Under the single transferable vote system there is every inducement to vote for a crank, and he generally manages to amass a considerable number of late preferences. For most of us are cranks when probed as deep as a ninth or tenth preference.
James Creed Meredith is one of those neglected figures in Irish history who managed to play a important role in several fields of intellectual endeavour (including scence fiction), as well as setting up the legal system of the nascent Irish state, ending his career as a member of the Supreme Court from 1937 to 1942. In 1913, he turned his talents to the question of electoral systems for a devolved legislature in Ireland. The Home Rule Bill then being discussed provided for an all-Ireland Parliament with a 164 member House of Commons and a 40-member Senate; I confess that I was utterly unaware of the electoral arrangements prescribed, which were that the Senate should be entirely nominated at first and then elected by proportional representation from the provinces, and that the House of Commons would be elected from constituencies of a variety of sizes, with those electing three or more members (nine seats with 31 MPs) doing so by proportional representation (incidentally the multi-member seats were mainly urban, and the smaller districts rural, which - whether by accident or design - would surely have resulted in under-representation of the Left). The Irish Parliament had the power to change the House of Commons electoral system once established, but there was no such power as regards the Senate.
Meredith, a strong supporter of proportional representation over first-past-the-post (as we now call it), makes an equally strong argument that the single transferable vote is not a good system in general and that it was unsuited to Ireland in 1913 in particular. His preferred option instead is a modification of the voting system which then applied - and in fact still does - in, of all places, my adopted country, Belgium: an open list system, with the additional points that parties should be allowed to form tactical coalitions to pool their votes, and that the final seats should be allocated by Droop quota and largest remainder rather than by the d'Hondt method.
Meredith makes the point that STV is not in fact a proportional system, and that its supporters are entirely disingenuous about it effect on party politics; and much of this part of his critique remains pretty valid today, and in some respects it has been born out by Irish experience in the century since he wrote. It is undeniably true that STV encourages parties to put up fewer candidates than a list system does. It is also clear that the fewer candidates a party nominates, the greater the role of the party selectorate in choosing them. In my own Belgian village last October, four of the five parties put up a full slate of candidates for the 21 seats available, so we had 89 candidates to choose from (out of a population of just over 10,000, so roughly one inhabitant in a hundred was on the ballot paper); compare with Omagh in Northern Ireland in 2010, where there were also 21 seats up for grabs, but in three 7-seat areas with a total of 30 candidates (out of a population of just under 20,000, so one person in 600 was a candidate). In Oud-Heverlee, no party ran fewer than 5 candidates; in Omagh, no party other than Sinn Féin ran more than two candidates in any area (and the Shinners' largest slate was five).
From the point of view of party management, open lists are pretty ideal. You have a great excuse for candidate recruitment, candidates have every incentive to work hard at getting their own personal vote out (which benefits the party as a whole) and you don't have to worry about losing votes when they transfer away at later counts. It's not surprising that reform-minded Irish politicians today tend to advocate a move away from STV to a list system (missing the point that the real problem is when you insist that government ministers must also be burdened with constituency duties).
Where I part company with Meredith is that I don't agree that what is good for party managers is necessarily good for politics in general. I concede some of his points - including the argument that STV's favouring of moderate candidates against the extremes is in fact a strike against its claims to proportionality (it's a hit I am willing to take) - but it still seems to me that STV offers the voter more transparency and clarity over the process, and more influence over the result, than any list system ever can. Meredith makes much of the need to better integrate the reality of political parties into the electoral system; in fact a lot of that work has been done since 1913, with parties now registered legal entities, with certain statutory duties and obligations, in a manner undreamt of in 1983, never mind 1913.
Still, it's a very interesting book for us psephological anoraks. I do not believe I have read a more robust denunciation of STV from the pro-electoral reform side of the debate. I suspect that Meredith, by pointing out the impracticality of province-wide elections for the Senate (including 14 for Ulster, presumably the same for Leinster) may have had one immediate effect on the Home Rule legislation - the Senate ended up in the Act as a body whose members were to remain nominated but (after ts first term) by the Irish rather than the British government, making it an unabashed rubber stamp and effectively demonstrating, a century ago, that Ireland really doesn't need a second chamber at all.