The College try: Changes in Electoral College scheme could dramatically affect outcome of presidential ballots

Lawmakers in Pennsylvania are considering switching the way the state divides up its electoral votes before the 2012 election. Pennsylvania now gives all of its electoral votes to the winner of the popular vote--the same method that 47 of the other 49 states also use.

But Pennsylvania is now considering following the lead of the two outlier states, Maine and Nebraska. These states--which have far smaller populations and influence on the national election scene than Pennsylvania does--follow an apportionment system that provides one electoral vote for each House district that a national candidate carries, and two electoral votes to the winner of the state's popular vote.

The way that states divvy up their electoral votes is very consequential, for the simple reason that electoral college--and not you, the U.S. electorate--determines the winner of any presidential election. That was, of course, the indelible lesson of the 2000 election, which went to George W. Bush based on electoral college apportionment as Al Gore won the popular vote. And electoral college considerations also govern the behavior of the predictions markets. When I analyze those markets to conclude that Barack Obama has a 48.8 percent likelihood of winning re-election, I am referring to his likelihood of winning the Electoral College, not the popular vote.

The basic scheme of representation in the Electoral College follows the allotment of congressional districts: that is, rather than hewing to the one-person, one-vote model of popular elections, it assigns one elector's vote for each district represented in the House of Representatives, and two per each state's senate delegation. This creates obvious distortions in electoral clout when the College assembles to designate a new president. California, for example, has 12.1 percent of the U.S. population, but just 10.2 percent of the nation's presidential electors. At the other end of the spectrum, Wyoming has 0.18 percent of the U.S. population but 0.56 percent of electors--more than triple the representation of the state's population. You can see where your state stands in the table below, which shows each state's standing in the electoral in ascending order of state population:

For every state but Maine and Nebraska, the Electoral College gives all of its electors to the winner of the state's popular vote. That means that if Candidate A obtains 49.9 percent of the vote and Candidate B obtains 50.1 percent of the vote, Candidate B gets 100 percent off the state's Electors. In Florida in 2000, when the Supreme Court finally halted the recount, George W. Bush had 2,912,790 votes and Al Gore had 2,912,253 votes. That margin was enough to win Bush the presidency, since it also won him the entire slate of Florida's 25 electors.

And even with Maine and Nebraska adopting a slightly more proportional approach to doling out electoral votes, those states have largely followed the same winner-take-all script. Only once, in 2008, has either Maine or Nebraska not given all of their electoral votes to one candidate. This is largely a function of simple math: With just three (for Nebraska) and two (for Maine) House seats, it's mathematically impossible for the winner of the state's popular vote not to win most of its electoral votes as well. A candidate could dominate one district in Nebraska and lose the other two, but he or she would still get three electoral votes to the other candidate's two electoral votes.

In 2008, Obama received 55 percent of Pennsylvania's vote and all of the crucial swing state's 21 electoral votes; under the new plan he would have received just 11 of the 21 electoral votes. Since Democratic candidates tend to win their districts by larger margins than the Republicans, a proportional system in more populous states could work to the GOP's advantage in presidential races. There's a strong likelihood that Democratic presidential candidates could win the state's popular vote by rolling up higher margins of victory in dense, urban congressional districts and get less than half of the state's electoral votes.

By itself, in the near-term, Pennsylvania's change will make it more likely that a candidate can win the national popular vote and lose the Electoral College. Recent experience suggests that Obama would be well positioned to carry Pennsylvania's popular vote in 2012, since Democratic candidates have won Pennsylvania's vote in five straight elections. But since Republican candidates have an advantage in many of the smaller states that have disproportionate electoral votes, a weaker Democratic showing in Pennsylvania's electoral-college tally could produce any number of variations on the 2000 experience, where the winner of the national popular vote does not win the presidency. Or to put things a bit more mathematically: This rule change is likely to decrease the proportion of electoral votes that a Democratic candidate receives relative to his or her national popular vote.

There are some important caveats to such predictions, however. First, Republican could regain control of Pennsylvania. The state just elected a conservative Republican senator and the GOP currently holds 12 of the state's 19 House seats. (Pennsylvania lost one seat in the 2010 Census, so it will have only 18 House seats and 20 electoral votes in this election.) Second, a move to proportional electoral college representation in Pennsylvania could spur different approaches to electoral form aimed at increasing the role of the popular vote. Some reformers have proposed eliminating the electoral college all together; others have proposed inducing individual states to pledge their electoral votes to the winner of the national popular vote. Meanwhile, other reformers seek to move the entire Electoral College on to the Nebraska-Maine model that Pennsylvania is weighing now--a move that could well further dilute the impact of the popular vote on a presidential election's actual outcome. All one can confidently predict right now is that a new electoral-college scheme for a state as large and influential as Pennsylvania will likely trigger changes of some kind across the rest of the system.

As Americans ponder the prospect of such reforms, here are some quick and useful side-notes on the present composition of the Electoral College:

First, beyond the disproportionate distribution of electoral votes, millions of American citizens actually have no Electoral votes. Washington DC has no congressional representation, but was granted the equivalent amount of Electoral votes in the 23rd amendment to the Constitution. Puerto Rico, with more than 3.7 million people, would receive seven electoral votes if it was granted the same rights or was a state; according to the 2010 Census, its population is just greater than Connecticut and just less than Oklahoma. Its citizens currently have zero electoral votes.

Second, in case you are wondering, there are 538 electoral votes, making it possible to for any given presidential ballot to end up in a tie. In the event of a tie, the vote will go to the current House of Representatives; House members would then determine the election by state. Since the Republicans presently control more states, a statistical tie in the Electoral College would likely translate into a Republican victory. (The closest that the United States has come to an Electoral College tie was the 1876 presidential election, where a deal to suspend Reconstruction in the South staved off the prospect of a House vote deciding the contest between Republican Rutherford B. Hayes and Democrat Samuel Tilden, the winner of the popular vote.)

Third, nothing in the Constitution mandates that the popular vote need have any sway in designating presidential electors. A state legislature could decide to appoint the electors itself, as was the norm for the first few U.S. presidential elections. Of course, any state lawmakers pulling such a move now would likely be unpopular in the next state legislative election cycle.

Polls on the national popular vote are very informative, but at least three times since the United States first started counting popular votes, the top vote getter has been denied the presidency--in 1876, 1888, and 2000. (We could nudge that total up to four, if we count the still-controversial 1824 vote.) So as we furiously monitor the many shifts in national political polling over the next 14 months of the 2012 election cycle, remember that the numbers that really matter are the ones governing the breakdown of the vote among the 538 votes in the Electoral College.

David Rothschild is an economist at Yahoo! Research. He has a PhD in applied economics from the Wharton School of Business at the University of Pennsylvania. His dissertation is in creating aggregated forecasts from individual-level information. Follow him on Twitter @DavMicRot and email him at PredictionBlogger@Yahoo.com.