c#\p \ 
 
 EFFECTIVE VOTING. 
 
 AN EXPLANATION OF THE BALLOT REFORMS USUALLY KNOWN AS “ PREFERENTIAL 
 VOTING ” AND “ PROPORTIONAL REPRESENTATION.” 
 
 By C. G. Hoag. 
 
 INTRODUCTION—THE TWO OBJECTS OF VOTING. 
 
 Many of the gravest evils of our politics in local, State, and Federal 
 Government are due in a considerable degree to our failure hitherto to 
 grasp the distinction between the two fundamentally different objects 
 : of voting and to adopt the right sort of ballot for carrying out each. 
 
 One of the objects to be carried out by voting is to make decisions 
 either between policies (measures) or between candidates for admin¬ 
 istrative positions. For that object majority voting is obviously 
 ! required by the fundamental principles of democracy. 
 
 The other object to be carried out by voting is to make up a deliber¬ 
 ative body fit to make decisions—subject or not to the operation of 
 the initiative and the referendum—on behalf of all the people. Now, 
 though the principles of democracy require that the decisions made 
 in such a body should be made by majority voting, they by no means 
 [require that the body itself should be made up by majority voting. 
 Iln making up such a body, as has been recognized universally— 
 though till recently very confusedly—each member’s right to a seat 
 should rest on his being the choice, not of a majority of all the voters 
 represented by the body, but merely of such part of them as we call 
 a ‘‘constituency.” 
 
 This fundamental distinction between the two objects of voting is 
 the secret of fundamental electoral reform. It reveals to us in what 
 cases rational reform means replacing a plurality system by a majority 
 system and in what cases it means replacing our present constituency 
 system*, under which a member of a deliberative body is elected by 
 a plurality of a constituency, not by a majority-of-a-constituency 
 system but by a unanimous-constituency or “proportional” system. 
 
 Plurality voting should be replaced by majority voting wherever, 
 as I have said, a decision is to be made. That means wherever 
 measures are voted on in deliberative bodies, wherever measures are 
 voted on directly at the polls, and wherever administrative officials 
 are selected or removed by deliberative bodies or are elected or 
 recalled at the polls . 1 
 
 Whenever the object of an election is to make up a body fit to make 
 decisions or choose administrative officers, our present system of 
 fvoting in fixed arbitrary constituencies should be replaced by a 
 katiopal constituency system in which neither pluralities nor majori¬ 
 ties have any place. 
 
 I i But see passages in Part II where it is suggested that such oflicials should not be elected at the polls 
 it all. 
 
 3 
 
4 
 
 EFFECTIVE VOTING. 
 
 Part I. 
 
 MAJORITY VOTING-COMMONLY CALLED “PREFERENTIAL VOTING.” 
 
 That the ballots should reveal the will of the majority except when 
 their object is the election of a decision-making body is obviously of 
 considerable importance. And in order that they may reveal that 
 will, they must give the voters the opportunity to express their will 
 so fully that the will of the majority can be learned from them no 
 matter how many the competing policies or candidates may be or 
 what the grouping of the voters at the particular election in question. 
 
 Ballots that allow such an adequate expression of the voter’s will 
 with this object in view have usually been called “preferential 
 ballots,” and the systems of applying them have usually been called 
 systems of “preferential voting.” Those names, however, are 
 ambiguous, for giving the voter an opportunity to express several 
 preferences on the ballot is also a feature of the Hare system of voting 
 to make up a representative body. I have thought it best, therefore, 
 to call such ballots majority ballots and such a system majority voting . 
 
 Electoral systems designed to reveal the will of the majority by a 
 single election are in use for final elections in Queensland, Victoria, 
 and Western Australia, Grand Junction (Colorado), Spokane (Wash¬ 
 ington), and elsewhere, and for primary elections in Wisconsin, 
 Minnesota, North Dakota, and possibly elsewhere. 
 
 In respect to these systems two criticisms must be made. The 
 first concerns their application: They are applied, in the places men¬ 
 tioned, not only to the election of administrative officials, where 
 majority voting is desirable, but to the election of deliberative bodies, 
 where majority voting, as I have said and as everybody who under¬ 
 stands proportional representation knows, is absolutely out of place. 
 The second criticism concerns the systems themselves: Every one of 
 them is defective—that is, every one of them may fail to reveal the 
 will of the majority truly. 
 
 Leaving the point raised by the first of these criticisms to be covered 
 later in this paper, I propose to discuss briefly now the defects of 
 these systems, to explain a system that lacks those defects, and to 
 consider the whole problem in its practical aspects. 
 
 The system used in Queensland and that used in Wisconsin and 
 Minnesota are different modifications of the more complete ahd less 
 defective Ware system. Without discussing the defects of the 
 modifications, therefore, which were injudiciously devised to simplify 
 the Ware system, let me explain the latter sytem itself, which is used 
 in Western Australia. 
 
 Under the Ware system the voter is allowed to indicate his prefer¬ 
 ence among the candidates—as many preferences as he pleases—by 
 putting the figure 1 opposite the name of his first choice, the figure 
 2 opposite the name of his second choice, and so on. “The first 1 
 count is only of the first choice votes. If no candidate has a majority, 
 the lowest candidate is excluded and Ms votes only are scrutmized 
 again and added to the votes of the other candidates as the prefer¬ 
 ences indicate. The candidates are thus successively excluded 
 until only two are left, of whom the higher will have a majority 
 
 1 The uncredited quotations in this paper are from a valuable paper on “Preferential Voting” in Equity 
 July 10, 1910, by William Hoag, of Boston, Mass., then Secretary of the American Proportional Repre¬ 
 sentation League. 
 
EFFECTIVE VOTING. 
 
 5 
 
 vote and be elected.” The defect of this Ware system, which is 
 shared by the more defective modified forms of it mentioned, con¬ 
 sists in the fact that a candidate who, according to the ballots as 
 marked, is really preferred by the majority to any other of the candi¬ 
 dates, taken singly, may possibly be dropped after one of the counts 
 because at the bottom of the poll then, when his real strength is 
 not completely revealed. 
 
 The system of majority voting used, in slightly different forms, 
 in Grand Junction, Spokane, and elsewhere is distinctly different 
 from those on the Ware principle. It offers the voter an opportunity 
 of indicating on the ballot his first choice (by a cross in a square 
 opposite the name), his second choice (by a cross in a second square 
 opposite the name), and his u other choices” (by crosses in a third 
 column of squares opposite the names). Following is a ballot 
 marked correctly by a voter whose first choice is D, whose second 
 choice is B, and whose u other choices” are A and G. 
 
 
 First 
 
 choice. 
 
 Second 
 
 choice. 
 
 Other 
 
 choices. 
 
 A 
 
 
 
 X 
 
 B 
 
 
 X 
 
 
 c 
 
 
 
 
 D 
 
 X 
 
 
 
 E 
 
 
 
 
 F 
 
 
 
 
 G 
 
 
 
 X 
 
 The rules for counting the votes under this system, as used in 
 Grand Junction, are substantially as follows: “A count is first taken 
 of first choices. If no candidate has a majority, the candidate 
 lowest on the first Count is excluded and a second count is taken of 
 all the ballots, counting first and second choices for the remaining 
 candidates. The highest candidate at this count is elected if he 
 has a majority of all the ballots. But if no candidate receives such 
 a majority, the candidate having the smallest number of first and 
 second choice votes is excluded and a third count is taken of first, 
 second,” and other choices. The candidate having the most votes 
 on this third count wins. 
 
 The rules of counting the ballots in Spokane differ from those in 
 Grand Junction in one respect: In Spokane, when the second or the 
 third count is made, there is no dropping of the candidate who stands 
 lowest on the preceding count. This is unquestionably an improve¬ 
 ment, for the dropping of the lowest man, though necessary in sys¬ 
 tems of the Ware type, is not necessary at all in this one. 
 
 Considering this system, then, only in the better form in which it 
 has been adopted, that of Spokane, we have to ask whether it is or 
 is not defective. Is it sure, in other words, to reveal the will of the 
 majority truly? Is it not. The reasons are to be found in two de- 
 
6 
 
 EFFECTIVE VOTING. 
 
 fects. The first may be described as that of letting a voter’s second 
 choice completely offset or neutralize his first choice, so far as the 
 relative standing of those two candidates on the second count is con¬ 
 cerned, as certainly happens at the second count under this system, 
 or of letting his third choice offset or neutralize his first and his sec¬ 
 ond, so far as the relative standing of those three candidates on the 
 third count is concerned, as certainly happens at the- third count 
 under this system, etc. The second defect is allowing u other choices ” 
 lower than third choices to be taken into account before it is found 
 out whether some candidate has not a majority when only first, 
 second, and third choices have been taken into account. 
 
 The first of these defects is inherent in this system. To get rid of 
 it, in other words, would mean abandoning the system altogether. 
 The question arises, then, how important, practically, this defect really 
 is. From the point of view in my explanation of it above it seems 
 gross. And there is no question but that it may, like the inherent 
 defect of the Ware system—the dropping of the lowest man—bring 
 about an erroneous result. But, on the other hand, if considered 
 from another point of view, it does not look quite so bad. Though 
 it is true that my second-choice mark on the ballot does com¬ 
 pletely neutralize my first-choice mark, so far as the relative standing 
 of those two candidates on the second count is concerned, it is true 
 also that the two marks help on that count both those candidates as 
 against all the rest. To mark a candidate as second choice under the 
 Spokane system, then, means helping the candidate thus marked in 
 case no candidate gets a majority on the first count, and in that case 
 only, equally with the candidate marked as first choice as against all 
 the rest on the ballot. It is therefore only in case the struggle for 
 the office turns out to be between the candidate of my first and the 
 candidate of my second choice, and is not settled until the second or 
 a subsequent count, that my marking the candidate of my second 
 choice will prove to have tended to thwart my will. The defect is 
 considerable, but its importance can easily be overestimated; it does 
 not prevent the Spokane system from being a vast improvement 
 on plurality voting for all elections whose object is the making of 
 decisions. 
 
 The second defect of the Spokane system can easily be eliminated, 
 and in my opinion it should be. The excuses for its presence in the 
 system are three: First, it is argued that for most voters the desig¬ 
 nating of the exact order of preferences below the second is impossible. 
 Secondly, it is argued that all that the voter needs, beyond the op¬ 
 portunity to record his first and his second preference, is the oppor¬ 
 tunity to vote against the candidates whom he regards as positively 
 objectionable, and that this last opportunity is given him by the 
 u other choice” column of squares. Thirdly, it is argued that it 
 would encumber the ballot to put on it more than three columns of 
 squares. 
 
 The third of those arguments can be met by saying simply that 
 the form of the ballot and the method of indicating preferences should 
 be the same under the Spokane system as under the Ware; in other 
 words, that there should be but one column of squares and that the 
 voter should be asked to record his preferences in that column by 
 the figures 1, 2, 3, etc. This method is supposed to be difficult for 
 the voters only by those who have never seen it in use. In fact it is 
 
EFFECTIVE VOTING. 
 
 7 
 
 not difficult at all, even to voters of little education. On this point 
 there is conclusive evidence. See the testimony of Earl Grey, 1 re¬ 
 cently Governor General of Canada, before the Royal Commission on 
 Systems of Election, and that of the Hon. John McCall, 2 agent general 
 of Tasmania (where the method mentioned is used in connection with 
 proportional representation) before the same commission. 
 
 The first and the second of the excuses reveal a failure to appreciate 
 the fundamental importance, in the long run, of giving the voter at 
 least the opportunity of expressing his will fully. Any restriction of 
 the voter’s freedom to express his will fully not only may but must 
 result in a transfer of political power from the voters, who should 
 have it all, to somebody else to whom chance or shrewdness delivers 
 it. Any such restriction, therefore, is a vitiation of democracy at 
 the fountain head. Moreover, the idea that there is sure to be a 
 longer gap, in the voter’s opinion, between his third choice and his 
 fourth—or between his sixth and his seventh, for that matter—than 
 between his first and his second is a mere assumption and in fact is 
 not true. The writer happened to be arguing this point about the 
 time of the national party conventions of 1912 with a man of Demo¬ 
 cratic and Progressive tendencies, who insisted that there was sure 
 to be a longer gap between a voter’s first choice and his second than 
 between two choices lower down in his scale. When, however, I 
 mentioned the names of the principal aspirants for the Presidency, 
 and asked him where, in his own list of preferences among those 
 aspirants, the big gap came, he admitted in an instant that it came 
 between the fourth one and the fifth. A system of voting based on 
 what voters 11 aren’t intelligent enough to do” or u probably don’t 
 want to do anyway” is not a system that will prove satisfactory in 
 the long run. What we must come to in the end is a system that 
 permits the voter to express his will as fully as he wants to. 
 
 To improve the Spokane system, then, provide only one column 
 of squares for the voter’s marks, and let him indicate his preferences— 
 as many or as few as he pleases—by the figures 1, 2, 3, etc., thus: 
 
 A. 
 
 3 
 
 B. 
 
 2 
 
 C. 
 
 
 D. 
 
 1 
 
 E. 
 
 
 F. 
 
 
 G. 
 
 4 
 
 
 
 Then take into account only first, second, and third choice votes on 
 the third count; only first, second, third, and fourth choice votes on 
 the fourth count; and so on. With those improvements the Spokane 
 system will be at its best, with only the one defect that I have men¬ 
 tioned as inherent in it. 
 
 The question now arises: Is there not some system of voting by 
 which the will of the majority can be determined by a single balloting 
 with absolute certainty? There is. It is known as Nanson’s method, 
 
 1 P. 183 of the British Government Blue Book, entitled Minutes of Evidence taken before the Royal 
 Commission on Systems of Election. The Blue Book is numbered “ed. 5352, ” London, 1910. Price, Is. 8d. 
 
 2 P. 188 of the same Blue Book. 
 
8 
 
 EFFECTIVE VOTING. 
 
 because invented by Prof. E. J. Nanson, of the University of Mel¬ 
 bourne, Australia, and it is explained and defended by him on pages 
 121-141 of the Blue Book of the British Government entitled “ Re¬ 
 ports from His Majesty’s Representatives in Foreign Countries and in 
 British Colonies Respecting the Application of the Principle of Pro¬ 
 portional Representation to Public Elections.” 1 With this method 
 is used the same preferential method of voting that is used with the 
 Ware method and that I have advocated for use with the Spokane 
 method. In the count a— 
 
 first choice is counted one, a second choice two, a third choice three, and so on, as 
 far as choices are indicated; all other candidates are given a number one greater than 
 the number given the last choice of the voter. The additions are then made and all 
 candidates excluded whose total is more than the average. Another scrutiny is then 
 taken of all the ballots to ascertain the order of preference of the remaining candidates, 
 another similar count made, and all above the average excluded again. This process 
 is continued until only one candidate remains, and he is elected. The successful 
 candidate will almost always be the choice of a majority of the voters over each of the 
 other candidates, and, if not, he will be the choice either over more candidates than 
 any other, or, if only the choice over an equal number, he will be preferred to those 
 others by a greater majority or judged inferior by a smaller. In short, all that can be 
 accomplished by a comparison two by two— 
 
 That is, a comparison of every candidate with every other taken 
 by himself— 
 
 can be accomplished by this abbreviated method of Nanson’s. The method may 
 therefore safely be regarded as a complete solution of the problem of ascertaining the 
 will of the majority. * * * The voting is simplicity itself, but the counting is 
 
 often more tedious than with Ware’s method because all the ballots have to be scru¬ 
 tinized at every count. The number of counts required is controlled by the number 
 of candidates. Four candidates might make two counts necessary, 8 candidates 
 three counts, 16 candidates four counts, 32 candidates five counts, and so on, one addi¬ 
 tional count being possibly necessary for every doubling of the number of candidates. 
 * * * It should be realized, however, that each count accomplishes all that could 
 
 be accomplished by holding another election; and it will not be questioned that 
 another count is far simpler and less expensive than another election. 
 
 We have now considered three methods of voting designed to reveal 
 the choice of the majority among several candidates (or measures) by 
 means of a single balloting. Let us now review comprehensively 
 the points of likeness and those of difference between them. 
 
 The three systems all differ from the nonpreferential voting hitherto 
 common in this country in that they give the voter the opportunity of 
 recording not merely one single desire—the desire to support a certain 
 candidate (or measure), but as much of his will as he cares to express in 
 respect to where his support is to go under the various circumstances that 
 may be found to have arisen when the ballots begin to be counted. In 
 other words, they all differ from the nonpreferential system in 
 that they give the voter not merely the privilege of casting a ballot, 
 but the privilege of making it effective. 
 
 The three systems are not necessarily different in respect to the 
 method by which the voter records his will on the ballot. For the 
 Ware and Nanson systems provide for the indication of the voter’s 
 preferences by means of the figures 1, 2, 3, etc.; and the Spokane sys¬ 
 tem, as I have said, could and should be changed so as to do so. 
 
 The only essential difference between the three systems, then, is 
 the rules for counting the ballots—that is, for deducing from them 
 what is to be regarded as the will of the majority. 
 
 1 London, 1907. Price, Is. 3d. This Blue Book is designated “Miscellaneous No. 3 (1907).” The inclu¬ 
 sion of such a paper in a hatch of reports on proportional representation is an example of the common con¬ 
 fusion of that system with the use of a preferential ballot for majority voting. The confusion was not Prof. 
 Nanson’s, of course, hut that of some Government official. 
 
EFFECTIVE VOTING. 
 
 9 
 
 I have already said that, so far as insuring a correct result is con¬ 
 cerned, the Ware and Spokane rules are fallible and the Nanson system 
 infallible. Let me now support that assertion by a concrete example, 
 taken from the Equity article by Mr. William Hoag referred to above. 
 
 Suppose A, B, C, D, and E are candidates for a single office that is 
 to be filled by majority vote, and suppose the ballots cast in the elec¬ 
 tion, 15 in number, are marked as indicated below, where each line 
 of figures running up and down the page represents the preferences 
 of one of the 15 voters as expressed on his ballot. 
 
 A. 222142 4521 21 
 
 B. 3 1454334331 34 
 
 C. 44 315113423421 
 
 D. 134223421154342 
 
 E. 513 31252 421 3 
 
 With these ballots cast, the Ware rules would elect E. With the 
 same ballots, the Spokane rules would elect A. (With the same bal¬ 
 lots, it may be noted in passing, the nonpreferential ballot still preva¬ 
 lent in this country would elect C.) But a careful comparison of the 
 ballots will show that the candidate who is unquestionably preferred 
 by a majority of the voters, as against any other of the candidates taken 
 singly, is D. From a two-by-two comparison it appears that— 
 
 D defeats A singly. 8 to 7 
 
 D defeats B singly.:. 10 to 5 
 
 D defeats C singly.. . 8 to 7 
 
 D defeats E singly. 8 to 7 
 
 Well, then, do the Nanson rules elect D? Let us carry them out 
 and see. 
 
 First scrutiny, Nanson rules. 
 
 A. 22214254521521 5= 43 
 
 B. 35145433433153 4= 51 
 
 C..... . 4 4 5 3 1 5 1 1 3 4 2 3 4 2 1= 43 
 
 D. 134223421 15434 2= 41 
 
 E 
 
 5 3= 47 
 
 225 
 
 As the average is 45, and B and E are higher than the average, they 
 are excluded and the ballots again examined. The preferences appear 
 as follows on the — 
 
 Second scrutiny, B and E excluded. 
 
 A. 21113133321311 3= 29 
 
 C. 13 3 3 3 1 3 1 1 2 3 2 1 3 2 1= 32 
 
 D. 122222221 13223 2= 29 
 
 90 
 
 The average is 30, and C is excluded because his total is above 30. 
 
 1 The figure 3 appears here in place of the figure 4, which the person who cast this ballot actually put 
 opposite ihe name of candidate C, because candidates B and E have now been eliminated and the figures 
 used in this second count under the Nanson rules must be those that indicate the preferences of the voters 
 among Ihe remaining candidates only. It will be found that this table and the following one are made up in 
 accordance with this principle. It will be noticed, also, that a count ing clerk could read off the figures of 
 this table from the preceding table as fast as another clerk could write the figures in this table down. To 
 read them off conveniently, the reading clerk would have only to draw a line through the figures, indicating 
 the preferences expressed by all the voters for candidates B and E. 
 
 The complete text of the Nanson rules is printed elsewhere in this issue of Equity. 
 
10 
 
 EFFECTIVE VOTING. 
 
 Third scrutiny , C excluded . 
 
 A. 21112122221211 2=23 
 
 D. 12221211112122 1=22 
 
 A is excluded and D elected. 
 
 Fortunately the Nanson rules can give nothing but the same cor¬ 
 rect result in any case, for they cover fully the mathematical law that 
 governs all cases possible. 
 
 So much for the correctness of the three methods of counting the 
 ballots: The Ware and Spokane methods are fallible and the Nanson 
 method infallible. We have now to compare the three methods in 
 another respect, their relative ease of operation in elections of various 
 sorts. 
 
 The Ware rules require the bringing together to the central electoral 
 authorities either of all the ballots themselves or of a full transcript 
 record of them in order that the particular ballots cast for a candidate 
 who is eliminated may be made effective, each for the candidate 
 marked on it as preferred by the voter among the candidates who can 
 be helped by his vote. The Nanson rules make the same requirement 
 in order that the Nanson provisions for the second and any necessary 
 subsequent counts may be carried out. The Spokane rules require 
 at the central electoral bureau neither the ballots themselves nor a 
 full transcript record of them; all they require there is a record of the 
 number of first-choice votes, the number of second-choice votes, the 
 number of third-choice votes, etc., that each candidate received at 
 each precinct. So far as this difference goes, therefore, the Ware and 
 the Nanson rules are less acceptable, especially in the case of elections 
 covering an extended territory such as an entire State, than the 
 Spokane rules. 
 
 The work of actually counting the votes is somewhat greater, on the 
 average, with the Nanson rules than with either the Ware or the 
 Spokane. I say this for what it is worth, hastening, however, to add 
 that the additional work required by the Nanson rules should, in my 
 opinion, be regarded as negligible in comparison with the infallibility 
 of the Nanson rules. On the average the work of counting under the 
 Ware rules and that under the amended Spokane rules seem to be 
 about the same. 
 
 In what cases then, if any, should each of the three sets of rules be 
 used ? On this point I can not pretend to speak for anyone besides 
 myself. 
 
 The Nanson rules, without question, should be used whenever 
 decisions are to be made in representative bodies or in assemblies of 
 any sort; for in such cases there is no need of trying to avoid the 
 trifling extra work required to make sure that the result will be cor¬ 
 rect. They should also be used whenever decisions are to be made 
 at the polls unless the opposition to the extra work they require— 
 which of course is slight in comparison with that involved in a second 
 election—is insurmountable. 
 
 The Ware or the Spokane rules—the latter amended of course— 
 should be used whenever decisions are to be made at the polls where 
 the opposition to the extra work required by the Nanson rules is 
 insurmountable. That opposition will naturally be strongest of 
 course in cases where the territory covered by the election is not com¬ 
 pact like that of a municipality, but extended like that of a State; and 
 in such cases the rules in favor of which the Nanson ndes will be 
 
EFFECTIVE VOTING. 
 
 11 
 
 rejected will be likely to be the Spokane rules, which do not require 
 the bringing together of the ballots for the completion of the count, 
 rather than the Ware rules, which do make that requirement. 
 
 In many cases, therefore, the contest, as soon as the Nanson system 
 is understood and appreciated, will probably be between that infallible 
 system, wdiich requires a little more work and the bringing together 
 of the ballots, and the (amended) Spokane system, which does not 
 require the bringing together of the ballots. The Ware rules, how¬ 
 ever, are worthy of support wherever the Nanson rules can not be 
 1 ted and the Ware rules happen to be preferred to the (amended) 
 
 :ane 
 
 My putting of the (amended) Spokane rules on a par with the Ware 
 thus is not in accordance, I know, with the views of some experts on 
 “preferential voting”—prominent members of the American Propor¬ 
 tional Representation League—who are at least as well qualified as 
 I am to express an opinion on the point. No course is open to me, 
 however, but to present the matter as I see it, and though the Ware 
 rules seem to me to have one advantage over the amended Spokane* 
 the latter have over the Ware two advantages which, taken together, 
 seem to me to balance the one on the other side. 
 
 The advantage of the Ware rules is that they are similar to those 
 for the counting of the ballots under the very best system of “propor¬ 
 tional representation,” as it is usually called; that is, of electing delib¬ 
 erative or decision-making bodies by unanimous constituency instead 
 of by mere plurality or majority of a constituency. This similarity 
 will be evident to the reader when he comes to the explanation in 
 Part II of this article of the Hare system of proportional representation, 
 for that is the best system of proportional representation, and the 
 rules of counting under it are similar in method, though different in 
 object, to those of the Ware system of majority voting. As we shall 
 doubtless eventually adopt the Hare plan of electing representatives 
 in deliberative bodies, the advantage of adopting a majority system 
 under which the procedure in counting the ballots is similar is obvi¬ 
 ously considerable. 
 
 Of the two advantages of the amended Spokane over the Ware 
 rules one has already been mentioned, the fact that neither the bal¬ 
 lots themselves nor any but the simplest records of them need to be 
 sent from the precincts to the central electoral authorities for the 
 completion of the count. This practical advantage is of real impor¬ 
 tance in State-wide elections, and it will be thought by many to be 
 of some importance in large cities. The other practical advantage of 
 the amended Spokane rules is their extreme plausibility on being 
 explained in the briefest way. 
 
 So far as liability to error in determining the will of the majority is 
 concerned, the improved Spokane system is about on a par, in my 
 opinion, with the Ware. Neither, as we have seen, is infallible; but, 
 on the other hand, neither is likely to result in any serious misinter¬ 
 pretation of the majority’s will. Under the Ware system it is surely 
 improbable that a candidate desired by a majority, as agamst any 
 other one candidate, will be dropped as the lowest of all at any stage 
 of the process of counting, and it is certain that any candidate elected 
 by the Ware system will have, if not an absolute'majority as against 
 every other one candidate, at least exceedingly strong support. 
 Under the amended Spokane system also it is highly improbable that 
 
12 
 
 EFFECTIVE VOTING. 
 
 a candidate desired by a majority, as against every other one candi¬ 
 date, will fail to receive a majority of first choices at the first count, 
 a majority of firsts and seconds at the second count, a majority of 
 firsts and seconds and thirds at the third count, and so on throughout 
 all the counts made, and it is certain that any candidate elected by 
 this system will have, if not an absolute majority as against every 
 other one candidate, at least very nearly that. 
 
 IMPORTANCE OF MAJORITY VOTING AND ITS RELATION TO PROPOR¬ 
 TIONAL REPRESENTATION. 
 
 The superiority of any one of these majority systems over our old 
 plurality system is great, for one of them indicates the will of the 
 majority infallibly by a single election, and the two others indicate 
 that will more accurately by a single election than the old plurality 
 system does by final election and primaries together. Any one of 
 them makes primaries unnecessary. Any one of them makes running 
 for office more attractive to men of the highest type. Any one of 
 them allows the voter to express his own real will freely without fear 
 of 11 throwing his vote away,” for if the candidate of his first choice 
 can not be elected his second or some lower preference will at once come 
 into play to make his ballot effective in helping the candidate he likes 
 best of those who can be helped by it. In this way any one of the sys¬ 
 tems conduces to freedom in expressing the will on the ballot and to 
 justice in securing to that will, as expressed, its due weight in the 
 making of the decision at stake. In short, for the making of decisions 
 by the ballot, any one of the three systems may fairly be called, in 
 contradistinction to our old system of plurality voting in two elections, 
 democracy in voting instead of its mere semblance, effective voting 
 instead of the mere casting ol ballots that may count for nothing 
 toward deciding the issue according to the voters’ real will. 
 
 But whenever the object of voting—to come back to a point I 
 made at the outset—is not to make a decision at all but to make up 
 a body fit to make decisions in the name of all the voters—in other 
 words, a representative or deliberative body—then majority voting, 
 even by an infallible system, is only one degree less absurd and less 
 disastrous to democracy than plurality voting itself. Indeed, the 
 advantage of using a good preferential majority system instead of the 
 old system (of the unchangeable vote, the two elections, and the plu¬ 
 rality rule) is trifling in comparison with the disadvantage of using 
 even the best of majority system when the object of the voting is the 
 election of a representative body. For when that is the object, it is 
 almost sure to be defeated, and the principles of democracy grossly 
 violated, unless a system of voting be adopted by which no candidate 
 can be elected to the body unless he is supported not merely by a 
 majority but by approximately the whole of a constituency. And 
 that brings us to the second part of this article. 
 
 Part II. 
 
 UNANIMOUS-CONSTITUENCY VOTING-“ PROPORTIONAL,” OR TRUE, 
 
 REPRESENTATION. 
 
 The misrepresentation of the people in their legislatures is unneces¬ 
 sary. It is due to our custom, which will seem rediculous to our 
 grandchildren, of defining the constituency of each member by an 
 
EFFECTIVE VOTING. 
 
 13 
 
 arbitrary geographical bne and then allowing a plurality of the voters 
 within the designated area to elect the “representative.” Such a 
 method virtually disfranchises not only all who vote against the can¬ 
 didate elected, who frequently comprise from 40 to 60 per cent of the 
 whole electorate, but also—to a less degree—those who voted for 
 that candidate m the final election not because they preferred him 
 to all others but only because they disliked him less than anv of the 
 others who had come through the primaries as officially recognized 
 candidates. Thus our present system makes it certain that most of 
 the ballots will be thrown away if marked for the voter’s first choice, 
 therefore discouraging the expression of his real will on the ballot, 
 and it makes it probable that a large percentage of the ballots, even 
 as marked, will be utterly ineffective. It is perfectly feasible, how¬ 
 ever, by changing the nature of the constituency from a unit according 
 to geography to a unit according to will, and by allowing the voter 
 to express his will on the ballot adequately instead of inadequately— 
 that is, by expressing his second, third, and other preferences as well 
 as his first, or by voting for a list of candidates instead of for only one—* 
 to make nearly every ballot effective in the make-up of the repre¬ 
 sentative body. To be exact, at the municipal election of 1909 in 
 Cape Town, South Africa, where our present system is still used, only 
 42 per cent of the votes cast were effective in electing members of the 
 council, whereas at the elections of the same year in Johannesburg 
 and Pretoria, where the Hare system of proportional representation 
 was used, 96 per cent and 99 per cent, respectively, of the ballots cast 
 were effective in electing members. 
 
 So much for the relative justice of the two systems from the indi¬ 
 vidual voter’s point of view. Now consider it from the point of view 
 of the several political parties. In New Jersey in 1912 a congressional 
 delegation 1 was chosen from the 12 districts of New Jersey in the 
 proportion of 1 Republican to 11 Democrats. In round figures 
 the situation was: Total vote, 335,000; Democratic, 168,000; Repub¬ 
 lican, 96,000; Progressive, 71,000. Percentage of Democratic vote^ 
 50; of Republican, 29; of Progressive, 21. Voters virtually dis¬ 
 franchised, that is, those whose ballots did not have the slightest 
 effect on the make-up of the delegation—Democrats, 13,000; Repub¬ 
 licans, 82,000; Progressives, 71,000; total virtually disfranchised, 
 nearly 50 per cent of the whole number. According to the ballots, 
 as actually marked, there should have been elected 6 Democrats, 3 
 Republicans, and 3 Progressives; according to the rules of our present 
 electoral system, there were declared elected 11 Democrats, 1 Repub¬ 
 lican, and no Progressive. How much the result was further vitiated 
 by the fact that many of the voters did not dare to mark their ballots 
 for their real first choice for Congress, for fear of throwing away their 
 votes, nobody knows. 
 
 In the momentous British general elections of January, 1910— 
 
 there were 144 constituencies in which the successful member was returned by a 
 majority of less than 500. Of these constituencies 69 seats were held by the Ministeri¬ 
 alists [the Liberals, Irish Nationalists, and Labor Party men] and 75 by the Unionists 
 [Conservatives]. The majorities were in some cases as low as 8, 10, and 14. The 
 aggregate of the majorities in the Ministerialist constituencies amounted to 16,931, 
 and had some 8,500 Liberals in these constituencies changed sides, the Ministerialist 
 majority of 124 might have been annihilated. On the other hand, the Unionists held 
 
 1 On this point I select my examples from national instead of from State elections because national issues 
 are more generally understood. This example is taken from an article entitled “ Shall the people rule? ” 
 by John E. Eastmond in the Trend Magazine for February, 1913. 
 
14 
 
 EFFECTIVE VOTING. 
 
 75 seats by an aggregate majority of 17,389, and had fortune favored the Ministerialists 
 in these constituencies their majority would have been no less than 274. Such is the 
 stability of the foundation on which the House of Commons rests, such the method 
 to which we trust when it is necessary to consult the nation on grave national issues. 1 
 
 The number of Socialist votes for Congressmen in 1908, relatively 
 to the number cast for Congressmen by other parties, entitled the 
 party to several members, but it elected only one. In 1912 the party, 
 though it doubled its vote, did not elect a single Congressman. 
 
 If such virtual disfranchisement of a large percentage of the voters 
 were necessary, it would be foolish to spend good printer’s ink pro¬ 
 testing against it; but it is quite unnecessary. There is no reason 
 why nearly every voter in New Jersey should not have in the legisla¬ 
 ture a member for whom he voted and for whom he wanted to vote. 
 And the legislature need not have many members either. All that 
 is necessary is to get rid of the arbitrarily defined geographical con¬ 
 stituency (for each member)—half or two-thirds of the voters in 
 which, perhaps, vote against the member sent up by it and are there¬ 
 fore certainly misrepresented—a constituency defined as a group of 
 voters united in wanting to be represented by a certain person and 
 numerous enough to deserve a representative. 
 
 Before considering the means by which such a unanimous-con¬ 
 stituency system of representation is carried out, consider the results 
 of two of the recent parliamentary elections in Tasmania, where such 
 a system—the Hare—has been in use for several years. 
 
 1912. 
 
 
 Votes. 
 
 Seats 
 
 actually 
 
 obtained. 
 
 Seats in 
 proportion 
 to votes. 
 
 Labor. 
 
 33,634 
 40,252 
 
 14 
 
 13.66 
 
 Nonlabor. 
 
 16 
 
 16.34 
 
 
 1913. 
 
 
 Votes. 
 
 Seats 
 
 actually 
 
 obtained. 
 
 Seats in 
 proportion 
 to votes. 
 
 Labor. 
 
 30,896 
 34,676 
 
 14 
 
 14.14 
 
 Nonlabor. 
 
 16 
 
 15.86 
 
 
 Surely John Stuart Mill’s eloquent tribute to the Hare system in 
 his Autobiography is not extravagant. He writes: 
 
 This great discovery, for it is no less, in the political art, inspired me * * * with 
 new and more sanguine hopes respecting the future of human society—by freeing the 
 form of political institutions toward which the whole civilized world is manifestly 
 and irresistibly tending (i. e., democracy) from the chief part of what seemed to 
 qualify, or render doubtful, its ultimate benefits. * * * I can understand that 
 
 g ersons, otherwise intelligent, should, for want of sufficient examination, be repelled 
 om Mr. Hare’s plan by what they think the complex nature of its machinery. But 
 anyone who does not feel the want which the scheme is intended to supply; anyone 
 who throws it over as a mere intellectual subtlety or crochet, tending to no valuable 
 purpose, and unworthy the attention of practical men, may be pronounced an incom¬ 
 petent statesman, unequal to the politics of the future. 
 
 Take a specific case. Suppose the legislature were to be composed 
 of 25 members in a single chamber. (A second chamber is useless, 
 
 1 J. H. Humphreys: Proportional Representation, Methuen & Co., London, 1911, p. 27. 
 
EFFECTIVE VOTING. 
 
 15 
 
 obviously, if the first one is made truly representative of the entire 
 electorate.) 
 
 In such a case, under our present system, we should divide the 
 State into 25 districts, in each of which one member would be elected 
 by a plurality of the voters. By adopting one of the majority prefer¬ 
 ential systems explained in the first part of this article we might, of 
 course, insure the election of each member by a majority of the voters 
 of his district instead of by a mere plurality. But that would not 
 help matters much; for even though every member were elected by 
 a majority of the voters of his district, the number of the unrepre¬ 
 sented or misrepresented in the whole State might still be very great—• 
 any number, indeed, up to half of all the voters. 
 
 But suppose we abandon the single-member geographical con¬ 
 stituency altogether, substituting for it a constituency of approxi¬ 
 mately the same number of voters who are unanimous in desiring a 
 certain person as representative. Then we shall be able to reflect truly 
 in the legislature the interests and opinions, in other words the will 
 of all the voters of the State. First throw the geographical districts 
 together into one or more large districts. You mat poll the whole 
 State as one district, electing all 25 members at larA on one ticket, 
 or you may divide the State into two or three district, each of which 
 would elect as many members as its population entitled it to. For 
 the purposes of explanation I shall suppose that we choose to divide 
 the State into three large districts, of which one, let us say that in 
 which Newark is situated, is to elect nine members, the others eight 
 each. Now, how is each of the districts, for example that of Newark, 
 to elect its members ? 
 
 There are several types of proportional or unanimous-constituency 
 representation. All are based on the fundamental principle that if 
 a large district is to send a certain number of members, say nine, to the 
 legislature, each of the nine largest groups of voters who can unite on a 
 candidate should be allowed to send one. But this fundamental 
 principle can be applied in various ways. 
 
 One way is to provide that the candidates be nominated in lists 
 by the several parties, any group of voters who wish to act together 
 for this purpose being considered a “ party’ ’ for the particular election 
 in question. (If party names are allowed on the ballots, some pro¬ 
 vision must be made, of course, by primaries or otherwise, for deter¬ 
 mining which group of petitioners is entitled to use any party name 
 which two or more such groups may claim. If party names are not 
 allowed on the ballots, this difficulty is altogether avoided, yet with¬ 
 out in the least preventing any party or faction, old or new, from 
 electing as many members as its voting strength entitled it to.) 
 This is commonly called the “list system.” It is used for the election 
 of parliamentary chambers in Belgium, Finland, Switzerland (the 
 Councils of about half the Cantons), Sweden, and Wiirtemberg. 
 
 Following are provisions suitable in a general way for the election 
 of the group of nine members to be sent to the legislature by the 
 Newark district that we have taken as an example: 
 
 List System of Proportional Representation. 
 
 NOMINATION AND ELECTION PROVISIONS. 
 
 Section 1 . Nomination of candidates for the legislature in each district shall be by 
 petition, signed by electors, who have signed no other petition to nominate any candi¬ 
 date for the legislature at the same election, to the number of . The petitions shall 
 
16 
 
 EFFECTIVE VOTING. 
 
 include the domicile addresses of the candidates. Each such petition may nominate 
 as many persons not nominated by any other petition, as there are seats to be filled 
 from the district. Each petition shall be signed, filed, and verified in the manner 
 prescribed by general law, shall contain the signed consent of each candidate, and 
 shall be filed with the election authorities at least 20 days prior to the election. 
 
 Sec. 2. The several lists of candidates—all the persons nominated by one group of 
 petitioners being considered one list—shall appear on the ballots, without party names 
 or emblems, in the order in which they were filed with the election authorities. The 
 names and domicile addresses of the candidates on each list shall be printed on the 
 ballot in the alphabetical order of the surnames. 
 
 Sec. 3. The votes shall be taken on ballots used for the filling of no other offices. 
 The form of the ballot shall be substantially as shown in table on following page. 
 
 Sec. 4. The total number of valid ballots for representative in the legislature cast 
 in the district shall be divided by the number of representatives to be elected from 
 the district, and the quotient shall be the “quota.” 
 
 Sec. 5. The number of full quotas contained in the total number of valid ballots 
 cast in the district for the candidates of a list shall be the number of candidates on said 
 list to be declared elected on the first assignment of seats. 
 
 Sec. 6. After this first assignment of seats, the remaining seats, if any, shall be 
 assigned as follows: The number of valid ballots for the candidates of each list shall 
 be divided by the number of seats, plus one, already assigned to such list, and to 
 the list showing the largest quotient shall be assigned one additional seat. This 
 performance shall be repeated until the number of seats assigned shall be that to 
 which the district SlTentitled. 1 
 
 Sec. 7. The candidates to be declared elected from any list shall be those indi¬ 
 vidually receiving we largest number of valid votes. 
 
 . [Form of ballot.] 
 
 FOR MEMBERS OF THE LEGISLATURE. 
 
 Directions to Voters: 
 
 Mark a cross (x) opposite the name of one candidate only for whom you want to vote. (If the 
 candidate you vote for is found to be elected without your vote, or if he is found to have too few 
 votes to be elected with it, your vote will be counted for some other candidate on the same list.) 
 
 A ballot is spoilt if more than one name is marked. If you spoil this ballot, tear it across once, 
 return it to the election officer in charge of the ballots, and get another from him. 
 
 CANDIDATES FOR THE LEGISLATURE. 
 
 * List 1. List 2. List 3. 
 
 
 [Domicile address.] 
 
 
 
 B. 
 
 
 
 a u 
 
 o . 
 
 
 
 a u 
 
 P.... 
 
 
 
 s. 
 
 
 
 T. 
 
 
 
 V. 
 
 
 
 Y. 
 
 •> 
 
 [Domicile address.] 
 
 D. 
 
 
 
 it (( 
 
 L. 
 
 
 
 M. 
 
 
 
 U U 
 
 N. 
 
 
 
 u u 
 
 Q. 
 
 
 
 u u 
 
 u. 
 
 Sec. 8. A ballot marked for more than one candidate shall be set aside as invalid; 
 but no ballot on which the voter has clearly voted for one and only one candidate 
 shall be set aside as invalid because the mark is in a form other than that of a cross 
 (X) or for any other merely technical reason. 
 
 1 The reasons for these provisions are explained below in connection with a concrete illustration. The 
 provisions are essentially the same as those governing the assignment of seats to lists in the system of 
 proportional representation used for parliamentary elections in Belgium. 
 
 * The squares for the voters’ cross would be at the right of the names, of course, in States in which cus¬ 
 tom would require such a change. 
 
EFFECTIVE VOTING. 
 
 17 
 
 Sec. 9. To any vacancy that shall occur, otherwise than by the operation of the 
 
 recall, in the delegation of representatives to the legislature from a district, the - 
 
 fhere name the proper election authorities] shall appoint, to fill out the unexpired 
 term, that candidate from the list on which the vacating representative was nominated 
 who of all the unelected candidates on that list received most valid votes. 
 
 Consider liow these rules would work out in a concrete case. Sup¬ 
 pose the Newark district, which is to elect nine of the twenty-five 
 members of the legislature, polls valid votes as follows: 
 
 List 1. 
 
 Candidate A. 20, 000 ^ 
 
 Candidate B. 30, 000 ^ 
 
 Candidate O. 2, 000 
 
 * Candidate P. 26, 000 *<•''' 
 
 Candidate S. 600 
 
 Candidate T. 20 
 
 Candidate V. 15, 000 * 
 
 ^ Candidate Y.,. 5,100 
 
 98, 720 
 
 List 2. 
 
 Candidate C. 5,000 
 
 Candidate G. 2, 000 
 
 Candidate H. 3,450 
 
 Candidate J. 20,110 IS 
 
 Candidate K. 12,001 
 
 42, 561 
 
 List 3. 
 
 Candidate D. 10,000 
 
 Candidate L. 5, 000 
 
 Candidate M. 11, 002 ^ 
 
 Candidate N. 20, 425 ^ 
 
 Candidate Q. 30,204 
 
 Candidate U. 1,104 
 
 77, 735 
 
 The ballots being marked as shown by these figures, the total vote 
 of the district would be 219,016. The quota would be that number 
 divided by 9, which is 24,335^. 
 
 As that last number is fully contained in the vote of the first list 
 four times, that list is given four seats on the first assignment. The 
 four candidates on the first list to receive these seats are (1) Candi¬ 
 date B, (2) Candidate P, (3) Candidate A, and (4) Candidate V. 
 
 As the quota is fully contained once in the vote of the second list, 
 that list is given one seat on the first assignment, the candidate 
 p receiving it being Candidate J. 
 
 As the quota is fully contained three times in the vote of the third 
 list, that list is given three seats on the first assignment, the candi¬ 
 dates receiving them being Candidates Q, N, and M. 
 
 Applying now the rule of section 6 above, we find that dividing 
 98,720 by 5 gives us 19,744; dividing 42,561 by 2 gives us 21,285^; 
 and dividing 77,735 by 4 gives us 19,433f. We therefore assign the 
 ninth seat to the second list, the candidate receiving it being Candi¬ 
 date K. 
 
 The method of determining to which lists the seats not filled by 
 the first assignment shall be given may seem arbitrary, but in fact 
 it is not. Of course, if the voters happened to divide into the three 
 S. Doc. 142, 63-1 - 2 
 
18 
 
 EFFECTIVE VOTING. 
 
 groups on the lines of full ninths, that is, if the first list received 
 exactly four-ninths of the votes, the second list exactly two-ninths, 
 and the third list exactly three-ninths, the first assignment would 
 fill all the seats, and the constituency filling each one would be a 
 full, as well as unanimous, quota. But the voters can not be expected 
 to divide themselves into ninths so perfectly; and as that is the case, 
 the rules of the election system must provide for giving the seats 
 to the nine largest groups that actually do form. We can now see the 
 reason for dividing the total vote of the first list by 5, that of the 
 second list by 2, and that of the third list by 4 to determine to which 
 the ninth seat should be assigned. We divide the total vote of the 
 first list by 5 to find out how large each group would be if five instead 
 of four were made out of the first list’s vote. We divide the total vote 
 of the second list by 2 to find out how large each group would be 
 if two groups instead of one were made out of that list’s vote. And 
 after dividing the third list’s vote by 4 for a similar reason, we 
 assign the next seat (the ninth) to that one of the three lists “which 
 shows the largest quotient,” as the rules put it, in other words to 
 that one of them which, if it were assigned one more seat, would be 
 electing each of its members by the largest part of a full quota. 
 
 Compare the errors of this election with those of an election of 
 representatives under our present single-member geographical con¬ 
 stituency system. Under the latter the error involved in the election 
 of a member often amounts, as we have seen, to from 40 to 60 per 
 cent of all the voters of a constituency, and it would often amount 
 to nearly 50 per cent of them even if a majority preferential system 
 were substituted for our usual system of pluralities and primaries. 
 In the election we have examined, on the other hand, the error, in 
 the case of the election of each member, is only the difference between 
 the number of votes that elect him and a full quota. In the case of 
 Candidates B, P, A, and V, for example, the error is the difference 
 between 24,680 (the number of times that 4 goes into 98,720) and 
 24,335U which is the full quota. In the case of candidates J and K 
 the error is the difference between 21,285^ (the number of times that 
 2 goes into 42,561) and 24,335U the full quota. In the case of Can¬ 
 didates Q, N, and M the error is the difference between 25,91 If (the 
 number of times that 3 goes into 77,735) and 24,335U the full quota. 
 
 It should be noticed that every member elected by this system is 
 elected by a constituency that may fairly be called unanimous. 
 Take the case of Candidate V. It is true, of course, that many of 
 the 21,285^ votes by which Candidate V is elected were cast by voters 
 who preferred one of the other candidates, B, P, O, S, T, or Y. Yet 
 every one of those voters indicated, by marking candidates in the 
 first list after reading the “Directions to voters” at the top of the 
 ballot, that he wanted to help elect some other candidate in that list 
 if his vote could not be effective for the particular candidate he 
 marked. It may be objected that the list itself did not precisely 
 suit every one of the voters who decided to vote for some candidate 
 on it. That is true* but it must at least have nearly suited every 
 politically intelligent voter who supported it, for otherwise the 
 unsatisfied would have taken advantage of the opportunity, easily 
 afforded them by section 1 of the election provisions, to nominate one 
 or more other lists. Within the limits of the flexibility of a list or 
 
EFFECTIVE VOTING. 
 
 19 
 
 party system, therefore, the constituency electing Candidate V— 
 and the same may be said of any of the eight others elected—may be 
 said to be unanimous. 
 
 The proportional system set forth above provides, therefore, a 
 means by which at a single election the voters of a large district can 
 form almost perfectly unanimous constituencies each one of which 
 contains approximately the same number of voters that there would 
 be in an entire single-member district under our present system. 
 Under the proportional system explained, therefore, the delegation 
 of members from one of the large districts—for example, that of 
 Newark—would represent the voters of that part of the State almost 
 perfectly; and as the same would be true also of the two other large 
 districts of the State, the legislature as a whole would represent 
 almost perfectly the voters of the State. 
 
 Besides the list system of proportional or true representation, there 
 are two others equally worthy of our attention in this country at the 
 present time. 
 
 One of these is the Hare system, which is advocated for the election 
 of city councils and fully explained in Pamphlet No. 2, 1 of the Ameri¬ 
 can Proportional Representation League, which will be mailed by 
 either secretary of the league to any address on receipt of 4 cents 
 in stamps. This system is often called “the single transferable 
 vote”; but that name is not altogether satisfactory, for in most 
 other systems also, including the list system explained above, the 
 vote is “single,” each voter’s ballot actually counting for only one 
 candidate, and it is “transferable,” though only within the limits 
 of one list. A more accurate name for the Hare system would be 
 the unrestricted-preferences system , for its distinguishing characteristic 
 is that it allows each voter to indicate, unrestricted by any party 
 or list lines, his personal preferences, as many or as few as he pleases, 
 among all the candidates on the ballot whose names are printed in a 
 single hst, the rules for counting the ballots under the Hare system 
 being suitable, of course, for making every ballot effective, so far as 
 possible, according to the individual preferences expressed on it 
 
 This is the system in use for the election of the Parliament of 
 Tasmania and the Senate of South Africa, and incorporated by 
 special votes of the British Parliament in the “Home rule” bill 
 for the election of the senate and the house proposed for Ireland. 
 That it is the most flexible and perfect of all systems of voting for 
 representatives is hardly open to question. It requires, however, 
 for the carrying out of the transfers required to make wasted votes 
 effective, the bringing together to a central place of all the ballots 
 cast in the large district which elects several members. As this 
 requirement will doubtless be felt by some people to be at present 
 a somewhat serious obstacle to the introduction of this system in 
 elections covering an extended territory, the list system just explained 
 and the “schedule system,” explained in the April (1913) Equity, 
 are worthy of our careful attention now in this country, especially 
 in respect to State and Federal elections, in which each district 
 would be very extensive. 
 
 1 For some slight amendments that should be made in the Ilare rules as printed in that pamphlet and in 
 the last issue of Equity, see an article elsewhere in the present issue. 
 
20 
 
 EFFECTIVE VOTING. 
 
 Under the schedule plan the candidates 7 names are printed on 
 the ballot in a single list, as under the Hare plan, instead of in 
 several lists, as under the list plan. But under the schedule plan each 
 candidate really stands for a list, for the distinguishing feature of 
 the plan is that every ballot that can not help elect the candidate 
 for whom it was cast is to be counted to help some other candidate 
 on a list published before the election as that of the candidates who 
 are to receive, in the order in which their names stand on the list, 
 any votes cast for the candidate in question which can not help him 
 either because he has a full quota without them or because he is found 
 to have no chance of being elected with them. 
 
 The differences between the schedule plan and the list plan are 
 these : 
 
 (1) The former lends itself to the Australian form of ballot, 
 whereas the latter lends itself to the party-list form of ballot. 
 
 (2) The former offers the voter many lists, a different one for each 
 candidate, without making the ballot physically cumbersome, 
 whereas the latter either restricts the voter to a few lists or makes the 
 ballot cumbersome. Neither plan, of course, makes the ballot “ long 77 
 in the political sense of that word; for politically a ballot is 11 long 77 
 that is hard to vote so as to make the voter’s will effective, and a 
 ballot is “ short 77 that is easy to vote so as to produce that result; 
 and it is very easy for the voter to vote so as to make his will effective 
 with either a schedule or a list plan ballot. 
 
 (3) The schedule plan allows a candidate’s name to be on more 
 than one list, whereas the list plan, unless the rules are made rather 
 complicated, does not. 
 
 Either plan, if carried out with suitable rules, may be expected to 
 give nearly perfect results. Either will elect a representative body 
 incomparably more useful, because incomparably more truly repre¬ 
 sentative of the voters, than those we have had in the past. To a 
 legislature elected by either may safely be intrusted the power to 
 select the chief administrative official of the State—not a governor, 
 with veto powers, but a sort of State manager with purely adminis¬ 
 trative duties—from anywhere in the country or the world to serve 
 indefinitely, on a professional basis, so long as he is satisfactory to 
 the truly representative legislature and to the people. Either may 
 be made the basis, therefore, of a form of State government at once 
 truly democratic and highly efficient.. 
 
 As for the election of the single-chambered council or commission 
 by which our cities should be governed under the “ manager plan, 77 
 for that I have been advocating the Hare system because it gives 
 the individual voter even greater freedom in the expression of his 
 will than the list or the schedule system. Wherever, however, the 
 Hare system is thought too hard to explain to the public, or the 
 opposition to bringing together all the ballots of the city—or of one 
 of the large districts, if the city is divided into such districts—is 
 insurmountable, either the list or the schedule system is to be recom¬ 
 mended, with confidence that either will give general satisfaction 
 and make the excellent “manager plan,” which for efficiency is cer¬ 
 tainly the best general plan of city government, truly democratic. 
 
EFFECTIVE VOTING. 
 
 21 
 
 CONCLUSION. 
 
 To sum up a few of the points covered by this article, there is no 
 place at all hi a plan of government suited to modern conditions for 
 either primary elections or plurality voting. When the object of the 
 voting is to make a selection among more than two candidates for 
 one nonrepresentative office, or to make a decision for one of more 
 than two alternative measures or policies, the voting should be done 
 by one of the majority systems explained in the first part of this 
 article, preferably by the Nanson system, where the counting of the 
 ballots by that system would not be seriously inconvenient. When 
 the object of the voting is to make up a selection-making or decision¬ 
 making body—that is, a deliberative, representative, legislative, policy¬ 
 determining body—the voting should be done by a unanimous con¬ 
 stituency or proportional system, such as the Hare, the list, or the 
 schedule. In all majority voting the voter should be allowed to 
 express his will as fully as he pleases. In the case of unanimous- 
 constituency voting he should be allowed to express his personal 
 preferences with the utmost freedom possible; in other words, the 
 Hare system should be preferred to the list or the schedule except 
 where public opinion is not sufficiently educated to demand the Hare. 
 The marking of majority ballots should in all cases be by the figures 
 1, 2, etc.; that of unanimous-constituency ballots by the same method 
 (in the case of the Hare system) or by a simple cross (in the case of 
 the list or the schedule system). Finally, the representative body 
 of a State or a city elected by a system insuring true representation 
 should be given the power to select and to retain in office during its 
 pleasure indefinitely, subject only to the application of the recall by 
 the people where that institution is in use, a chief administrative 
 official of professional experience and attainments. 
 
 o 
 
3 0112 061891765