The efforts of parties to give effect to the reforms in which they have been more particularly interested have so far ended in failure. In 1905 Mr. Balfour introduced a Bill for the redistribution of seats, unaccompanied by any reform of the franchise. The measure was met with the cry of "gerrymander!" and its disappearance with the fall of the Government was regretted by few. In 1907 the Liberal Government attempted to deal with the franchise problem, apart from any scheme of redistribution. It endeavoured in Mr. Harcourt's Plural Voting Bill, a highly complex measure, to give effect to the principle of "one man, one vote." This Bill was strongly opposed on the ground that the reform was partial in character. If, said the opponents of the measure, it is unfair that one elector should have twelve votes whilst another elector has but one, it is equally unfair that the vote of an elector in one constituency should be twelve times as valuable as the vote of an elector in another constituency. The justice of the argument must be admitted, and explains why the rejection of the Plural Voting Bill by the House of Lords aroused comparatively little public feeling. Yet the rejection of this Bill has focussed attention upon the deficiencies of our franchise laws, and the eyes of all politicians are turning towards that more comprehensive measure of electoral reform which cannot be indefinitely postponed. Such a measure has been categorically promised by Mr. Asquith on more than one occasion. So far back as 1908, soon after his accession to the Premiership,[1] he made the following public declaration: "I regard it as a duty, and indeed as a binding obligation on the part of the Government, that before this Parliament comes to an end they should submit a really effective scheme for the reform of our electoral system."
The simplification of the franchise.
What are the lines on which a really effective scheme can be framed? The fate of the partial measures already referred to is at least an indication of the difficulties which will attend any attempt to carry an incomplete scheme. It may be assumed that an effective scheme must deal with the three problems named: franchise (including registration), redistribution, and three-cornered contests. Each of these factors must be dealt with as simply as a due recognition of the problem to be solved will allow. The complexity of Mr. Harcourt's Plural Voting Bill was due to the fact that we possess no less than twenty[2] different franchises. But the remedy is easy. "If," said the late Sir Charles Dilke, "they wanted to cheapen the cost, to remove the disgrace from this country of having registration more full of fraud and error than anywhere else, they could only do so by some simple franchise. All registration reform was condemned to failure until they made up their minds on a simple and easy basis for the franchise, sufficiently wide to enable them to absorb all existing franchises." Such a simple franchise is to be found in manhood suffrage, which would admit of the easy transfer of electors' names from the register of one electoral division to another. The chief objection to this solution, which arises from the fear that the most numerous class in the country may monopolise representation, may be met by linking the adoption of a simple franchise with a system of election which shall give due representation to minorities.
Redistribution.
Redistribution must be treated with like boldness, but before considering the principle on which this reform must be based, it would be well to give some indication of its urgency. Here are the figures of four of the largest and four of the smallest English constituencies as given in the Parliamentary Return of 1911:— *** Constituency. Electors. Constituency. Electors. Romford (Essex) 55,951 Durham. 2,698 Walthamstow (Essex) 42,029 Bury St. Edmunds 2,878 Wandsworth 39,821 Whitehaven 2,989 Harrow (Middlesex) 38,865 St. George's, Tower Hamlets 3,252
Should be automatic.
It will be observed that an elector in Durham has twenty times the political power of an elector in the Romford Division of Essex. Nor are these discrepancies confined to England. There are great divergencies between the electorates of individual constituencies in Scotland and Ireland, and any measure of redistribution which attempted to deal effectively with these would necessarily have to be of a far-reaching character. Even were it possible to effect a readjustment by the creation of parliamentary areas containing an equal number of electors, so rapid are the changes in the electorate that the scheme would be out of date almost before it came into force. Mr. Ellis T. Powell has published a valuable table entitled "the process of electoral evolution,"[3] in which he has arranged the constituencies in the order of their size as measured by the number of electors who were on the registers in 1886, and again in 1906. The table shows how remarkable has been the change in their relative importance. The rapidity of the change is still further indicated by a comparison based upon the 1908 register. Any one who has the curiosity to count the number of constituencies which retained the same position on the list both in 1906 and 1908 will find this to be the case in nineteen constituencies only out of a total of 481. So great, indeed, has been the change since 1901, the date of the last census, that no satisfactory scheme of redistribution could be framed upon the population figures of that year. It would seem that the only satisfactory principle upon which the problem can be solved is that of an automatic redistribution of seats on the completion of every census, but the difficulties associated with such a solution, if the present system of single-member constituencies is retained, are so overwhelming as to render it almost inadmissible. True, the South African Constitution provides for the automatic redistribution of seats after every quinquennial census,[4] and the Canadian Constitution contains a similar provision, but the inconveniences attaching to a rearrangement of boundaries are not so great in new countries as those which obtain in an established country. Moreover, as time goes on, the inconveniences associated with rapid changes in boundaries will be felt more and more both in Canada and in South Africa. For local authorities[5] rightly complain of the difficulties which arise from the creation of different areas for different purposes and the consequent overlapping of boundaries, and these difficulties would increase were fresh parliamentary divisions created every ten years. The problem which would be involved in the creation of new parliamentary divisions for London is such as to render a satisfactory scheme almost impossible. Apart, however, from these considerations, the difficulties of another kind attendant upon the creation of new constituencies are so great that it is quite easy to understand the unwillingness of the leaders of both political parties to embark upon schemes of redistribution. The influence of boundaries upon the political fortunes of parties is so well known that any rearrangement, whether in the metropolis or in the large towns, would probably be looked upon with very grave suspicion, and the more so that in several towns party organizations have already endeavoured to obtain the maximum of party advantage under existing conditions.
Secures neither one vote, one value nor true representation. Further, it has been proved beyond question that a redistribution of seats will, if single-member constituencies are retained, fail to accomplish the end which its advocates have in view, namely, one vote one value. For redistribution can only secure equality in the size of electoral districts, and this is not the same as equality in the value of votes. With equal electoral districts it would still be possible in two adjoining constituencies for one member to be returned by a large majority and the other by a small majority. In Wales it might still happen that a Conservative vote would be valueless for the purpose of obtaining representation. Equality in vote value is only secured when the votes of electors of all parties are equally effective. This can only happen when the representation of parties is brought into agreement with their voting strength.
The Royal Commission on Electoral Systems entered very carefully into the probable effect of redistribution upon the representation of parties within the House of Commons, and came to the conclusion that, so far "as facts can be adduced to test it, the theory that the varying size of constituencies accounts for the exaggeration of majorities falls to the ground." This conclusion—and the Commission could hardly have come to any other—is in agreement with the opinions expressed both by Mr. S. Rosenbaum, of the Royal Statistical Society,[6] and by Mr. J. Rooke Corbett, of the Manchester Statistical Society.[7] The following summary of the results of Mr. Corbett's analyses of the eight General Elections 1885-1910 shows conclusively that redistribution would fail to remedy the inequalities in representation arising from a system of single-member constituencies: