Tyler Cowen in an aside at Marginal Revolution.
Most countries don’t use range voting. Ireland and Tasmania have had some experience with the Single Transferable Vote (STV) system. What happens is that a bunch of candidates run for each post, party identification is weak, and reps emphasize constituency service. That’s probably the major dominant effect, namely that most systems of range voting weaken political parties.
Not quite so, I don’t think, having spent a reasonable amount of my youthliving under, voting in and studying the Irish STV system (although I am not a parties scholar – Matthew Shugart or Simon Jackman may have better and more sophisticated things to say). The effect on constituency service is undeniable – Irish politicians neglect parish pump politics at their peril, and typically compete on the number of fixes, medical cards etc that they can deliver to their consituents. But this has typically gone together (at least until the very recent past – we will see what happens in the next election) with strong, and indeed near-tribal levels of party identification. Voters have typically strongly identified with the one or the other party. Fragmentation of the party system has been relatively low – while smaller parties occasionally emerge as challengers, they are usually re-absorbed into one of the three main parties sooner or later if they do not wither away of their own accord. There have been (I think – am not sure where to find the figures on this) more independents in the Dail (Irish parliament) in recent years running and winning on single local issues such as hospital location etc – but they usually are implicitly identified with the one or the other of the main parties, and tend to vote the party line on issues other than the one that they have run on.
Dail voting takes place under a whip, with punishments of suspension or expulsion from the party for TDs (members of the Dail) who vote against the party line. There are high levels of intra-party competition among TDs and potential TDs within constituencies – the most serious threat that a TD faces is from ambitious members of her own party who would like to take her seat. However, this competition usually involves a combination of appeal to local loyalties and constituency service. It does not lead to ideological division, or (usually) to a weakening of the party on the national level.
None of this is to say that Tyler is wrong, necessarily – but one of the few significant empirical examples we have would seem to argue that quite strong parties can be maintained in an STV system. Whether this is an aberration or a manifestation of some deeper logic of STV (when combined with other specific aspects of party competition) is very hard to say, given the tiny N that we have.
Update: See further John Coakley and Michael Gallagher’s book-length survey on Irish politics.
In this regard, the Irish experience would seem to run counter to the widely held theory that PR-STV leads to weaker party discipline and party organization. … Taagapera and Shugart argue that “if strength of party organization is desired, STV is inapprorpriate, because either list PR (even with preference voting) or plurality (in the absence of US-style primaries) gives far more leeway to party elites in deciding who the party’s representatives may be.’ … Katz … puts forward a more absolute version of the theory, hypothesizing that ‘Where intraparty choice is allowed, parliamentary parties will tend to be disunited’ … Blais pushes the argument even further: ‘There is strong evidence that the single transferable vote leads to a weaker party system … Electoral competition within the party hinders unity and cohesion … the single transferable vote, like preferential voting in general, is detrimental to the development of a responsible party system.’ … Both authors point to the prevalence of intra-party personalistic competition in the area of constituency service as evidence for their theory. However, both are also forced to declare that the equally evident unity of Irish political parties is ‘illusory’ (Katz …) and ‘superficial’ (Blais …). Neither author explains what illusory or superficial party unity is and neither provides evidence to demonstrate its existence or its consequences. It would seem more sensible to note that intense intra-party competition in the area of constituency service can coexist with a very substantial degree of party cohesion and party discipline, and that the latter are products of constitutional structure and inherited modes of politics and are not undermined by PR-STV.
So in short – many political scientists agree with Tyler – but Ireland is a highly awkward case for their arguments. Also, I should note that I hadn’t read Coakley and Gallagher before writing this post to my shame (I do know John, but haven’t been keeping up with the literature on Irish politics in the last decade or so), and that the striking similarity of our arguments should be taken as an indicator of how compelling the empirical evidence is on this question, rather than as a product of groupthink.
{ 8 comments }
I think the only other country which uses the PR-STV system is Malta. Malta has had over time an almost perfect 50/50 party divide, which is also problematic for this theory.
Actually, this might be something that Monkey Cage might like to highlight.
The Irish Parliamentary Joint Committee on the Constitution is seeking submissions on the Irish electoral system.
http://www.oireachtas.ie/viewdoc.asp?DocID=12767&&CatID=36
“Interested individuals and organisations are invited to submit submissions in writing to the Committee by 30 November 2009. Views expressed in the written submissions and presented at the public hearings will contribute to the Committee’s deliberations and will inform its report, including conclusions and recommendations, which will be presented to the Houses of the Oireachtas and the Government.”
This is the first I have ever seen of STV being claimed as a form of “range voting,” but whatever.
When we ask whether an electoral system “weakens parties,” obviously we want to know “relative to what.” Naturally, Irish parties have very strong voting cohesion in the Dail (and Maltese parties even more, I believe). And there is the “tribal” aspect of party ID that Henry mentions; I have never heard it put that way before, but I am no expert on Ireland and I won’t dispute it.
So it might be a bit odd to think of Irish parties as “weak.” Yet I think they are, comparatively. Two facts that Henry mentions point in that direction. First, that many small parties emerge, exist for a while, and then get reabsorbed into a major party or “whither away.” I am not aware of any systematic accounts, but I’ll be that has happened more often in the past several decades in Ireland than almost anywhere else in the “advanced” democracies, with the exception of Japan, which formerly had the electoral system most similar to STV (single NON-transferable vote, or SNTV).
And then there are the independents. Ireland has long had an unusually high number of nonpartisan candidates elected. Again, Japan would be the only case with more, I am almost certain. And, again, SNTV is about as similar to STV as you can get, although SNTV clearly undermines party control over candidates even more than STV does.
You certainly do not see so many winning independents in any first-past-the-post system that I can think of, among the “advanced” cases (UK, Canada, formerly New Zealand, or the USA for that matter).
This party-weakening argument is by no means undermined by the observation that the Irish independents who are elected “usually are implicitly identified with the one or the other of the main parties, and tend to vote the party line on issues other than the one that they have run on.” Quite the contrary, this observation goes to the heart of the incentives of a multi-seat personalistic electoral system. Japan, again, regularly had, under SNTV, independents elected and then actually reabsorbed into the long-ruling LDP. The ability of independents to win, knowing that the ruling party will have little choice but to deal with them, is just another symptom of a system that increases the value of personal reputation-seeking over the cultivation of the party vote. It makes the boundaries of “party” a bit more permeable and harder to define with the precision we typically see in other systems.
Nor is this latter point about the relative balance of the personal and the party vote undermined by the observation that intraparty “competition usually involves a combination of appeal to local loyalties and constituency service. It does not lead to ideological division, or (usually) to a weakening of the party on the national level.” The literature on electoral systems like STV and SNTV (as well as open-list systems like Finland’s) makes quite clear that this is not a contradiction. Members affiliate with parties for their collective pursuit of policy (as well as government office where that depends on holding a legislative majority, of course). Nonetheless, under electoral systems that place a premium on the personal vote, they must distinguish themselves from co-partisans in their district. They do so not on matters that would divide the party, but on those that don’t: local services, or other areas where the individual can specialize and “claim credit” as an individual.
Were STV to be adopted in a presidential system, such as the USA, I have little doubt that we would see the major parties be even weaker than they are now, because the imperative of cohesion as a means to control the government does not exist here. At the same time, there is little doubt that STV would let niche parties like the Greens and Libertarians gain representation. In fact, I think it would look a lot like Ireland, but with lower ideological content and differentiation of the big parties than we see on the Emerald Isle, and a lot less voting cohesion in the legislature, while perhaps more differentiation and discipline of the smaller.
This may be of interest to Monkey Cage readers of this thread.
The Irish Parliamentary Committee on the Constitution is holding hearing and seeking submissions in relation to the electoral system. More details are available here:
http://www.oireachtas.ie/Viewtxt.asp?UserLang=EN&DocID=12774
Malta currently has a provision that ensures a majority of seats, and hence control of the government, to the party with a majority (plurality?) of first-preference votes. That obviously undermines STV’s “unwasted vote” advantage (whereby a vote for a smaller party can transfer if that party is unelectable). That is a very big un-STV feature.
However, this was not always in place, yet two parties have dominated for most of Maltese independence.
But “weak parties” are perfectly consistent with “few parties.” These are different dimensions of a party system, and one should expect that where parties are weaker internally, there would be fewer of them–competition then turns to non-ideological factors, internal dissent can be tolerated to some greater extent than under other electoral systems, etc.
I do not know much about Maltese party politics. However, I do know that party ID in the electorate is strict. In that sense, parties are actually “stronger” than Irish. That means that few votes transfer across party lines. But that only makes competition WITHIN the party fiercer. Parties nominate many more candidates per seat than in Ireland, because they pretty much have no reason to fear that they will lose seats they could have won but for transfers to other parties’ candidates.
Matt – I had hoped you would be chiming in on this. But what does this have to say for the questions of party cohesion that Katz et al. point to? I am not aware of any systematic study of line voting in the Dail – but the impressionistic evidence would suggest that Irish party cohesion in the legislature is not unusually weak by any reasonable measure. Perhaps there is more on this in the pieces that Coakley and Gallagher are disagreeing with (I imagine that Katz knows a fair amount about Irish politics given his frequent collaborator) – but I would like to know what. I suppose part of my problem here is that I am not seeing what the operational definition of ‘weakness’ is. The most obvious definition to me would be something like ‘inability to coordinate parliamentary votes along party lines’ – but this is a definition that I don’t think fits the Irish case particularly well. I suspect that you are using a different definition than me here, which may be a standard one (I have not studied parties since my M.A. thesis) in the literature, but which could perhaps be brought out more explicitly for the benefit of non-specialists.
Yeah, STV is certainly not a form of range voting (aka score voting).
But there are proportional systems that are related to score voting and are superior to STV in some objective senses.
http://scorevoting.net/RRV.html
http://scorevoting.net/Asset.html
Also, the single-winner form of score voting is just vastly superior to the single-winner form of STV, called “Instant Runoff Voting”. Especially when it comes to how the systems deal with voter strategy.
Once voters maximally exaggerate on the front-runner candidates, score voting leaves them free to safely support candidates they like BETTER than those front-runners. E.g. you give Gore a 10 and Bush a 0, then you can still give Nader a 10 if you like him better than Gore. With IRV, top-ranking Gore means you MUST bury Nader.
That may not be a problem when “Nader” is a weak candidate who you KNOW will be eliminated. But say you were a Republican in the last Burlington VT mayoral election, who preferred the Democrat to the Progressive. Your faction could have gotten the Democrat instead of the Progressive by top-ranking the Democrat instead of the Republican. So in the future, that type of voter will learn to strategically top-rank their favorite between the Progressive and the Democrat, since the GOP functions as the (strong, not weak) “third party” in a left-leaning city like Burlington. Then a GOP candidate can’t win, even if he’s moderate enough to actually be the most overall representative candidate in a future election. He won’t be able to win because voters will not want to “throw away” their vote.
Kind of a tangent from the focus of this post, but relevant nonetheless.
Clay Shentrup
San Francisco, CA
206.801.0484
Henry, I agree that party “strength”/”weakness” is not all that useful a metric. I have stopped using it in my work, after some youthful indiscretions. My point was that a party can simultaneously (a) maintain voting cohesiveness in parliament while (b) allowing its members to provide for constituent services of various types, and sometimes even (c) tolerate “independents” operating on the fringes of the party or even outside it, yet cooperating on matters of mutual interest.
Now, I think a lot of folks would call a party that has characteristics b & c somewhat “weak” even if it also exhibits characteristic a, whereas a party that had only characteristic a would be a good deal “stronger.” But we should probably dispense with “strength” as our conceptualization.
There clearly are different dimensions of party organization and behavior here. And the point I was trying to make is that several of the stylized facts mentioned in the post are consistent with STV and its tendency to give members a strong (that word again) incentive to cultivate a person vote.
In general, cohesiveness in voting does not tell us much about the priorities of the individual members of the legislature who comprise the party, other than that they work out whatever internal differences in priorities they may have before they take votes. It tells us nothing about how they work those out, but the more the personal vote matters to legislators, the more party leaders are likely to have to expend time and resources offering concessions to get members to toe the line. How that works in Ireland is something I am not privy to.
(Funny how any time someone uses the words “range voting” on a blog, a certain fan of the concept always appears!)
Comments on this entry are closed.