elections. It can be stated without fear of successful rebuttal that this was never intended by the Framers of the Constitution. As written, the Constitution makes no provision for the unit rule under the general ticket system, but neither is there any prohibition against it. The practice has been made legal by general usage. Under this proposal, the electoral college would be abolished but the states would retain the electoral votes to which they are presently entitled. (This vote would be turned over automatically and en bloc to the winner in any given state, thereby perpetuating the weighted system of voting which obtains under the present system.

President Johnson endorsed this automatic plan in January 1965, pointing out the dangers of unpledged electors. Also, his proposal was directed at the possibility that the election might be thrown into the House of Representatives, where each state has a single vote. Since two of the other plans include similar provisions, we omit mention of this part of the amendment.

The most compelling argument against this automatic plan is that its adoption would serve as an obstacle to any meaningful reform, if one concedes that reform is needed. Meanwhile, the automatic vote is not consonant with the other provisions of the Constitution. Moreover, the bloc system of voting puts a premium on fraud because the juggling of a few votes can swing the electoral votes of the entire State.

THE POWER IN THE STATES SHOULD BE RETAINED

The amending power under the Constitution is deliberately slow and not to be taken lightly. As one studies the various proposals to amend the electoral provisions of the Constitution, one is struck by the fact that no serious effort has ever been made to make those provisions work. No constitutional amendment would be necessary were the states to abide by the exact provisions of the Constitution and of their own choice abolish the unit rule.

It cannot be stated too often that it is within the power of the states, without constitutional amendment, to institute that part of the district plan which would give every voter the equivalent of three votes for President. These three votes would correspond to his representation in Congress: One vote would be for an elector from his district; the remaining two votes would be for electors chosen at large in the state and correspond to his two Senators.

However unlikely this possibility may seem at the moment, perhaps their own self-interest will one day point the way, since any constitutional amendment regarding the electoral process carries with it the threat of further intrusion of the Federal Government in the rights of the states.

It would seem wiser to continue the States' discretion as to the manner in which electors are chosen. It should be noted, however, that this entire matter is complicated by recent decisions of the Supreme Court in regard to reapportionment of congressional districts and state legislatures. Nevertheless, when, the issues are finally resolved, it would appear that before surrendering further rights to the Federal Government the states should consider most carefully the restoration of their important role in the selection of a President on a fair and equitable basis.

Undoubtedly, there are some parts of the electoral college machinery that need repair and/or replacement, but there is nothing wrong with it beyond the power of the states to remedy. Since the Constitution already provides that each state shall appoint its electors in whatever manner the legislature thereof may direct, it is both possible and desirable for every state to preserve the political effectiveness of all of its areas and for all of its people by having its presidential electors chosen separately by the voters of each congressional district, with only two to be chosen at large as is presently the case with United States Senators. Such a practice would give every voter in the United States exactly three electoral votes and would end the weighted voting which now obtains.

This procedure would be in keeping with the intent of the Constitution. Moreover, as was once pointed out:

"The constitutional integrity of this country depends upon the constitutional integrity of its constituent states. One of the last bulwarks of defense for the vanishing rights of the states of the Union is now found in the constitutional provisions which lodge control of elections generally, and of presidential elections particularly, in the several states of the Union."

Electoral College a Bastion of Freedom, Manion Forum Network Broadcast, November 20, 1964.

No less important to the states and to the people is the fact that the Constitution provides the authority, if it does not actually command, the states to assume responsibility for the election of the leadership of the Union which they, themselves, established by ratification of the Constitution. Were they to assume this constitutional authority once more, it would be possible to hope that the balance of power between the Federal Government and the states might be restored. Unless the effort is made, there will be no halt to the trend to increasingly centralized Government and the corresponding relentless whittling away of the rights of the States and the people. Responsibility for the course to be taken lies in the hands of the American people. Let's keep the electoral college!

STATEMENT FOR THE RECORD OF THE HEARINGS OF THE SUBCOMMITTEE ON CONSTITUTIONAL AMENDMENTS OF THE U.S. SENATE

Distinguished Chairman and Members of the Committee: I appreciate this opportunity to make a statement.

Anyone who has dealt with the question of electoral reform is aware of the various proposals and the arguments put forth for them by worthy proponents. For my part, I am gratified that the election of the president might soon come to be done by direct popular vote. Such a system is the only one under which it is impossible to have a non-plurality president. It is the only system under which all the voters have equal power in determining the choice.

First, I would like to comment briefly on the inequities in power of the various voters to influence the result under other proposals. I was coauthor (with L S. Shapley) in 1962 of a paper entitled "Calculating The Electoral College Exactly" in which, I believe, it first came to be known that there was an inequity in the system (although small and mathematical) beyond those knowingly built-in by our founding fathers. In that paper, the results pertained only to the power of the separate states under the unit rule and not to that of the individual citizen-voter of those states. Since that time, studies have been made, using mathematical models, on relative individual power. I do not want to add here to an already long and well-argued debate, but I would like to say that the exact nature of the inequities, or where they lay, is not yet certain, in particular, whether they favor the voters of large states or small, or depend upon circumstances. For mathematical models are not a substitute for, but only an aid to, real-world analysis by other means.

It may suffice to say that the possibility of a non-plurality President is entirely equivalent to the existence of inequalities in voting power among independent voters-that as long as there are inequalities there is the possibility of a "miscarriage" of the popular will, and as long as there is the possibility of a "miscarriage" then there are inequities. Only by using a system of direct popular election is all doubt about inequities in the voting power of citizens dispelled. Under that system no belief could be held, true or erroneous, in the advantage of one particular section of voters over another.

I now address myself to an entirely different question, that of determining an appropriate "runoff" rule.

Rather than allow a pure plurality to carry the election in all cases, it seems desirable in some contingencies that there be a runoff between leading candidates. Everything possible should be done to impart a maximum aura of legitimacy to the presidency. It is, however, considered that runoff elections are hazardous to the political body, debilitating for the candidates, and costly to the country. and they should be resorted to only in exceptional cases. It is also reasoned that any strong showing by a third party can only be the result of drawing strength from both other leading sides, and hence that a runoff is less necessary than might otherwise be believed. Still, a plurality of, for instance, 33% or even less does not suggest a majority president. But between the extremes of a pure plurality and an absolute majority can be found many possibilities.

Some threshold is called for which will constitute a sufficient plurality. The threshold figure of 40% has been suggested. If such a number is to be used, the number 40% is historically reasonable. But the passage of time, evolving political knowledge on the part of the electorate, and evolving technology of elections, may modify what is considered reasonable in that respect. As an alternative to any absolute threshold, I would like to suggest a relative threshold that will work according to the closeness of the election in a certain prescribed way.

*Reprinted (in part) in Game Theory and Related Approaches to Social Behavior (M. Shubik, editor), Wiley and Sons, New York, 1964.

A reason for the desirability of a relative threshold is indicated by the following hypothetical election. Say that candidate A leads candidate B by only a few tenths of one percent, but that about ten percent of the vote is spread among other candidates. This hypothesis is within the bounds of realism. Each of A and B have around 45% of the vote. Yet there is little question that if the most popular candidate is to win, then the issue is still in doubt. The absence of a runoff, in the case where a third party may not be evenly divided among the leading candidates, would cast doubt on the legitimacy of the election. Only a runoff between A and B will determine who would be the popular choice. Yet if the threshold of a sufficient plurality is 40%, there would not be a runoff.

A more compelling reason yet for providing a runoff in such a contingency can be advanced. Lacking the provision for a runoff in the electoral system, a third party candidate trying to decide whether to run, might be tempted to try to siphon sufficient strength from one of the leading contenders to change the result of the election. With the presence of a relative runoff provision, this particular threat is no longer so tenable. It is perhaps a hallmark of a well-designed runoff rule that such tactical considerations would be mitigated.

The rule to be proposed is a simple one. If the difference in the number of votes between the two leading candidates is large enough so that, in a decision between just the two of them, it would require more than % of the remaining votes to provide the second of them with a majority, then the plurality of the leading candidate would be considered sufficient for election. To give a numerical example using %, if there are 2500 voters than a majority is 1251. If A has 1200 votes, B has 1000 votes, and C has 300 votes, then the plurality of A would be sufficent for election because more than 2% of the outstanding 300 votes would be required to give B a majority.

The rule can be more formally stated as follows:

Let A, B, and C, respectively, be the number of votes of the highest candidate, the number of votes of the next highest candidate, and the total number of votes of all the remaining candidates. Then, if the sum of B and % of the votes of C is not larger than the sum of A and 3 of the votes of C, the plurality of A would be sufficient.

For a sufficient plurality, if A+C is to be at least B+%C, it follows that a simpler, but equivalent, statement of the rule is:

A is sufficient if A is at least B+%C

C.

The C in the rule above is the result of substracting C from The plurality that would be sufficient to win the election is a variable which depends on the closeness of the election. Of course, as a consequence of the rule, a majority is always sufficient. But there is also a minimum plurality required in order to avoid a runoff. When the fraction % if used in this way, it can be shown that the minimum plurality necessary for election, which occurs when B and C are equal, is 40%. Since B and C are not in general equal, it is true that the sufficient plurality is generally required to be higher.

It would perhaps be most graphic to indicate what would constittue sufficient and what would constitute insufficient pluralities under such a rule. The following table shows a range of examples:

TABLE 1.-SUFFICIENT AND INSUFFICIENT PLURALITIES UNDER (2/3) RELATIVE THRESHOLD RULE

I The election of 1968 (Nixon 43.3 percent, Humphrey 42.9 percent) is covered a fortiori by this example.

(23)

(20)

(17)

(14)

(11)

The fraction % in the rule, of course, is arbitrary (just as is the 40% absolute threshold). But it does correspond to a time-honored proportion for strong consensus, and the currently considered 40% threshold may conceivably have been suggested implicitly by the minimum plurality to which gives rise.

Instead of %, any fraction between 1 and 1⁄2 could be used (not lower, because A is greater than B). Let any such fraction be r. Each value of r gives rise to a different rule, only one of which can be chosen. The rule would become: A is sufficient if A is at least B+(2r−1) C

When the fraction r is used, it can be shown that the minimum plurality 100r required is %. A range of feasible rules and their associated minimum 1+r

pluralities is shown in the following table:

TABLE 2.-DIFFERENT VALUES OF EACH OF WHICH WOULD

1 A runoff would be necessary if the plurality is below this minimum. It may be necessary otherwise.

Corresponding to Table 1 for r=%, two tables now follow which show a range of sufficient and insufficient pluralities for two other cases of r:

1 The election of 1968 (Nixon 43.3 percent, Humphrey 42.9 percent) is covered by this example.

(15)

(11)

1 The election of 1968 (Nixon, 43.3 percent; Humphrey, 42.9 percent) is covered by this example. Tables 1, 3, and 4, put together, show that the value of r= is likely the most reasonable and most corresponding to political intuition.

It remains only to say that the relative threshold rule is no more difficult in practice than the absolute threshold rule. When election returns are given, in the case of the absolute threshold rule it must be calculated whether A is at least 40 percent of the sum of A and B and C. Under the relative threshold rule, it

need only be calculated whether A is at least the sum of B and C. The relative threshold rule is equally workable, and if anything, it is simpler.

Under this proposed relative threshold rule, only one election in the last 100 years would have resulted in a runoff. It is true that that election is the one just past. But as has been suggested, under those or perhaps equally likely circumstances (had a few votes been changed), the runoff would have been justified.

STRIPPED OF MY RIGHT

Almost two hundred years ago, the government of the United States promised me the right to one vote in the presidential and vice-presidential elections of this country. But first, of course, there were qualifications demanded of me:

(1) That I be a U.S. citizen and

(2) That I be of a certain age limit, usually at least 21 years old.

I meet these specifications easily. I was born in the U.S., and citizenship here is my natural birthright. I also surpass the 21 year age limit by one year. However, an outmoded state law deprived me of casting my first vote in the 1968 elections. Why? It is very simple. In the summer of 1968, my husband and I graduated from Purdue University in Lafayette, Indiana. In September of the same year, he took the liberty to accept employment with a company near Nashville, Tennessee. Because we moved from one state in this "Union" to another less than six months before the elections, I was stripped of my right to vote.

Unfortunately, the problem which frustrated me is far from unique. In fact, millions of American citizens were denied this precious privilege in November because of such antiquated residency laws. After all, the United States is a mobile country. People are constantly choosing new jobs and being transferred to different places of residence. In the past year, more than 1% of all Americans between the ages of 21 and 29 moved to new locations; a large number of older citizens also moved. But despite these telltale statistics, the barrier still stands: 33 states require newcomers to live within their boundaries for at least one year before granting them what is already their right to vote in the presidential elections; Mississippi requires two years; and most of the remaining 16 states demand at least six months.' This is highway robbery. It is something which we should not expect, which we should in fact reject, in the country known as the greatest democracy in the world.

The alarming fact is that the problem is not just a newly-revealed one. In 1964, 70,600,000 Americans voted for the office of President. More than 1⁄2 of the eligible voters who failed to cast their ballots that year (a group numbering more than 20,000,000) were forcefully prohibited from the polls for no better reason than their inability to meet these arbitrary residence requirements. And in 1960, more than 8,000,000 eligible voters out of the 33,000,000 persons who moved during that year were unable to enter the voting booth. John F. Kennedy won in 1960 by the slim margin of less than one vote per election precinct. Since there are 166,137 precincts in the country, it is more than interesting to speculate on the difference that the 8,000,000 rejected voters could have made.3

This year, although exact numbers are not yet known, it was again a very tight race. If the laws had been changed between 1960 and 1968, another man could be entering the White House on January 20, 1969. And this difference could possibly be of prime importance to our country and to the world.

Of course, it is natural that a state would create a residency requirement for a few months for local and state elections, so that the newcomer may be familiar with the candidates before he votes. However, I have not asked to vote for these officials; I am only demanding my right of franchise in the presidential elections of the country which is mine, no matter what state I am living in.

How is the problem solved? There are several possibilities. The U.S. Senators and Representatives could drive an amendment through the Congress, stating that the residency requirement be discarded completely. Communities could continue registration until the last few days before the elections, leaving just enough time to examine the rolls for new registrants and other changes which might encourage fraudulous votes. An alternative to this solution utilizes the absentee ballot. In

1 Alfred Balk. "Our Election-Day Paper Curtain," Saturday Review, June 8, 1968, p. 34. 2 Balk, p. 34.

3 Brendan Byrne, Let's Modernize Our Horse-and-Buggy Election Laws, Grass Roots Guides on Democracy and Practical Politics, The Center for Information on America. Booklet No. 5, p. 3.