On proportional allocation
Proportional division of electoral votes would not work in practice.
No amount of campaigning can realistically affect more than one electoral vote in most states under the proportional approach. A third of the states have only 2, 3, or 5 electoral votes, where one electoral vote would correspond to between 20% to 33% of the popular vote. Two-thirds of states have 11 or less electoral votes (where one electoral vote would correspond to 9% of the popular vote).
If you run through the states, it turns out that 2 electoral votes might be in play in California under the proportional approach, 1 vote might be in play in about 18 states, and 0 electoral votes could be realistically affected during the course of a real-world campaign in about 30 states. The number of battleground states would be about the same as now (about 18). These states would simply become winner-take-one states (except for California, which might be winner-take-two). Two-thirds of the states would remain mere spectators. It would be as pointless for a presidential candidate to give 30+ states any campaign attention, or policy attention.
On a practical note, any state that unilaterally adopted this system would be virtually eliminating whatever influence it might now have, as became obvious during the campaign in Colorado in November 2004, where voters rejected the proportional plan by 2-to-1.
What"s worse, even if 5 or 10 or 20 or 49 states adopted the proportional plan, it still would not work. If just one closely divided battleground state, such as Ohio or Florida (with 20 or 27 electoral votes) did not go along, then the entire presidential election would be controlled by the one hold-out state because all the other states would be winner-take-one or winner-take-nothing states. Even if a candidate won all 18 states where 1 electoral votes might be at stake in a proportional system, Ohio or Florida's 20 or 27 votes would be what would matter.
Finally, the proportional system would not accurately reflect the nationwide popular vote. In 2000, if every state used this system, Bush would have been elected.
The bottom line is that the proportional system would neither create a 50-state campaign where every vote matters and it would not accurately reflect the national popular vote of all the voters of all 50 states.
The simplest way to make every person"s vote matter equally in a presidential election is to run that election in the same way as every other election is run in the United States, namely, the candidate who gets the most votes wins the office.
First, proportional allocation will be done in all states only by constitutional amendment. As it is, it ONLY will be done in a state where the party passing policy has candidates that lose the presidential vote there -- e.g., it only will move when Tom Delay types are behind it because it's always going to be primarily backed by people trying to help their party.
Second, even if done in all states, it would leave many states in the position of political activity having no impact, as no degree of political activity would be likely to "swing" an electoral vote. Your state would have to be right near a tipping poingt.
Third, even if those states that are near a "tipping point," would never be able to tip more than one. So it completely neuters the principle of one-person, one-vote, as activity in a big state would never give you any more gain in electoral vote than activity in a small state.
Nope. Go with a national popular vote where the proportions are just what we want: one person, one vote.
On congressional district allocation
Allocating electoral votes by congressional districts would be far worse than the current system.
First, there are only 41 presidentially competitive congressional districts in the US. This is only 9% of the districts. That"s even less competition than produced by the existing statewide winner-take-all system. The congressional district system would concentrate a presidential campaign into an small accidental of districts. Under a current system, about a third of people live in the 16 closely divided battleground states.
Second, district system would not accurately reflect the nationwide popular vote. In 2000, if every state used this system, Bush would have been elected, even though Gore had 537,000 more popular votes nationally. In Michigan, Bush carried 10 districts in 2004, while Kerry won only 5, yet Kerry won the statewide popular vote very comfortably.
Third, every vote would not be equal. Montana (with 900,000 population) has one congressional district, while Wyoming (with about 350,000) has one.