Should the Rules Used for Picking the President Be Altered?
An Analysis of Two Current Proposals

Second, there is the initiative in Colorado to allot that state's electoral college votes not on a "winner-take-all" basis (as the state currently does), but rather proportionally to the statewide popular vote.

Operation Arnold: The Move to Amend the "Natural Born" Clause

As John Dean discussed in his column last week, there is currently an effort to alter the Constitution to permit foreign-born U.S. citizens to become candidates for the Presidency. As Dean notes, the drive now underway may well be related to the popularity of both Republican California Governor Arnold Schwarzenegger, born in Austria, and Democrat Michigan Governor Jennifer Granholm, born in Canada.

Senator Orrin Hatch of Utah has introduced a specific proposal to change Article II of the Constitution to permit each naturalized citizen "who has been for 20 years a citizen of the United States" to be eligible for the White House. Meanwhile, other potential proposals would insist on 14 or 35 years of citizenship before Presidential eligibility.

The fate of these proposals is unclear, but to my mind they make a great deal of sense. Originally, the "natural born" qualification was inserted out of concerns about rich Dukes and Earls coming into America and taking over elections with their unbeatable riches. And, of course, such concerns simply do not apply with nearly the same force in today's world.

Should Amendments Like These Be Phased In, or Instantly Effective?

Some people who are skeptical about Senator Hatch's proposal (and even some who are supportive) have suggested that, at a minimum, any amendment of this kind enacted today should not become operative for a long period of time so that the amendment would not affect individuals like Schwarzenegger and Granholm - whose personal appeal may be putting the issue on today's agenda. They argue that such a "phase in" period would ensure that any enacted amendment stemmed from enduring values, not transitory personal charisma.

It is not clear to me, however, that a phase-in period is itself supported by deep constitutional values. Certainly, almost all of our twenty-seven constitutional amendments have taken immediate effect upon ratification. And even most of the amendments affecting the electoral process - for example, the amendments prohibiting race and gender discrimination in voting - have been implemented in the election directly following amendment ratification.

One notable exception to this trend is the Twenty-Second Amendment, which limits Presidents to two full terms. By its own specific words, the Twenty-Second Amendment does "not apply to any person holding the office of President when this [amendment] was proposed by Congress." Thus, Harry Truman, President in 1947 when the amendment was proposed, could have served more than two four-year stints.

But the reasons for delaying the effective date of the Twenty-Second Amendment - the desire not to insult or destroy the settled expectations of a known and specific existing incumbent - do not apply here.

Why the Twenty-Second Amendment Analogy Fails Here

Immediate implementation of a new amendment for naturalized citizens would surely not insult or stain anyone's reputation. Indeed, to the contrary, delaying the implementation would run some risk of insulting those like Schwarzenegger and Granholm.

Ordinarily, the Constitution prefers general rules that are not formally crafted either way around known or named individuals. Yet there are permissible exceptions from this generality principle in one direction: person-specific rules like the Twenty-Second Amendment, designed to benefit particular individuals, are very different than exceptions designed to single out particular individuals for inferior treatment.

Furthermore, with respect to the settled expectations of existing politicians, it is hard to see how anybody's reliance on the current exclusion of naturalized citizens from Presidential service should matter. It is possible that a few persons who may harbor Presidential aspirations might fear the additional competition created by an amendment permitting naturalized citizens to serve. But these fears, however rational they may or may not be, do not amount to the kind of reliance that we ought to constitutionally credit.

(In the same vein, the Seventeenth Amendment, which provided for direct election rather than legislative selection for U.S. Senators, did not interfere with the remainder of Senate terms that had already begun as of the date of ratification. But it did apply to all Senate candidates, including incumbent Senators, at the next regularly scheduled Senate election. Thus, the possibility that the incumbents might dislike the competition the new direct election system would create did not convince the ratifiers to delay its effectiveness.)

Of course, if supporters of the Hatch amendment feel they need to delay its implementation in order to increase its chances for passage, they could opt to do so. My point is simply that there is no separate, constitutional reason for a phase-in. Moreover, such a political calculation might well be wrong: The steam behind the amendment seems to derive from enthusiasm over people like Schwarzenegger and Granholm, so that a phase-in might actually decrease chances of passage.

Colorado's Amendment 36: Moving Away From "Winner Take All"

The second proposal I shall note today relates not to who can be President, but rather to how the presidential elections are decided. Colorado, this November, might become the third state to depart from the winner-take-all system for allocating electoral votes.

Almost every state currently awards all of its electoral college votes to whichever Presidential candidate receives the most popular votes in the state. Only two states - Nebraska and Maine -- hold open the possibility of divvying up the electoral vote pie between more than one candidate. Each of these two states awards two electoral college votes to the person who gets the most popular votes statewide, and then awards one electoral college vote for each congressional district in which a candidate receives a plurality of votes.

Now on Colorado's ballot is Amendment 36, which would divide Colorado's 9 electoral college votes among the Presidential candidates in direct proportion to the percentage of the popular vote they receive. Thus, for example, suppose candidate A wins the state 55%-45%. Then she would receive 55%, or 5, of the of state's 9 electoral votes, leaving 4 electors for Candidate B.

Opponents criticize Amendment 36 on many grounds. Most forceful is the idea that Colorado is minimizing its own clout in the national election by rewarding the candidate who appeals most to the state's voters and their needs with only a very small electoral vote victory (most likely 5-4 in today's tight election world.) Critics argue that unless and until all states are willing to allocate electoral votes proportionally, Colorado should not render itself electorally irrelevant vis-a-vis other states. To do so is unwise unilateral disarmament.

Could Colorado's Amendment 36 Constitutionally Affect November's Election?

Critics of Amendment 36 also point out that, if enacted, the measure is sure to be challenged in court. One possible basis for challenge is that the Colorado initiative, itself scheduled for a vote on November 2, 2004, by its terms applies to the 2004 Presidential election - which, of course, is occurring that very same day. Yet voters may not realize that when they cast their ballots. And this "retroactivity" question (the Amendment itself unwisely uses the legally loaded word "retroactively") may create wrinkles under both state and federal law.

More fundamentally, however, the initiative could be attacked as violating Article II of the U.S. Constitution, which provides that "each state shall appoint, in such Manner as the Legislature thereof may direct" a number of electors to the electoral college. (Emphasis added).

Three members of the U.S. Supreme Court in Bush v. Gore expressed the view that "Legislature" in this clause means the elected "Legislature" of the state -- not the courts of the state, and (perhaps) not the people of the state acting in their direct democracy lawmaking capacity.

It is not clear how those who believe Presidential selection power is vested in the distinct bodies known as state "Legislatures" would react to a popularly-enacted Colorado initiative measure.

A Comparison to Maine and Nebraska

It does seem likely that Colorado's moving away from the winner-take-all method could disadvantage it in the national competition for candidate attention. But Colorado's proposed straight proportionality approach may, at least, be preferable to the Maine and Nebraska approach - which makes use of a congressional district-by-district basis on which to award electors.

At first blush, allocating electors district by district would seem to be a move very analytically similar to - indeed, along the spectrum toward - proportionality. But mathematicians such as Michael Neubauer and Joel Zeitlin of California State University, Northridge have demonstrated that while this is true in large states, it's not true in small states. In addition, they have shown that a district-by-district approach, applied today, tends to disfavor the Democratic party.

In big states, the district-by-district method tends to move toward roughly proportional results. So, for example, in California in 2000, Bush lost the state but won 42% of the statewide popular vote. Under the district-by-district approach, he would have won nineteen of the state's fifty-four electors, because he got a plurality of voters in 19 of the state's congressional districts. Similarly, in Illinois in 2000, Bush lost the state but won 43% of the statewide popular vote. Under the district-by-district method, he would have won 41% of the state's electors because he won 41% of the congressional districts.

But what about small states? There, the district-by-district method, empirically, would not move much toward proportionality. As Neubauer and Zeitlin report, in the 10 smallest states, "only one elector would have been allocated differently" in 2000 had these states followed the Maine/Nebraska model. And because small states tend to be Republican these days, the end result of a national move toward a district-by-district approach overall would favor Republicans and hurt Democrats.

Indeed, if every state followed the Maine/Nebraska approach in 2000, Bush would have beaten Gore in the electoral college by a margin of 289 to 249, which much larger than the margin by which Bush actually won. This result seems counterintuitive, given that Gore - not Bush -- won the nationwide popular vote. While the move toward more equitable distribution within each state would seem analytically a step in the direction of a true nationwide popular election, the counting of results on a state-by-state basis creates numerical anomalies.

Would Colorado's proportionality approach, if followed in all states, operate similarly? Perhaps less so, because the potential manipulation of congressional district lines is taken out of the picture, but more empirical analysis would need to be done.

The Link Between Presidential Elections and House Districts More Generally

An even more arresting empirical observation is made by the mathematicians Neubauer and Zeitlin. They start by pointing out that the number of seats in the U.S. House of Representatives has been fixed by Congress since 1941 at 435. The size of the electoral college is equal to the size of the House plus the size of the Senate (plus, since the enactment of the Twenty-Third Amendment in 1961, the number of Senators and House members the District of Columbia would have were it a state). For this reason, the size of the electoral college has been fixed at 538 for over 40 years.

In 1941, when the current House size of 435 was reached, each House member represented about 300,000 citizens. Today, each House member represents almost twice that, or about 572,000 citizens.

Thus, if we wanted the same ratio of House members to people in the U.S. as existed in 1941, we would need today around 830 U.S. House members -- a large number to be sure, but perhaps no more unwieldy than 435.

But here's the startling point. Suppose we had kept expanding the size of the House to 830 - as Congress could have done, but has chosen not to do for six decades - and allocated those new seats to the states whose population would warrant them. In that hypothetical scenario, Gore would have won the electoral college vote in 2000 by a significant margin of 471 to 463 (the overall electoral college would have 934 votes - 830 to reflect the size of the House, 100 to reflect the size of the Senate, and 4 to reflect the size of DC.)

What's the "punch line" -- as Neubauer and Zeitlin put it? It's this: In a sense, "the winner of the 2000 Presidential election was determined in 1941 when the House size was fixed at 435. Had the House size been set at 500 in 1941 (and not been changed since) then Gore would have won the 2000 election!"

All of this leads me to the biggest sensible reform of the electoral college - and one that is surprisingly lacking in momentum: Scrap the entire thing and move to a true national election, in which the winner is the candidate for whom the largest absolute number of Americans vote. What could be simpler and more democratic?