Republicans in Pennsylvania are considering a proposal that would award 18 of the state’s 20 electoral votes to the winner of each of its congressional districts, leaving the remaining 2 to the winner of the state at large.
Had the proposal been in place in 2008, when Pennsylvania had one more electoral vote prior to reapportionment, Barack Obama would have carried only 11 of the state’s 21 electoral votes despite winning Pennsylvania by a 10-point margin.

The plan would create the possibility, in fact, that the winner of the popular vote in Pennsylvania could come away with fewer electoral votes from the state. Pennsylvania’s congressional districts are heavily gerrymandered and the Democratic vote is concentrated in the dense urban areas of Philadelphia and Pittsburgh. Republicans have a 12-7 edge now in the state’s congressional delegation, and are in charge of its redistricting process. With Pennsylvania needing to lose one seat because of slow population growth, they will try to consolidate Democratic voters into six districts while reinforcing their own.
A couple of the Republican-held districts might still be in play. The one occupied by Representative Charles Dent, for instance, now designated as Pennsylvania’s 15th Congressional District, has been slightly Democratic-leaning in its presidential voting and was carried by Mr. Obama in 2008. But if Mr. Obama were to carry Pennsylvania narrowly, its electoral votes could plausibly split 9-11 or even 8-12 against him.
The Constitution provides broad latitude to states to determine how to allocate their electoral votes. In fact, Democrats have little recourse in Pennsylvania at all: Republicans control both branches of the state legislature as well as the governorship. And Pennsylvania has neither ballot initiatives nor recall elections. Democrats simply have to hope that Republicans will decide that the proposal is not in their overall best interest.
Fortunately for Democrats, the proposal is already drawing objections from some Republicans. And it’s easy enough to see why: there are all sorts of downsides to this plan for Pennsylvania Republicans.
Problem #1. The Electoral College split could work against Republicans and cost their candidate the election.
It’s easy to draw maps where the Pennsylvania plan could cost Mr. Obama an election he otherwise would have won. One case is if Mr. Obama were to lose Ohio, Florida, Nevada, Indiana and North Carolina from among the states he carried in 2008, holding the rest including Pennsylvania.
That would result in a 279-259 win for Mr. Obama in the Electoral College under the current rules. But, since he would probably lose between 10 and 12 of Pennsylvania’s electoral votes because of the congressional district split, that would put him below the 270 votes required to win the Electoral College.

But one can also come up with plausible maps that would work the other way. Take the map we just drew and make two small alterations: Mr. Obama holds onto Nevada, but loses Pennsylvania. (Seem unlikely? It isn’t: Mr. Obama carried Nevada by a wider margin than Pennsylvania in 2008.)

Under the old rules, this would be enough for the Republican candidate to become the next president, winning the Electoral College 273-265. If Pennsylvania’s votes were split, however, Mr. Obama would be very likely to carry at least five of them in Philadelphia and Pittsburgh. That would be enough for him to win the Electoral College 270-268.
Over all, somewhat more of these scenarios are likely to work for rather than against Republicans. That’s because of the way Pennsylvania’s congressional districts are drawn, and because Pennsylvania has historically been slightly Democratic-leaning relative to the rest of the country.
Still, the relative order of states shifts a little bit from election to election, and Pennsylvania is close enough to the razor’s edge that it could flip from Democratic-leaning to Republican-leaning.
As recently as 1996, Pennsylvania was what I call the “tipping-point state.” Bill Clinton carried 21 states and the District of Columbia by wider margins than he did Pennsylvania, giving him 256 electoral votes. There were 28 states, meanwhile, totaling 259 electoral votes, that Bob Dole either won outright or lost by a smaller margin than Pennsylvania.
Neither figure is sufficient for a majority of the Electoral College — so had the 1996 election been closer over all, Pennsylvania would have been the decisive state. Mr. Clinton could have won Pennsylvania but lost the Electoral College under the proposed rules. But the same is true for Mr. Dole.
Problem #2. The plan could undermine the integrity of the Electoral College, which is probably not in Republicans’ long-term best interest.
Two states, Maine and Nebraska, already award some of their electoral votes by congressional district, although their votes had never actually been split before 2008, when Mr. Obama won Nebraska’s 2nd Congressional District in Omaha to take 1 of the state’s 5 electoral votes. This quirk hasn’t led to the collapse of the Republic, or the Electoral College.
But there are some important differences between these states and Pennsylvania. First and most obviously, they are much smaller, having 9 electoral votes between them, versus the 20 that could be affected by Pennsylvania’s plan. And Maine and Nebraska are not usually swing states whereas Pennsylvania very much is, so their rules are exceedingly unlikely to change the winner of the election.
Second, the candidate winning the popular vote in Maine and Nebraska is guaranteed to also win a majority of its electors. Nebraska awards 2 electoral votes to the overall winner of the state, as well as 1 in each of its 3 congressional districts. But since it is mathematically impossible to win a state without winning at least one of its congressional districts, the worst that the statewide winner could do is 3 out of 5 votes total. In Maine, meanwhile, the winner of the state is guaranteed 3 of 4 electoral votes.
As we mentioned, that isn’t the case in Pennsylvania. It’s quite plausible that a candidate could win the popular vote there but lose most of its electoral votes.
Finally, the Pennsylvania plan is much more explicitly partisan. And because of the geography of the state, it would give redistricting and gerrymandering a much larger role in the outcome of presidential elections. (There isn’t much worth gerrymandering in Maine or Nebraska.)
So for all practical purposes, this would represent a paradigm shift in the way that presidential elections are decided.
There’s no reason that Republicans couldn’t adopt a similar proposal in Wisconsin or Michigan, where they also have control of the state government.
There aren’t any particularly good places for Democrats to retaliate this year. They control the state governments in places like Arkansas and West Virginia, but they are small states and might not allocate many electoral votes to Democrats no matter how they were split.
But if they swept the statehouse in Texas one year, or in Georgia, or Missouri, what would stop them from adopting the Pennsylvania plan if this were the new normal?
It would be one thing if all states adopted the same method, some of which could have advantages to the current approach, and which are sometimes favored by political scientists and good-government advocates.
But what you’d probably wind up with instead is a patchwork of procedures for awarding electoral votes: some states would retain the current winner-take-all method, but some would use the congressional district split. Some might divide their votes proportionately, or use other hybrid approaches. And the method could change each time that the state government changed hands.
It’s perfectly Constitutional, for that matter, for states to award their electoral votes through the state legislature, as many states did in the early years of the Republic, without taking a popular vote at all. If Republicans want to all but guarantee that they win the presidency next year, there is nothing stopping them. They control the state governments in 21 states totaling 242 electoral votes. All they need to do is have their state legislatures pick Republican electors in those states, and then for their candidate to win by popular vote in Alaska, Arkansas, Kentucky, Mississippi and West Virginia, which would get him to 270 electoral votes.
That, of course, is an extreme and unlikely example. But the Pennsylvania play would undermine the integrity the Electoral College, which is already fairly unpopular.
Eight states and the District of Columbia, totaling 132 electoral votes, have signed on to a plan called the National Popular Vote Interstate Compact, a proposal that would use a workaround to guarantee that the winner of the national popular vote would win the Electoral College once states totaling a majority of 270 electoral votes had committed to the plan. Such efforts would presumably gain steam if the Pennsylvania plan were adopted, as would proposals for changes to the Constitution.
And that’s probably not in Republicans’ best interest: the Electoral College will generally favor them.
Take the 50 states and the District and Columbia and divide them into three roughly-equal sized piles:
In the red pile go all of the states that John McCain carried, which currently have a total of 170 electoral votes.
Into the purple pile go states with 182 electoral votes: all states that voted for both parties at least once between 2000 and 2008, plus Pennsylvania, Wisconsin, Michigan and Minnesota, where the results are usually close.
The remaining 186 votes are in Democratic base states and go into the blue pile.
The blue states award a total of 4.1 electoral votes per million voters, based on turnout in 2008. The red states, by contrast, award 4.9 electors per million votes.

So Republicans get about 20 percent more Electoral College mileage out of their base states than Democrats do, relative to the number of votes cast. This reflects the fact that Republican states tend to have smaller populations, and the Electoral College slightly favors low-population states. (California has 66 times as many people as Wyoming, but awards “only” 18 times as many electoral votes.)
In any particular election, this advantage may flip. Our simulations in 2008 suggested that Barack Obama was more likely to lose the popular vote but win the Electoral College rather than the other way around, a reflection of the fact that his campaign over-performed in swing states because of its superior turnout operations.
Over the medium-to-long term, though, the Electoral College should increase the number of Republican presidents — an advantage they would presumably want to keep rather than undermining the system.
Problem #3. The plan could motivate Democrats and lead to higher Democratic turnout both in Pennsylvania and nationally.
I don’t entirely buy that President Obama has problems with his base — his numbers are still pretty decent with most Democratic core voters. But they have eroded some, even among African-Americans. And even if these voters still support Mr. Obama, they may not turn out, or they may turn out but decline to devote their time and money to his campaign.
I can’t think of many better ways to motivate these voters than to convince them that Republicans are trying to steal the election, and remind them of what happened in 2000, themes that will become prominent should the Pennsylvania plan come to pass.
Is the possibility of winning 10 or 12 extra electoral votes a good trade if the price, say, is motivating $100 million in additional donations to Mr. Obama and the D.N.C.? Or boosting Democratic turnout by 5 percent?
These effects are tricky to quantify. But considering that the Pennsylvania plan will help Republicans only under a relatively specific set of circumstances — if the election is close over all, if Pennsylvania is close to the electoral tipping point, and if Mr. Obama rather than their candidate wins the state — it’s hard to know who gets the better of the deal.
Problem #4. The plan would significantly reduce Pennsylvania’s influence in the election campaign.
The first three problems concerned the Republicans’ broad self-interest in electing Republicans to the White House both in 2012 and in future years. But Republicans in Pennsylvania also have narrower and more parochial things to worry about.
If the Republican plan were adopted, its two statewide votes would still be worth going after, as might the electoral votes three or four of its congressional districts. The others would be foregone conclusions, heavily advantaging one or the other party, which would lose them only in the event of a national landslide.
Thus, Pennsylvania would effectively demote itself to something like New Mexico in the electoral pecking order — a state with five or six swing votes rather than 20. That means fewer favors from Washington, fewer visits from the candidates, less of a windfall for the state’s economy, and less face-time for its politicians.
Problem #5. The plan would probably become unpopular in the state over time, potentially costing some Republican office-holders their jobs.
That reduction in influence could also make Pennsylvania voters unhappy. Consider what happened in Colorado in 2004, when there was an initiative on the ballot to allocate the state’s electoral votes proportionately.
Initially, the initiative had plurality support in the polls. Once opponents began to point out that it would vastly reduce Colorado’s influence, however — essentially, only 1 of its 9 electoral votes would be in play — its support collapsed to the point that it was resoundingly defeated, 66-34, on Election Day.
The Pennsylvania and Colorado cases are not identical: Colorado’s plan would have split its votes proportionately according to the popular vote rather than by congressional district.
The experience suggests, however, that opponents of the bill are likely to have the stronger case in the court of public opinion: the argument that Pennsylvania has been made less influential in determining the next president will tend to carry the day.
Since there is no ballot initiative to vote on in this case, voters might instead take their frustrations out on Republican politicians throughout the state.
Republicans control 112 of the 203 seats in the Pennsylvania State Assembly, with 102 votes being required to pass the bill. Republicans hold a 30-20 majority in the Pennsylvania State Senate, meanwhile, so would need to hold 26 of their members to pass it there.
My guess is that they will not be able to keep their coalition together: it’s not clear if this plan is in their best interests. There are reasons why states have been loathe to tinker with the Electoral College despite having the latitude to do so.