Allocating electoral votes by congressional district would increase (not decrease) the number of Americans who are ignored in presidential elections. District allocation would not accurately reflect the nationwide popular vote. District allocation would not make every vote equal. It would make a bad system even worse.

Under the congressional-district approach (as currently used in Maine and Nebraska), the voters elect two presidential electors statewide and one presidential elector for each of a state's congressional districts.

As to competitiveness, in the 2000 presidential election, there were only 55 congressional districts (out of 435 districts) in which the difference between George W. Bush and Al Gore was 4% or less in the district. Similarly, in 2004, there were only 42 congressional districts in which the difference between George W. Bush and John Kerry was 4% or less in the district. That is, only about a tenth of the population of the country lives in a congressional district that is competitive in presidential elections. In contrast, about a third of the country's population currently lives in a state that is competitive in presidential elections. One reason why the congressional-district approach is so much less competitive than the existing statewide winner-take-all approach is that congressional districts are often gerrymandered in favor of one party or the other in many states. Also, in many states, congressional districts are gerrymandered on a bipartisan basis so that most districts are strongly partisan (thus protecting incumbents of both parties). If electoral votes were allocated by congressional district, state legislatures would have even greater incentives than they now do to gerrymander districts.

As to accurately reflecting the nationwide popular vote, a second-place candidate could easily win the Presidency under the congressional-district approach. If the congressional-district approach had been applied to the results of the 2000 presidential election, then Bush would have received 288 electoral votes (53.3% of the total number of electoral votes), and Gore would have received 250 electoral votes (46.5% of the total). That is, the congressional-district approach would have given Bush a 6.8% lead in electoral votes over Gore in 2000. Nationwide, Gore received 50,992,335 popular votes (50.2% of the two-party popular vote), whereas Bush received 50,455,156 (49.7% of the two-party popular vote). Under the existing system, Bush received 271 electoral votes in 2000 (50.4% of the total number of electoral votes)—a 0.8% lead in electoral votes over Gore. In summary, the congressional-district approach would have been even less accurate than the existing statewide winner-take-all system in terms of reflecting the will of the voters.

In the 2004 presidential election, George W. Bush carried 255 (59%) of the 435 congressional districts, whereas John Kerry carried 180. Bush also carried 31 (61%) of the 51 jurisdictions entitled to appoint presidential electors. If the congressional-district approach had been in place nationwide for the 2004 presidential election, Bush would have won 317 (59%) of the 538 electoral votes in an election in which he received 51.5% of the two-party popular vote.

As to making every vote equal, there is a wide disparity in the number of votes cast in various congressional districts. For example, Wyoming (with a population of 453,588 in 1990) and Montana (with a population of 799,065 in 1990) each had one member in the House of Representatives (and hence three electoral votes). In many states, there is a three-to-one disparity in the number of votes cast in particular districts (due to factors including population changes since the last federal census and the level of turnout).

The congressional-district approach could be implemented in two ways. First, an individual state could decide to allocate its electoral votes in this manner (as Maine and Nebraska currently do). Second, a federal constitutional amendment could be adopted to implement the system on a nationwide basis. Of course, passing a constitutional amendment requires an enormous head of steam at the beginning of the process (i.e., getting a two-thirds vote in both houses of Congress). There have been only 17 amendments since ratification of the Bill of Rights. The last time that Congress successfully launched a federal constitutional amendment (voting by 18-year-olds) was in 1971.

There is a prohibitive, additional political impediment associated with the adoption of the congressional-district approach on a piecemeal basis by individual states. In 1800, Thomas Jefferson argued that Virginia should switch from its then-existing district system of electing presidential electors to the statewide winner-take-all system because of the political disadvantage suffered by states that divided their electoral votes by districts in a political environment in which other states use the winner-take-all approach:

Indeed, the now-prevailing statewide winner-take-all system became entrenched in the political landscape in the 1830s precisely because dividing a state's electoral votes diminishes the state's political influence relative to states employing the statewide winner-take-all approach.

The "folly" of individual states adopting the congressional-district approach on a piecemeal basis is shown by the fact that there were only 55 congressional districts in which the difference between George W. Bush and Al Gore was 4% or less in the 2000 presidential election. Suppose that as many as 48 or 49 states were to allocate electoral votes by district, but that one or two large, closely divided battleground states did not. The one or two state(s) retaining the winner-take-all system would immediately become the only state(s) that would matter in presidential politics. Thus, if states were to start adopting the congressional-district approach on a piecemeal basis, each state adopting the approach would increase the influence of the remaining winner-take-all states and thereby decrease the chance that the remaining states would adopt that approach. A state-by-state process of adopting the congressional-district approach would bring itself to a halt.

For more details, see sections 3.3 and 4.2 of this book.

A system in which electoral votes are divided proportionally by state would not accurately reflect the nationwide popular vote and would not make every vote equal.

Every vote would not be equal under the proportional approach. The proportional approach would disadvantage certain states in relation to other states. For example, Montana and Wyoming each have one congressman and three electoral votes. However, Montana has almost three times as many people as Wyoming. The proportional approach would disadvantage fast-growing states because electoral votes are only redistributed among the states after each federal census. The proportional approach would penalize states with high voter turnout (e.g., Oregon).

Moreover, the fractional proportional allocation approach does not ensure election of the winner of the nationwide popular vote. In 2000, for example, it would have resulted in the election of the second-place candidate.

The proportional approach could be implemented in two ways. First, an individual state could decide to allocate its electoral votes in this manner. For example, Colorado voters considered a ballot initiative to do this in 2004 (but rejected it by a 2-to-1 margin). Second, a federal constitutional amendment could be adopted to implement the system on a nationwide basis. There are significant differences between the two approaches.

If a federal constitutional amendment were adopted along the lines of proposals that have been introduced in Congress previously, the electoral votes of each state and the District of Columbia would be divided proportionally according to the percentage of votes (carried out to three decimal places) received in that state by each presidential slate. Such a system would not accurately reflect the nationwide popular vote and would not make every vote equal; however, it would make voters relevant in all 50 states and the District of Columbia.

If, on the other hand, individual states were to adopt the proportional system on a piecemeal basis, the electoral votes would necessarily be rounded off to the nearest whole number. A presidential elector is, after all, a person, and a person's vote cannot be divided into fractions. Absent a constitutional amendment, there is no way for an individual state to introduce fractional voting into the Electoral College. This rounding-off has counter-intuitive effects. In particular, there would be even fewer battleground states under this system than under the current system. This counter-intuitive result comes from the rounding-off to whole numbers. States have an average of only 11 electoral votes (and, in fact, two-thirds of the states have fewer than 11). Thus, one electoral vote would correspond to 9% of the popular vote in a state with 11 electoral votes. One electoral vote would correspond to 33% of the popular vote in a state with three electoral votes. Campaigning is rarely capable of shifting more than 8% of the vote during a typical presidential campaign. Thus, the only battleground states would be those where popular sentiment in the state fortuitously hovers right at the boundary where one electoral vote might be shifted. The vast majority of the states would not be poised anywhere near the boundary point. Thus, presidential campaigns would ignore them. In the states hovering right at the boundary, the only "battle" in these states would be for one electoral vote. That is, the proportional system would be, in effect, a "winner-take-one" system. Among the 50 states and the District of Columbia, California is the only jurisdiction where as many as two electoral votes might be in play under the proportional approach.

The whole-number proportional approach does not accurately reflect the nationwide popular vote and does not ensure election of the winner of the nationwide popular vote. If the whole-number proportional approach had been in use throughout the country in the 2000 presidential election, it would not have awarded the most electoral votes to the candidate receiving the most popular votes nationwide. Instead, the result would have been a tie of 269–269 in the electoral vote, even though Al Gore led by 537,179 popular votes across the nation.

There is a prohibitive, additional political impediment associated with the adoption of the proportional approach on a piecemeal basis by individual states. Any state that enacts the proportional approach on its own would reduce its own influence. This was the most telling argument that caused Colorado voters to agree with Republican Governor Owens and to reject this proposal in November 2004 by a two-to-one margin. This inherent defect cannot be remedied unless all 50 states simultaneously enact the proportional approach (as would be the case with a constitutional amendment). This inherent defect could not be remedied if, for example, 10, 20, 30, or even 40 states were to enact the proportional system on a piecemeal basis. Suppose that as many as 48 or 49 states were to allocate electoral votes proportionally, but that one or two large, closely divided battleground winner-take-all states did not. The one or two state(s) continuing to use the winner-take-all system would immediately become the only state(s) that would matter in presidential politics. Thus, if states were to start adopting the proportional approach on a piecemeal basis, each additional state adopting the approach would increase the influence of the remaining winner-take-all states and thereby would decrease the chance that the remaining states would adopt the approach. A state-by-state process of adopting the proportional approach would bring itself to a halt.

For more details, see chapter 4 of this book.