projected changes we could see in 2010 based on likely changes over the next two years. Now, EDS set up several different models: one that projected population shifts based on what has happened since 2000, and four additional models that took a midterm approach; focusing on population shifts from 2004 onward, 2005 onward, 2006 onward and 2007-2008. For our purposes, we'll take the model with the most information: the full model based on the changes since the last Census.

That version shows Arizona (2 seats total), Florida (2) and Texas (4) gaining additional seats on top of the gains shown in the map above. It also has South Carolina picking up a seat.

[Click Map to Enlarge]

On the opposite side of the ledger, Ohio is likely to lose a second seat, while Missouri once again loses the seat it lost on the initial version of this map. Illinois and Minnesota round out the remaining list of states either losing population or not growing at rates fast enough to stay apace of the states gaining seats. If this projected map had been used in the November election that would have netted McCain an additional two electoral votes from the first map above. In other words, playing on the map immediately above, the 2008 election would have come to a count of 358-180 in favor of Obama. Again, this isn't a big shift in an election like we just witnessed. In a more competitive election, however, such a shift could prove consequential. Recall that the generic ballot question favored the Democrats throughout the 2008 cycle. If that divide had been more even, a seven electoral vote swing could have been decisive.

So what does all this mean? Well, not much. Projections are nice, but they aren't the real thing. I really ought to look back at how well these EDS projections were prior to the last Census. That might be a factor to throw into a more inclusive model -- similar to the house effect that FiveThirtyEight factored into their electoral college projections late in the game in 2008. With EDS cautioning that there could be volatility in the forecasting over the next two years, you have to take these numbers with a grain of salt. However, it is interesting to note what the map might look like in four years' time.


Recent Posts:
The Race for RNC Chair

Backloading in 2012? Arkansas is Moving Closer

The 2012 Presidential Primary Calendar

9 comments:

  1. Democrats still seem to have an advantage. I know this is a rather crude way of assessing this, but I made a map in which states which went more Democratic than the country went blue and those more Republican than average were red, in an attempt to see who would win the average election that was tied in the popular vote. In the current map, Democrats would have 278 EVs and Republicans 247, with Virginia essentially mirroring the national popular vote and therefore a tossup. Under the changes from your second map, if my math is right, it would still be 272-253. Even though my method doesn't take into account a number of important factors, that seems to be enough of a gap to illustrate that even after the changes, the electoral map will structurally favor Democrats.

    ReplyDelete
  2. Jack--interesting! Your new numbers are close enough that the home states of the candidates become very important. If McCain wasn't from Arizona, it pretty clearly would have gone blue this time. On the other hand, if he'd been from Ohio, it might have gone red.

    Josh--It's interesting to me that the shifts are taking party away from the big, traditional swing states: Ohio, Missouri, and Pennsylvania lose. Even though Florida gains, it doesn't make up for that. I think 2000 or 2004 style campaigns, where one state becomes the entire focus (Florida in 2000 or Ohio in 2004) become somewhat less likely, as it becomes easier for a group of small states to offset the loss of a big one.

    ReplyDelete
  3. Scott, I'm not sure that I would say that the likelihood of an election being concentrated in one state is shrinking — I think it's already gone. We saw quite a high number of states in play this year. In fact, I think the battlefield is going to shrink as we see a bit more polarization. We've seen some of the old bellwether states move away from that status; Missouri, for example, and other swing states like OR, PA and NH seem to be trending Democratic, while others such as MO, are trending Republican. Yes, those states are being replaced by others such as North Carolina and Indiana, but that already happened this year. The only state that wasn't a swing state this year but seems likely to be one in the future is Arizona. In short, the era of the one-state election is not just dying; it's already dead, and while we're quite far from it now, we may move somewhat closer in the next few years, before a new batch of swing states arrive.

    One more thought: If we use the state PVIs based on 2000 and 2004 — I don't have the data for 2004 and 2008 — Democrats have an even more pronounced 296-242 advantage. With the new map, that's still, if my counting is right, 291-247. Eliminating +/- 0 or 1 states would give Democrats an advantage of about 60 EVs.

    ReplyDelete
  4. Alright folks. I'm back. I hate to post something like this and then not be around to talk about it. But I was on the road over the weekend.

    A couple of things:
    1) Jack, I concur that structurally the Democrats do have an advantage in the electoral college (and likely in congressional seats following the 2010 census), but I'd temper that with one recent historical footnote. It wasn't that long ago that pundits were talking about the GOP having a similar advantage that would lead to their domination of Congress and an edge in the electoral college. In other words, things change and can do so rapidly. Caution aside, in the near term, I think it is relatively clear that the Democrats have an advantage, and unless, as I've said several times, they really mess things up, they are likely to stay there. Not, perhaps, to the degree they are now, but plus or minus a few seats or electoral votes here and there.

    2) To call the types of elections that we had in 2000 or 2004 dead is a bit strong. [Yes, this is more semantics than anything.] They are hibernating, perhaps, but they aren't dead. I agree that such elections don't look likely after last month, but let's see what happens once the Democrats assume power in a few weeks.

    Let's put it this way: The GOP is going to have to come up with a different strategy than the old confederacy plus some other states -- even with these changes -- to be successful. And that is going to be the question of the next two years.

    Good discussion, folks.

    ReplyDelete
  5. In which state is a one-state election most likely? Probably depends on the candidates. VA was the state closest to the national vote this year but is still trending Dem. It could still happen in FL; a lot of votes there and Obama only underperformed his national totals by a few points.

    ReplyDelete
  6. Jack,
    Let's look to the Electoral College Spectrum on this one. I'll put up a separate post shortly.

    ReplyDelete
  7. This is a very good question, Jack. I've been sitting on this without comment since Ballot Access News first started talking about it last month. I've dropped the ball on this one.

    ReplyDelete
  8. Oops. I hit publish accidentally.

    Yes, there will be an updated version when Obama signs his name to the bill.

    ...and it will make it to his desk.

    ReplyDelete