Proportional Representation: A Study in Methods of Election Part 2

Attercliffe 6,532 5,354 1,178 -- Brightside 5,766 3,902 1,864 -- Central 3,271 3,455 -- 184 Eccleshall 5,849 6,039 -- 190 Hallam 5,593 5,788 -- 195 ------------------------------------------- 27,011 24,538 2,473

It will be seen that the Ministerial majority in each of the Attercliffe and Brightside divisions was larger than the aggregate of the Unionist majorities in the other three divisions; yet the Unionists obtained three seats out of five.

In the same election the result of the contested seats in London (including Croydon and West Ham) was as follows:--

Parties. Votes Obtained. Seats Obtained.

Unionist . . . . . . 268,127 29 Ministerialist . . . . 243,722 31

The Unionists were in a majority of 24,405, but only obtained a minority of the seats. Had their majority been uniformly distributed throughout London there would have been an average majority for the Unionists of 400 in every const.i.tuency, and in that case the press would have said that London was solidly Unionist.

It may be contended that the foregoing are isolated cases, but innumerable examples can be culled from electoral statistics showing how a system of single-member const.i.tuencies may fail to secure for majorities the influence and power which are rightly theirs. In the General Election of 1895 the contested elections yielded the following results:--

GENERAL ELECTION, 1895 (Contested Const.i.tuencies)

Parties. Votes. Seats.

Unionists . . . . . . 1,785,372 282 Home Rulers . . . . 1,823,809 202

These figures show that in a contest extending over no less than 484 const.i.tuencies the Unionists, who were in a minority of 38,437, obtained a majority of 80 seats. In this election, if an allowance is made for uncontested const.i.tuencies, it will be found that the Unionists were in a majority, but in the General Election of 1886 the figures for the whole of the United Kingdom (including an allowance for uncontested seats made on the same basis[6]) were as follows:--

GENERAL ELECTION, 1886 (All Const.i.tuencies)

Parties. Votes Obtained. Seats Obtained.

Home Rulers . . . . 2,103,954 283 Unionists . . . . . . 2,049,137 387

This election was regarded as a crus.h.i.+ng defeat for Mr. Gladstone. He found himself in the House of Commons in a minority of 104, but his supporters in the country were in a majority. The results of the General Election of 1874--although the system of single-member const.i.tuencies had not then been made general--are equally instructive. The figures are as follows:--

GENERAL ELECTION, 1874

Parties. Votes Seats Seats in Obtained. Obtained. proportion to Votes.

Conservative . . . . . . 1,222,000 356 300 Liberal and Home Rulers . 1,436,000 296 352

From this it appears that in 1874, while the Liberals in the United Kingdom, in the aggregate, had a majority of 214,000 votes, the Conservatives had a majority of 60 in the members elected, whereas with a rational system of representation the Liberals should have had a majority of 52.[7]

Such anomalous results are not confined to this country; they are but examples of that inversion of national opinion which marks at all stages the history of elections based on the majority system. Speaking of the United States, Professor Commons says that "as a result of the district system the national House of Representatives is scarcely a representative body. In the fifty-first Congress, which enacted the McKinley Tariff Law, the majority of the representatives were elected by a minority of the voters." In the fifty-third Congress, elected in 1892, the Democrats, with 47.2 per cent, of the vote, obtained 59.8 per cent, of the representatives.

The stupendous Republican victory of 1894 was equally unjustified; the Republican majority of 134 should have been a minority of 7, as against all other parties.[8] Similarly in New South Wales the supporters of Mr.

Reid's government, who secured a majority of the seats at the election of 1898, were in a minority of 15,000. The figures of the New York Aldermanic election of 1906 show an equally striking contrast between the actual results of the election and the probable results under a proportional system:--

_A "game of dice."_

Parties. Seats Seats in Obtained. proportion to Votes.

Republican 41 18 Democrat 26 27 Munic.i.p.al Owners.h.i.+p Candidates 6 25 Socialist -- 2

It is unnecessary to proceed with the recital of the anomalous results of existing electoral methods. It has been abundantly shown that a General Election often issues in a gross exaggeration of prevailing opinion; that such exaggeration may at one time involve a complete suppression of the minority, whilst at another time a majority may fail to obtain its fair share of representation. M. Poincare may well liken an election to a game of dice (he speaks of _les coups de de du systeme majoritaire_,) for no one who has followed the course of elections could have failed to have observed how largely the final results have depended upon chance. This, indeed, was the most striking characteristic of the General Elections of 1910. In the January election there were 144 const.i.tuencies in which the successful member was returned by a majority of less than 500. Of these const.i.tuencies 69 seats were held by the Ministerialists and 75 by the Unionists. The majorities were in some cases as low as 8, 10, and 14. The aggregate of the majorities in the Ministerialist const.i.tuencies amounted to 16,931, and had some 8500 Liberals in these const.i.tuencies changed sides, the Ministerialist majority of 124 might have been annihilated. On the other hand, the Unionists held 75 seats by an aggregate majority of 17,389, and had fortune favoured the Ministeralists in these const.i.tuencies 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.

_The importance of boundaries_.

All these anomalies can be traced to the same cause--that with a single-member system the whole of the representation of a const.i.tuency must necessarily be to the majority of the electors, whether that majority be large or small. It directly follows that the results of elections often depend not so much upon the actual strength of political parties, as upon the manner in which that strength is distributed over the country. If that strength is evenly distributed, then the minority may be crushed in every const.i.tuency; if unevenly distributed any result is possible. In the latter case the result may be considerably influenced by the manner in which the const.i.tuencies are arranged. A slight change in the line of the boundaries of a const.i.tuency might easily make a difference of 50 votes, whilst "to carry the dividing line from North to South, instead of from East to West, would, in many localities, completely alter the character of the representation." [9] An example will make this statement clear. Take a town with 13,000 Liberal and 12,000 Conservative electors and divide it into five districts of 5000 electors each. If there is a section of the town in which the Liberals largely preponderate--and it often happens that the strength of one or other of the parties is concentrated in a particular area--the net result of the election in five districts will depend upon the way in which the boundary lines are drawn. The possible results of two different distributions may be shown in an extreme form thus:--

Const.i.tuency Libs. Cons.

1st. 4,000 1,000 Lib. victory.

2nd. 2,400 2,600 Cons. "

3rd. 2,300 2,700 " "

4th. 2,200 2,800 " "

5th. 2,100 2,900 " "

------ ------ 13,000 12,000

Const.i.tuency Libs. Cons.

1st. 2,600 2,400 Lib. victory.

2st. 2,600 2,400 Lib. "

3st. 2,600 2,400 Lib. "

4th. 2,600 2,400 Lib. "

5th. 2,600 2,400 Lib. "

------ ------ 13,000 12,000

_The gerrymander_.

With one set of boundaries the area in which the Liberals largely preponderate might be enclosed in one const.i.tuency. The Liberals might obtain a majority of 3000 in this const.i.tuency but lose the other four seats. If, however, the boundary lines were so arranged that each const.i.tuency included a portion of this excessively Liberal area, the Liberals might obtain the whole of the five seats. In both cases the result of the election would fail to give a true presentation of the real opinions of the town. The influence of boundaries in determining the results of an election has been clearly realized in the United States for more than a century. Professor Commons states that whenever the periodical rearrangement of const.i.tuencies takes place the boundaries are "gerrymandered." "Every apportionment Act," says he, "that has been pa.s.sed in this or any other country has involved inequality; and it would be absurd to ask a political party to pa.s.s such an Act, and give the advantage of the inequality to the opposite party.

Consequently, every apportionment Act involves more or less of the gerrymander. The gerrymander is simply such a thoughtful construction of districts as will economize the votes of the party in power by giving it small majorities in a large number of districts, and coop up the opposing party with overwhelming majorities in a small number of districts.... Many of the worst gerrymanders have been so well designed that they come close within all const.i.tutional requirements." [10]

Although the National Congress has stated that the district for congressional elections must be a compact and contiguous territory, the law is everywhere disregarded.

The word "gerrymander" has found its way into English journalism. It was used by Liberals in their criticism of Mr. Balfour's abortive redistribution scheme of 1905, and has been equally used by Unionists in 1909 in their criticism of Mr. Harcourt's London Elections Bill. On neither occasion was the word used in its original meaning, and, although its history is to be found in most works on electoral methods, the story may, perhaps, be repeated with advantage:--

"The term Gerrymander dates from the year 1811, when Elbridge Gerry was Governor of Ma.s.sachusetts, and the Democratic, or, as it was then termed, the Republican party, obtained a temporary ascendency in the State. In order to secure themselves in the possession of the Government, the party in power pa.s.sed the famous law of 11 February 1812, providing for a new division of the State into senatorial districts, so contrived that in as many districts as possible the Federalists should be outnumbered by their opponents. To effect this all natural and customary lines were disregarded, and some parts of the State, particularly the counties of Worcester and Ess.e.x, presented similar examples of political geography. It is said that Gilbert Stuart, seeing in the office of the _Columbian Centinel_ an outline of the Ess.e.x outer district, nearly encircling the rest of the country, added with his pencil a beak to Salisbury, and claws to Salem and Marblehead, exclaiming, 'There, that will do for a salamander!' 'Salamander!' said Mr. Russell, the editor: 'I call it a Gerrymander!' The mot obtained vogue, and a rude cut of the figure published in the _Centinel_ and in the _Salem Gazette_, with the natural history of the monster duly set forth, served to fix the word in the political vocabulary of the country. So efficient was the law that at the elections of 1812, 50,164 Democratic voters elected twenty-nine senators against eleven elected by 51,766 Federalists; and Ess.e.x county, which, when voting as a single district had sent five Federalists to the Senate, was now represented in that body by three Democrats and two Federalists." [11]

Mr. Balfour's scheme did not involve a political rearrangement of boundaries, and the word "gerrymandering" was thus incorrectly employed in relation to it, but so long as we retain a system of single-member const.i.tuencies a Redistribution Bill will always invite suspicion because of the possibilities of influencing the arrangement of const.i.tuencies which such a measure affords. Instructions are usually given to boundary commissioners to attach due consideration "to community or diversity of interests, means of communication, physical features, existing electoral boundaries, sparsity or density of population;" [12] but although such instructions are at once reasonable and just, they would not prevent, and indeed might be used to facilitate, a gerrymander in the American sense of the term were such a proceeding determined upon. It is quite conceivable that a mining district in which one party had a very large majority might be surrounded by an area in which the political conditions were more balanced, but in which the opposite party had a small majority. If that mining area was, in accordance with the wording of these instructions, treated as one const.i.tuency because of its community of interests and the surrounding area divided into three or more districts, the minority would in all probability obtain a majority of seats.

_ The modern gerrymander_

The new const.i.tuencies required by the South Africa Act of 1909 have been arranged with the utmost care,[13] but had the delegates to the South African National Convention adhered to their original proposal to abandon single-member const.i.tuencies, they would have secured for South Africa, among other invaluable benefits, complete security from the gerrymander, any possibility of which begets suspicion and reacts in a disastrous way upon political warfare. The gerrymander is nothing more or less than a fraudulent practice. But the United States is not the only country in which such practices take place. Their counter-part in Canada was described by Sir John Macdonald as "hiving the grits," and even in England, without any change of boundaries, practices have arisen within the last few years which have had their birth in the same motives that produced the American gerrymander. In boroughs which are divided into more than one const.i.tuency there is a considerable number of voters who have qualifications in more than one division. A man may vote in any division in which he has a qualification, but in not more than one. He may make his choice. In Edinburgh for many years, on both sides of politics, there has been a constant transfer of voters from one register to another in the hopes of strengthening the party's position in one or other division. It was even alleged that the precise moment of a vacancy in West Edinburgh (May 1909) was determined by the desire to ascertain the strength of the Unionist party in that division, to discover how many Unionist votes should be transferred for the purpose of improving Unionist prospects or of defeating the designs of their opponents. This allegation may be wholly unfounded, but the single-member system encourages such a proceeding, and the statement at least indicates how the voting power of a division may be manipulated. The mere possibility of such an action arouses the suspicion that it has taken place. Similar practices have, it is stated, been pursued in Bristol. Votes have been transferred from one division, where one of the parties was in a hopeless minority, for the purpose of strengthening its position in other divisions. An examination of the figures of the election in Birmingham in 1906 shows that in one division, Birmingham East, the Unionists narrowly escaped defeat. They won by a majority of 585 only.

In the other divisions the Unionists won by very large majorities. Must not the possibility of transferring surplus votes in strong const.i.tuencies to strengthen the position in weak const.i.tuencies prove an irresistible temptation to the agents responsible for the success of the party? They are ent.i.tled to make use of all the advantages at their disposal. In this way a new and more subtle form of the "gerrymander"

has arisen in England, and if we are to redeem English political warfare from proceedings which approximate very closely to sharp practices, we must so amend our electoral system as to give due weight to the votes not only of the majority but of the minority as well.

_The Block Vote_

The a.n.a.lysis of the results of majority systems would not be complete without some reference to the use of the "block" vote in the London County Council, the London Borough Council, and other elections. In the London County Council elections each const.i.tuency returns two members, and each elector can give one vote to each of two candidates. The Metropolitan boroughs are divided into wards returning from three to nine members, each elector giving one vote apiece to candidates up to the number to be returned. [14] Both in the London County and London Borough elections the majority, as in a single-member const.i.tuency, can obtain the whole of the representation. All the defects which arise from parliamentary elections again appear, and often in a more accentuated form. The figures of the two London County elections, 1904, 1907, disclose a catastrophic change in representation similar to that which characterized the General Election of 1906:--

LONDON COUNTY COUNCIL ELECTION, 1904

Seats in Parties. Votes. Seats proportion Obtained. to Votes.

Progressive and Labour 357,557 83 64 Moderate 287,079 34 52 Independent 12,940 1 2

Progressive majority over Moderates 70,478 49 12

LONDON COUNTY COUNCIL ELECTION, 1907 Seats in Parties. Votes. Seats proportion Obtained. to Votes.

Moderate 526,700 79 67 Progressive and Labour 395,749 38 50 Independent 6,189 1 1