To begin, consider states that would be expected to be safely Republican or Democratic. The model I use is based on trends evident in states from 1992-2012. (The particular beginning point, whether before or after 1992, does not matter much.) The estimates indicate that an evenly split popular vote in either 2012 or 2016 would yield 23 states, with 191 electoral votes, where the Republican vote margin exceeded 10 percentage points. It also predicts that 13 states (plus the District of Columbia), with 179 electoral votes, would have margins more than 10 points favoring the Democrats. In a close election, the outcome would be determined by the 168 Electoral votes in the 14 remaining “swing” states. To win, the Republicans would need 79 of these Electoral votes, and the Democrats would need 91.
Even though the Democrats would need more Electoral votes from the swing states, they would have a significant advantage. The Democrats won 13 of the 14 states in 2012 (all except North Carolina). Of course, they won the popular vote by 3.9 percentage points in 2012. Had the popular vote been even, I estimate that the Republicans would have had a better than 50 percent chance in just three of the states (North Carolina, Florida and Ohio) with 62 Electoral votes. In 2016, the Republicans’ chances would be marginally lower in all three. The estimated probability that they’d win all three states if the popular vote was evenly split is just 48 percent.
While swing state trends look good for the Democrats, the same is not apparent in the rest of country. In the remaining states, there is more movement toward the Republicans than the Democrats. In the 23 safely Republican states, continued movement toward the Republicans is occurring at a rate about 25 percent faster than the average movement toward the Democrats in the 14 safely Democratic seats. In fact in 10 of the 12 states that are changing fastest, the movement is toward the Republicans, not the Democrats. However, these trends have virtually no influence on the chances of victory in a presidential election because with a winner-take-all system like the Electoral College, additional votes in a state that is already safe for one party are “wasted.”
In 2000, the Republicans won the presidential election even though they lost the national popular vote. The same appears unlikely in 2016. To win enough Electoral votes for victory will require making inroads into states that increasingly lean toward the Democrats. As shown in the figure at the top of this post, the estimates from the model I am using suggest that to have a 50 percent chance of winning the Electoral College the Republicans would have to win the popular vote by a margin of between one and two percentage points.