In democratic theory, the ballot is the most perfect expression of the democratic commitment to the moral equality of persons. Every citizen, whether old or young, rich or poor, sophisticated or simple, enjoys the same endowment in an election: a single vote. The ballot not only gives citizens their voice in government, it also makes their voices equal.
In practice, however, democracies have erected all sorts of impediments to the ideal of equal voice, such as restrictions on suffrage, legislative malapportionments, and discriminatory gerrymanders. Among the most egregious impediments, however, are surely the systems of indirect election purported to filter and to refine the voice of the people. The Electoral College is one such system. This Commentary examines the effects of that system and the proposed reforms to it on the prospect of equal voice in elections.
I. The Electoral College Now
The Electoral College effectively divides the selection of the president into fifty-one separate elections in the fifty states and the District of Columbia. In every state except two, state electorates choose slates of electors pledged to support the winner of the statewide popular vote. Like all winner-take-all electoral systems, the Electoral College is exceedingly responsive: small changes in voter support magnify into large mandates in the Electoral College. Moreover, like all electoral systems divided into geographic districts, the Electoral College exhibits a representational bias, in that a candidate who falls well short of a majority of the popular vote can still win the presidency. Electoral College anomalies of this kind have occurred in four out of the fifty-five presidential elections held since 1789-and most recently in 2000.
Because the number of electoral votes for each state equals the size of its congressional delegation and the House of Representatives is apportioned roughly in proportion to population, the weighting of votes in presidential elections is not as grossly unequal as in U.S. Senate elections, where a single voter in the least-populous state, Wyoming, has sixty-eight times the power of a voter in the most-populous state, California. But the inequality in the value of each vote in the Electoral College is still considerable. In the five states with the largest electorates, the number of residents of voting age per electoral vote ranges from 445,500 in Pennsylvania to 486,400 in Florida. Meanwhile, in the five least-populous states, the ratio of voting age population to electoral votes ranges from 124,300 in Wyoming to 160,700 in Vermont. Thus, in the Electoral College, a vote cast in a less populated state is worth three to four times as much as a vote cast in a highly populated state. Although the Supreme Court has held that the Constitution does not allow such extreme deviations from the "one person, one vote" standard in legislative apportionment, they are perfectly tolerated in the apportionment of the Electoral College.
The Electoral College induces a second kind of inequality just as troublesome as the first: inequality of attention. The political arithmetic of the Electoral College encourages presidential campaigns to focus their efforts on states with a close partisan balance, whether large or small in population. In recent years, Florida (twenty-seven electoral votes), Ohio (twenty), New Hampshire (four), and New Mexico (five) have been the object of much solicitous attention. Meanwhile, except for fundraising visits, campaigns have slighted California (fifty-five) and Texas (thirty-four). And it has been decades since presidential nominees waged a vigorous contest for Kansas (six) or Rhode Island (four). From March to October 2004, George W. Bush and John Kerry combined to make forty-nine visits to Ohio, forty-six to Florida, forty-one to Pennsylvania, thirty-four to Wisconsin, and twenty-six to Iowa-but the candidates visited California only eight times. In the same election, voters in twenty-eight states, including California, Texas, New York, Illinois, New Jersey, and Georgia, saw fewer than 100 campaign advertisements per month over the entire course of the campaign, while voters in the swing states viewed an order of magnitude more. Judging by the two campaigns' efforts, a majority of the 200 million citizens eligible to vote in 2004 were not seriously solicited for their votes.
The inequality of attention has two consequences for equality of voice. First, where campaigns concentrate their efforts influences who turns out to vote. Campaigns mobilize voters both as a direct strategy-through get-out-the-vote efforts-and as an indirect consequence of the excitement and commitment they generate. Indeed, the closely-fought 2004 campaign stimulated the largest increase in voter turnout in four decades, a gain of 5.8 percentage points over 2000. The increase, however, was concentrated in the battleground states. Every one of the ten states still "in play" in October-as judged by advertising buys-posted increases in voter participation that well exceeded the national average.
Second, where campaigns concentrate their efforts affects who sets the campaigns' issue agendas. As a first-term president, George W. Bush departed from Republican free-trade principles to impose duties on imported steel and softwood timber, actions of keen interest in Pennsylvania, West Virginia, Ohio, Washington, and Oregon. As a candidate in 2004, John Kerry found it necessary to establish his bona fides as a friend of gun owners by taking time off from the campaign trail to hunt pheasants in Iowa. The Republicans allegedly whipped up gay marriage referenda to mobilize religious conservatives in eleven states including Arkansas, Ohio, Michigan, and Oregon. To be sure, the parochial concerns of the swing states did not monopolize the two campaigns' agendas; the war in Iraq, health care, and terrorism each received thorough discussion. Still, the Electoral College pushed the special concerns of voters in a handful of highly competitive states ahead of the concerns of the vast majority of the citizens whose misfortune it was not to live in Ohio or Iowa.
In sum, the Electoral College violates the principle of equal voice that is fundamental to elections as an institution. It favors the preferences of some voters over others by as much as a factor of four, based only on the happenstance of the population of the state in which they reside. Even more dramatically, it distorts the process by which presidential candidates attend to the wishes of the voters. The Electoral College encourages-yea, demands-that candidates attach nearly exclusive importance to the preferences and involvement of voters living in closely divided states.
II. Proposals for Electoral College Reform
The consideration of equal voice is essential to evaluating the different proposals for Electoral College reform. Reforms can hardly be justified apart from their effects on equality, a paramount value in a democracy. Equality is particularly important in discussing the most likely prospect for reform: awarding two "senatorial" electoral votes to the overall winner of the state as well as one electoral vote to the popular vote winner within each of the state's congressional districts. By aligning the electoral vote more closely with the popular vote in each state, the proposal would appear to improve the prospects for equal voice. But the proposal contains less for political equality than meets the eye.
First, in a purely mechanical sense, the proposal to allocate electoral votes by congressional district does nothing to address the existing bias in favor of voters in less populated states. Every state still gets a minimum of three electoral votes, even tiny Wyoming, with a voting-eligible population about the size of Denver's.
Worse, by making little difference in the states with the smallest populations, the proposal may in fact aggravate the bias against voters in the states with large populations. It would not alter the winner-take-all method of allocation for the eight states with a single statewide congressional district and three electoral votes. Moreover, the partisan balances in smaller states are so lopsided that only six of the thirteen states with four to six electoral votes might actually split them.
But in the states with the largest populations, the proposal would have a greater effect-probably for the negative. The reform might bring competition to areas that are not now contested at the state level, as I will discuss momentarily. But in the more populous states that are currently competitive in presidential elections, the proposal would markedly reduce their value as a prize and diminish campaigns' willingness to invest their resources there. Were electoral votes allocated by congressional district, the one-sided partisan composition of most districts would cause Democrats and Republicans to compete for eight rather than twenty-seven electoral votes in Florida and for five rather than ten electoral votes in Wisconsin. The winner-take-all tradition of the Electoral College took hold precisely as a strategy to increase individual states' leverage in the presidential election process. Allocating electoral votes by congressional district would diminish the influence of highly populated states far more than less-populated states and in that way affect the more populous states-and equality of voice-for the worse.
On the other hand, one might assume that putting congressional districts into play would encourage presidential candidates to broaden the reach of their campaign efforts to include more people in more places, thus stimulating voter turnout and attending to more numerous and more diverse constituencies. Perhaps it would. But such an improvement in competitiveness-and in political equality-is hardly guaranteed.
Suppose, to make a back-of-the-envelope calculation, that we take as a measure of competitiveness a presidential election decided by a plurality of less than ten percent of votes cast. By this standard, twenty-one states with 264 total electoral votes were competitive in the 2004 election, and indeed, twenty-two states-mostly, but not entirely, the same ones-drew the attention of campaign advertisers at some point in the 2004 campaign.
Against this standard, what might have been the effect of a rule to award one electoral vote for each congressional district and two for each state in the 2004 election? George W. Bush and John Kerry ran within ten percentage points of each other in ninety-five of the 435 congressional districts. Sixty-three of the ninety-five competitive districts were in states that were already competitive according to the ten point standard, leaving just thirty-two districts in which campaigns might have invested resources they did not put into the state. Among the thirty-two, of course, were significant numbers of districts in states that drew absolutely no attention in the 2004 campaign, including nine (of twenty-nine) in New York, five (of nineteen) in Illinois, three (of five) in Connecticut, and three (of five) in Arizona. By this measure, the proposed reform would have increased the number of electoral votes in play by about twelve percent.
While the campaigns might step up their activities in the thirty-two newly competitive congressional districts under the proposed rule, they might equally reduce their efforts in the twenty-one states that had been competitive under the old rule. As already noted, the abolition of winner-take-all vastly diminishes the value of winning each state, from, for instance, 20 electoral votes to 2 for Ohio, or collectively, from 264 electoral votes in 21 competitive states to 42. Further, because the overwhelming majority of congressional districts are uncompetitive even in two-party competitive states, the proposed allocation also would drastically narrow the field in which campaigns need to make intense efforts to win votes. The campaigns might take the resources formerly devoted to intense efforts to win 264 electoral votes in 21 competitive states and turn them instead to concerted efforts to win 95 competitive congressional districts with real but diminished efforts to win 42 senatorial electoral votes in the 21 competitive states. By this reckoning, the proposed reform would have reduced contestation by as much as forty-eight percent in 2004. Thus, it is by no means obvious that this step is in the right direction, either for competition or for equality.
This analysis suggests that the equality central to democratic theory and democratic practice cannot be achieved by half measures. If we want every vote to count equally, the only solution is to elect the president by direct popular vote. The chief executive is the only officer of the federal government who is responsible to the citizenry at large, and we cannot ensure equal responsiveness if one citizen's vote counts for three or four times as much as another's. Likewise, if we want every citizen to have an equal chance to enjoy the ministrations of at least one of the presidential candidates, the only solution is to elect the president by direct popular vote. Only a national popular vote gives campaigns the incentive to seek support in places and among demographics they will probably lose-or win decisively. Were presidents elected by the nation at large, Democratic candidates would have a reason to campaign intensively among poor Latinos in south Texas, while Republicans would be motivated to appeal nationwide to Latinos who might be attracted by the party's stand on social issues; similarly, Republicans would have a reason to mobilize white-collar commuters in Connecticut, while Democrats would be motivated to rally the prairie populists who still dot the Great Plains.
In short, citizens have equal voice only when every vote counts. And every vote counts only when candidates must seek support from all voters, no matter the partisan predispositions of their neighbors and no matter where they live.