of Dr. Bailey as a Nixon elector also more pointedly disfranchised the Nixon voters. Despite party discipline, custom and state laws which tend to bind electors to the candidates to whom they are pledged, the Constitution grants them a discretion that they sometimes insist on exercising.

The Electoral College system violates democratic principles by making the votes of some voters count for more than the votes of other voters. While more voters go to the polls in the larger states, they are able to influence more electoral votes. Studies show that, on balance, the voters in larger states have a better chance of influencing the outcome of an election. Wallace's third-party candidacy raised the risk of an electoral deadlock in which no candidate would have commanded an electoral vote majority. The decision would then have fallen to the Congress, with the consequent risk of political deals and possibly serious delay in naming the nation's chief executive, who conceivably would not have been the one a plurality of voters wanted.

Dr. Bailey becomes one of only five electors in American history who have voted clearly contrary to the wishes of voters selecting them, but other electors have switched in slightly different circumstances. There have also been independent elector movements as well as third-party candidacies. In 1796 a Federalist elector switched to vote for Thomas Jefferson rather than John Adams, and history records a voter then as complaining in language appropriate for many North Carolinians now:

"I chuse him to act, not think."

But, from then until now, the Electoral College system has resisted basic change. Dr. Bailey remains free to ignore the wishes of the voters. His defection, coupled as it is with Wallace's third-party candidacy which could have created a constitutional crisis, should alert the nation. It should produce new efforts for fundamental electoral reform.

[From the New York Times, Nov. 16, 1968]

ELECTORAL-VOTE WEIGHT

To the Editor: Much of the opposition to reform of the Electoral College comes from representatives of the less populous states who believe that the present system benefits their constituents because they have more electoral votes per unit of population than residents of the larger states.

What they fail to understand is that the unit-vote rule, under which each state's electoral votes go to the plurality vote winner, reverses the small state's advantage because the voter in the large state may potentially affect a very large number of electoral votes.

Computer calculations show, for example, that a voter in New York has over three times the chance of affecting an election as a resident of the District of Columbia and more than twice the chance of residents of seven other states. Voters in 32 states and the District of Columbia have less than average voting power under the present system. Reform can come only when voters in the 32 smaller states convince their legislators that they are tired of playing a secondary role in the Presidential elections.

JOHN F. BANZHAF III,
George Washington University,
Washington, November 11, 1968.

[From the St. Louis Post-Dispatch, Sunday, Nov. 24, 1968]

LETTERS FROM THE PEOPLE

LARGE STATE DECIDE

Much of the opposition to reform of the Electoral College comes from representatives of the less populous states who believe that the present system benefits their constituents because they have more electoral votes per unit of population than residents of the large states.

What they fail to understand is that the unit-vote rule, under which each state's electoral votes go to the plurality vote winner, reverses the small state's advantage because the voter in the large state may potentially affect a very large number of electoral votes.

Computer calculations show, for example, that a voter in New York has over three times the chance of affecting an election as a resident of the District of Columbia and more than twice the chance of residents of seven other states. Voters in 32 states and the District of Columbia have less than average voting power under the present system.

Reform can come only when voters in the 32 smaller states convince their legislators that they are tired of playing a secondary role in presidential elections.

JOHN F. BANZHAF III,
George Washington University,
Washington, D.C.

ILLUSION OF VOTING POWER

Voters in the smaller states who fancy that the electoral college gives them an advantage in choosing a president would do well to ponder the point made in our Letters column by Prof. Banzhaf of George Washington University. The advantage is an illusion. Under the present system, it is voters in the larger states who hold predominant power.

With the aid of a computer, Prof. Banzhaf worked out a method of measuring the "relative voting power" of voters in the various states, based on the relationship of population to the electoral votes cast by each state. In addition to the more glaring disparities, the study brings out interesting facts about the relative voting power of citizens living, for example, in Missouri and Illinois.

A Missouri voter has 1.7 times as much voting power as a citizen in the District of Columbia, but an Illinois voter has 2.5 times as much. In other words, the Illinoisan wields 47 per cent more influence on the choice of a president than does the Missourian. Is there anything equitable about that?

Various devices have been suggested for reducing such disparities without abandoning the electoral college. These would work to the advantage of the smaller states. Thus if the "winner-take-all" system were abandoned, and electoral votes for each state were proportionally distributed in accordance with the votes cast, the Missouri voter would enjoy a 7 per cent advantage in voting power over the Illinoisan. If electors were chosen by congressional districts, a Missouri vote would weigh 14 per cent more than an Illinois vote.

But of course there is no more justice in giving the Missourian an advantage in voting power than there is in giving it to the Illinoisan across the river. Neither the present system nor the proposed modifications of it meet the requirements of equal justice. The only way to give every presidential vote exactly the same weight, no matter where it happens to be cast, is to abandon the electoral college altogether and elect our presidents by direct popular vote.

[From "Is Electoral Reform the Answer?" by Alexander M. Bickel, Chancellor Kent pro fessor of law and legal history at Yale University, in Commentary, December 1968; an expanded version to be printed as "The New Age of Political Reform," by Harper & Row]

THE ELECTORAL COLLEGE

THE "HUMPTY-DUMPTY electoral college," as one of its critics has called it, is another old instituion put to interesting new uses. Now the chief target of reformers, the electoral college was unquestionably invented to serve ends most of which we no longer care to serve, and most of which it no longer serves. Only in form does it remain what it was invented to be. Pursuant to Article II of the Constitution and the Twelfth Amendment, it still consists of as many electors from each state as the state has senators and representatives, and it still convenes quadrennially to elect a President and a Vice President of the United States. But although it was probably intended and clearly not forbidden to act independently, it has hardly ever done so, certainly not in modern times. Electors compete for the office in a popular election, but with very infrequent exceptions, which have never proved significant, they do so in complete anonymity; electors are pledged to Presidential and Vice-Presidential candidates for whom, if they win by a majority or plurality, they cast their states electoral vote.

These features of the system, unforeseen and unintended by its originators and sanctioned by custom rather than by the Constitution, have in modern times made the large, populous, heterogeneous states-where bloc voting, as by ethnic or racial

minorities or other interest groups, often determines the result the decisive influence in Presidential elections. Recently, Mr. John F. Banzhaf, III has analyzed the various possible arrangements of electoral votes, and the circumstances in which any given state could change the result of an election.* He has also calculated the chances of a voter affecting the outcome in his state, and the chances that the outcome of a national election would then itself be altered. His conclusion is that voters in "states like New York and California have over two and one-half times as much chance to affect the election of the President as residents of some of the smaller states.” Pennsylvania, Ohio, Michigan, Illinois, and even the lesser industrial states, are also in advantageous positions. The reason is that while a voter in a large state has less chance of influencing the result in his state (because there are, of course, more people voting) he potentially influences a larger number of electoral votes; and so, despite the apparent dilution of his vote, he actually exercises much greater control over the outcome of the national election. This power he derives directly from the electoral college system.*

What we have known to be true, then, is true. We can now establish mathematically why modern Presidents have been particularly sensitive to urban and minority interests-modern Presidents of both parties, that is to say. have been more responsive to urban interests than have other factions in their parties. And only men who can be thus responsive are generally nominated and elected. Goldwater in 1964 is something of an exception; he was nominated. Mr. Nixon in 1968 was no exception. If he made some unfamiliar sounds, that was because of the particular urban mood of the day.

Hon. EMANUEL CELLER,

Chairman, House Judiciary Committee,
U.S. House of Representatives,
Washington, D.C.

THE NATIONAL LAW CENTER,

March 4, 1969.

DEAR MR. CELLER: You will recall that when I appeared before your Committee on Wednesday morning, February 26th, certain inquiries were made to which I could not reply without further study, and I was requested to communicate this information to the Committee as soon as possible. I am very happy to be able to comply and respectfully request that this letter be placed in the record as a supplement to my testimony.

OUTCOME UNDER PROPORTIONAL SYSTEM

One of the questions asked was whether and to what extent the results of past elections would have been changed had a proportional system been in use instead of the Electoral College. Appendix "M" of Peirce, The People's President, 358-59, shows how electoral votes would have been allocated under a proportional plan in the elections of 1864-1964. Based upon this table, my own (non-computer) analysis indicates the following:

Under a proportional plan requiring the winner to receive 50% of the vote, the outcome would have been different in twelve elections. In 1876 and 1888 the proportional system would have elected the candidate receiving the greatest number of popular votes-something the Electoral College failed to do. In ten other cases, however, the proportional system would have failed to elect a President as no candidate would have received a majority of the electoral votes.

Under a proportional plan which requires only 40% of the total electoral vote for election, the proportional system would have produced the same President as the Electoral College in seven of these ten cases. In the remaining three cases1880, 1896, 1960-the proportional system with a 40% requirement would have produced a President different from that produced by the Electoral College and contrary to the weight of the popular vote. However, in weighing this information, one should consider the footnote to the table in Mr. Peirce's book:

*J. F. Banzhaf, III, One Man, 3.312 Votes: A Mathematical Analysis of the Electoral College," Villanova Law Review, Winter 1968, pp. 303-346.

"NOTE. Caution must be exercised in interpreting the possible outcome of alternative electoral count systems in past elections, since the campaigns might well have been conducted in a different manner as the Presidential candidates and their managers sought to exploit the differing types of electoral bases that would have been involved."

ANALYSIS IN THE ALABAMA LAWYER

During my testimony I was asked about an article in the Alabama Lawyer and how that analysis of the Electoral College differed from my own. The pertinent portion of the article, Thornton, An Analysis of Electoral College Reform, 29 Ala. Law. 398, 406 (1968), reads as follows:

"Alabama has approximately 2% of the voting strength in the Electoral College. New York has 8%, or four times Alabama's power in electing a President. But in 1964, the popular vote in Alabama was less than 9% of the total popular vote for President throughout the country. New York's popular vote was 10.2% of the vote cast throughout the country, or over ten times Alabama's vote. Alabama would hardly want to make that trade."

As I indicated in my testimony this analysis overlooks the rather complicated problem of how an individual voter's ability to affect his state's electoral votes varies with population and apparently makes the all-too-common and incorrect assumption that it varies inversely with population. This is not surprising because the author's biographical statement indicates no mathematical background. Furthermore, his analysis should be read and evaluated in conjunction with the remainder of the article which argues against spreading political power among illiterates, paupers, criminals, Communists, and the insane, and in passing manages to criticize many of the Supreme Court's latest decisions.

The main thrust of Mr. Thornton's analysis is that in a state with a far less than average voter turnout those who do vote enjoy a higher than normal voting power. This, of course, is true and would continue to be true under a proportional or district system but not under a direct election. In any case the effect does not seem to favor either the larger or smaller states because there is little correlation between voter turnout and number of electoral votes per state. For example, the ten states with the lowest percentage voter turnout in 1964 averaged 9.6 electoral votes while the ten with the highest percentage voter turnout in the same year averaged 8.4 electoral votes. Thus, if one were to include this factor in the analysis there would be greater disparities in voting power under the present system with the real winners those in the large states with a small turnout (e.g. Texas), and the real losers those in the smaller states with large turnouts (e.g. Utah). The disparity in favor of the voters in the larger states and against those in the smaller states would in general still remain and the precise figures would vary from election to election. In viewing these figures, it shoud also be asked whether a permanent solution can be based upon such a transitory phenomena and whether states with a small turnout should continue to be rewarded with additional voting power.

ALTERNATIVE CALCULATION OF AVERAGE VOTING POWER

One further issue which was raised was whether it would not have been better to use a weighted average rather than a simple average when talking about "average" voting power. In other words, rather than adding the relative voting power of the citizens of each state (Column 4 of my tables) and dividing by the number of such entries, whether it would have been better to have multiplied the average voting power for each state by the population of the state and divided by the total population to obtain an average.

First, it should be pointed out that the critical figures are those in column 4 columns 5 and 6 are attempts to further explain and describe the basic figures, There are a large number of possible statistical measures which could be used for this purpose and one is no more "valid" or "correct" than any other. Any such measure is useful to the extent that the user understands what it means and how it was derived.

In each of the three tables attached to my testimony the footnote describing the appropriate column states clearly: "percent by which voting power deviated

from the average of the figures in column 4." By way of further description footnote 29 in the article which formed the formal portion of my testimony states: "29. An equally 'valid' average might be obtained by multiplying the voting power by the number of voters in each state and dividing by the total number of voters. This would be a per voter average rather than the per state average which has been used in these tables. See Banzhaf, Multi-Member Electoral Districts-Do They Violate the 'One Man, One Vote' Principle, 75 Yale L.J, 1309, 1330n41 (1966)."

The reference in the footnote is to an earlier article on multi-member districts in which I faced the same problem and solved it differently. As shown below with the illustration for Arkansas, I carefully explained that one could calculate at least two different averages. I explained both of these averages, computed both, and indicated deviations from each. Nevertheless, the result was confusion and disbelief that a computer could produce two different figures for deviation. For this reason in my application of this same technique of analysis to the Electoral College I computed only the simple average and indicated in a footnote how a weighted average of equal validity could be computed.

The average effective representation, computed on a per district basis, is 142 percent compared with a single-member district. Approximately 73 percent of the districts are below this average in effective representation.

2 The average effective representation, computed on a per voter basis, is 184 percent compared with a single-member district. Approximately 65 percent of the voters are below this average in effective representation.

Since the meaning of these two averages had now been explained at some length, I am now happy to advise the Committee that with the cooperation of Mr. Ray Thomas of the George Washington University Computer Center I have been able to perform the additional calculations and to present figures based on two different averages. The results are as follows.

With respect to the analysis of each of the three systems-Electoral College, proportional, and district-one may compute two equally valid averages. One called the simple or per-state average is computed by adding the figures in column 4 of each table and dividing by 51. The second, called the weighted or per-voter average, is computed by multiplying each figure in column 4 by the state's population and dividing the total by the total population.

Under the present Electoral College system, the per-state average relative voting power is 1.680. 52,953,391 people in 32 states and the District of Columbia-over 29% of the total population-have less than the per-state average. Under the same system, the per-voter average relative voting power is 2.158 98,313,875 people in 44 states plus the District of Columbia-over 54% of the total population-have less than the per-voter average.

Under the proportional system the per-state average relative voting power would be 1.620. 170,080,480 people in 35 states and the District of Columbiaover 94% of the total population-would have less than the per-state average. Under this system the per-vote average relative voting power would be 1.179. 117,166,706 people in 15 states-over 65% of the total population-would have less than the per-voter average.

Under the district system the per-state average relative voting power would be 1.540. 168,014,360 people in 34 states-over 93% of the total populationwould have less than the per-state average. Under the same plan the per-voter average relative voting power would be 1.22. 110,714,686 people in 13 statesover 61% of the total population-would have less than the per-voter average.