No, Illegal Immigrants Won't Elect Hillary
On Saturday, Paul Goldman and Mark Rozell authored a piece in Politico titled “Illegal Immigrants Could Elect Hillary Clinton.” Usually, a headline like that accompanies a story about voter fraud or back-of-the-envelope math on how immigration reform might create a huge bloc of new Democratic Hispanic voters (maybe even a “bonanza” for Democrats).
But Goldman and Rozell took a more interesting angle. They assert that the Electoral College is biased against Republicans because illegal immigrants (who can’t vote but are counted in the U.S. Census, thus influencing how many electoral votes each state gets) live mostly in blue states, and that this fact “significantly reduces the chances of the GOP winning the presidency”
Goldman and Rozell do the math on which states would gain and lose votes and find that if illegal immigrants and non-citizens were not counted towards electoral vote allocation, the GOP would likely gain four electoral votes and the Democrats could lose four. They then play out a plausible 2016 scenario in which the Republican candidate wins 266 electoral votes – falling exactly four votes short of the magic 270 needed to win the Oval Office.
The argument is interesting, but it’s ultimately quite speculative. Even if one accepts their demographic math, their assertion only works if the 2016 election is extremely close (which is far from a given), and even then there are good reasons to think that these four electoral votes won’t be the sole factor that makes the difference.
Uniform Swing – If 2016 Isn’t a Truly a Squeaker, This Doesn’t Matter
So let’s get an obvious fact out of the way: If this election isn’t close, then moving four electoral votes from one side to the other won’t matter. If Sen. John McCain had won four more electoral votes in 2008, he would have still lost, 361-177, to then-Sen. Obama. And if then-Vice President George H.W. Bush had lost four votes to then Massachusetts Gov. Michael Dukakis in 1988, Bush still would have won the Oval Office, 422-115.
Even in reasonably close elections, the winner can sometimes withstand a four electoral vote drop (see the 2004 election – President George W. Bush won the popular vote by 2.4 percentage points and won the Electoral College by 35 votes). So how close would the election need to be for illegal immigrants to swing it for Hillary?
It’s impossible to know the answer to that question with certainty, but we can use the 2012 election results to get a pretty good feel for it. Specifically, we start with the two-party popular vote share (Obama beats Romney, 52-47, for a four-point win) and then apply “uniform swing.” This term means that any swing in the popular vote is distributed evenly across the states. So if we want to know what the Electoral College might have looked like if Romney had won by 51-49, we simply add three points to his vote share (and thus take three from Obama) in every state. I applied uniform swing to estimate the Electoral College vote under the current allocation and under Goldman and Rozell’s scenario for eight different popular vote possibilities:
Uniform swing isn’t a perfect tool. Swing is never distributed perfectly evenly among states, as demographics change and different candidates appeal to different groups. (For example, if Florida Sen. Marco Rubio or former Florida Gov. Jeb Bush wins the GOP nomination, he might appeal to Hispanics more successfully than Mitt Romney did in 2012). That being said, it’s a simple, decent heuristic and most of the alternatives have significant assumptions baked into them.
The table suggests that the Republican candidate would need a roughly 1 percent win margin (that’s 50.5 percent for the Republican and 49.5 for the Democrat) in order for these four Electoral Votes to change the outcome of the race.
Elections that close don’t happen every four years. In 2000, the popular vote margin was within one point, but it wasn’t in 2004, 2008 and 2012. And before 2000 you have to go all the way back to 1968 (when Richard Nixon beat Democrat Hubert Humphrey and segregationist George Wallace) to find an election where the popular vote was within one point. Simply put, election results vary and most of them haven’t had the popular vote within one percentage point. That’s not to say 2016 couldn’t be that close – it could – but there’s no reason to assume this early in the cycle that it will be. And if the race isn’t very tight, then these four electoral votes almost assuredly will not make the difference between a win and a loss.
Even if the Election Is That Close, Illegal Immigrants Probably Won’t Make the Difference
So if you accept my argument so far, there’s a natural follow-up: If this election just so happens to be extremely close, will this specific issue – the four electoral votes Democrats “gain” because of illegal immigrants – give Hillary the White House?
The answer is both a yes and a no. The answer is yes in the sense that the GOP nominee could edge out the Democrat in the popular vote but still lose in the Electoral College by four votes. In that case, allocating electoral votes by citizen-only population rather than total population would have changed the outcome.
But the answer is also no, and the best way to see this is to think in terms of over-determination. Over-determination is a pretty simple concept – if something is over-determined, there are a bunch of different causes for it and it’s impossible to parse out just one that made it happen. The final result of a really close election will almost certainly be over-determined. Maybe a campaign picked the wrong messaging strategy or maybe a party picked a candidate who was too ideological and not quite electable enough or maybe one of the campaigns chose to focus on the wrong swing states – there’s a laundry list of factors that could put a candidate over the top or cause him or her to miss the mark in a really tight race (although when the race is less close and fundamentals such as the economy favor one party, factors like these often do not make a big enough difference to change the final election result). I don’t see a good reason to focus primarily on illegal immigration when a large number of other factors could just as easily shift the final result.
But What About the Electoral College Scenario?
Goldman and Rozell also sketch out a scenario in which the GOP nominee wins all the states Romney won plus Ohio, Virginia and Florida. This gets the candidate to 266 electoral votes, and he or she needs 270 to win. Rozell and Goldman believe that Iowa, Colorado and New Hampshire would be the next swing states a GOP candidate might target, but they also assert that these states would be tough pickups. The authors make this claim primarily by citing demographic trends in Colorado and recent election results – that the GOP won New Hampshire and Iowa only once (in 2000 and 2004, respectively) since the 1988 election.
Arguments like these are basically a restatement of the “blue wall” theory – that if you look at recent electoral results, Democrats have a “blue wall” of electoral votes that would be difficult to break through. I won’t go into detail rebutting this because Nate Silver dealt with it very well earlier this year. I won’t rehash all of his arguments, but he essentially uses uniform swing and related techniques to show that swing states the GOP hasn’t won recently (e.g. Iowa, New Hampshire, Colorado, Pennsylvania, etc.) aren’t actually out of reach. It just looks that way because Republicans have only won the popular vote once since 1988, and it’s hard to win swing states that lean slightly away from your party without winning the popular vote. And if Republicans get a solid win in 2016, the swing states will likely follow – whether they’re packed with illegal immigrants or not.