Last month, the Welsh Parliament approved the Senedd Cymru (Members and Elections) Act. This law establishes that from the next election (scheduled for 2026) the number of seats in the Welsh Parliament will rise from 60 to 96, who will be elected by closed-list PR from 16 six-seat districts, replacing the MMP system used ever since the assembly’s establishment. The Parliament’s term will also return to 4 years (it was changed in 2011 to 5 years by Westminster legislation aiming to avoid having devolved assembly elections at the same time as Westminster elections).
This reform is the result of a discussion held in various committees over the past few years. STV has long been considered for Wales, and was recommended by an expert panel, but the political committee stated a preference for closed list PR due to the ability to enforce lists zippered by gender (a requirement which will apparently be legislated by a separate act). I suspect the choice is also motivated by the way STV would take away power from party leaders, but this is speculation.
I think this will be the first UK assembly (as opposed to members of a supranational assembly) elected by closed list PR? UK electoral system experimentation continues…
Fruits and Votes 
Interesting to see this with Germany also moving away from MMP for their next election. (Though I don’t think Wales’ system was ever very proportional, it didn’t have a lot of topup seats IIRC. Wiki called it the Alternative Member System, which I think is like less-proportional MMP).
I always thought MMP was sort of a strange hybrid that didn’t really make conceptual sense. My old copy of Handbook of Electoral Systems says that in Germany the district reps were basically list politicians who’d make some vague noises about representing their district’s interest here or there. So really not like single district reps in the Anglosphere.
If you really want both single-rep districts and some degree of proportionality, I always thought dual member PR was underrated. Or my ‘3 for 5’ system
LikeLike
The common term in the UK for MMP is “additional member system” (AMS). I find the term objectionable for two reasons: (1) it implies the list (“top-up”) members are mere add-ons and thus implicitly less legitimate, and (2) they are not “additional” but compensatory (“additional” seems more like how the seats are determined under MMM/parallel).
The Welsh variant was indeed not highly proportional due to the limited share of compensatory seats (1/3 of the total) and also with only 40 single-seat districts lopsided results in that tier were likely and common. For instance, in 2021, Labour won 27 of the 40 districts on 40% of the nominal (candidate) vote; even after compensation they had 30 seats, so 50% of total, even though their party-list vote was only 36%. I suppose this disproportionality was a key reason for the move away from MMP, although I can’t say for certain.
As long as they were willing to go with such a large increase in assembly size, they might have considered MMP with 48 seats in each tier, which would have allowed for considerably better proportionality. I wonder if this was considered.
There is actually considerable evidence that German MPs do attend to district concerns. The key point is that it is not only the ones elected in the district that do so. The list members also make efforts to represent the district in which they had a nomination. (This is true in New Zealand, too, and more so–perhaps due to the FPTP legacy.) This is the sense in which it does make conceptual sense, as a PR system with local representation.
LikeLike
I wonder, under MMP system, if the members of the large parties have more incentive to attend to constituency service in their districts, because they are more likely to get their seats via the SMD part rather than via the list part.
LikeLike
I doubt it. Famously in all of the mixed systems I’ve ever heard of the SMD reps can still get into Parliament on the list even if they lose in their district. The Japanese even have a term for this, something that translates to ‘zombie candidates’. So you have less incentive to really win because you can get into Parliament either way. (After all, if you were in favor with the party enough to be their SMD rep, they probably like you enough to put you high on the list).
I’d actually be interested in hearing about any mixed systems that don’t let the SMD loser get in via the list
LikeLike
Patrick, I would say yes, they do. But it is probably less because they might win these seats than that having locally visible and active MPs (and candidates) helps them attract list votes.
Small parties do this, too, at least to some degree. (See my post on “Green contamination” from some years ago for one example.)
To the comment by hash872, allowing candidates to be elected either way–and be dually listed–is actually essential for getting the incentives right. So most MMP (although somewhat fewer MMM) systems allow this. If those who place second or lower in a district contest can’t also be on the list, then parties have to be careful not to “waste” their best personnel on district contests where they are less viable. When those who are likelier to win from the list rather than a district are also nominated in a district, they have incentive to be present before local voters even though they may not win there. The party can expect such a candidate/MP to help boost the party vote through his or her activities.
There’s actually a lot of research on these points, and quite a few posts at this blog on them as well.
Also, the term “zombie” is quite inaccurate, as are similar phrases like “back door MPs.” These are just cries of those who think only FPTP and similar systems are somehow legitimate. Note also that Japan is MMM, not MMP. This distinction matters, as MMM makes winning districts critical to a party’s seat total, whereas MMP usually does not. Japan also allows parties to set list ranks according to candidates’ relative performance in their district contests, whereas most other mixed-member systems use standard ranked, closed lists.
LikeLike
My Oxford Handbook of Electoral Systems says that over 80% of Bundestag members are dual candidates, running for both a district and a list. (Mr. Shugart was a contributor to this book, though he did not write this specific piece on Germany). The book makes it sound like Germany’s strong party system and loyalty to the local branch of a party are determining factors in service to a specific district, more than the voters per se. The author Zittel lists arguments both for & against a connection between district representation and constituency service
LikeLiked by 1 person
I strongly doubt they considered 48 district seats, as this would have required another round of boundary delimitation (districting). The 40 districts they had before were based on the old Westminster district map which overrepresented Wales with 40 seats. The new map, which will go into effect for Westminster next month, gives Wales 32 seats. The new Welsh Parliament map combines pairs of Westminster seats for 16 districts – 6 seats per district gives 96 seats. Anyway, I don’t remember any report on electoral reform in Wales considering a reform of MMP, the recommendation was always STV or some other new system.
Matthew, how proportional is the new system likely to be, based on your models? Is Labour likely to continue to be so overrepresented under it as before?
LikeLike
JD, I’ll get on those questions as soon as I am done with a certain call I have scheduled for right about now!
LikeLike
They should not be so afraid of boundary delimitation! But I also do see the practical problem.
So they could have gone with 32 single-seat districts and 40+ compensation seats. No need to go all the way to 96 seats, as a 32/40 should be quite a good deal more proportional than the MMP system they had.
LikeLike
I did a very rough estimate of what the results would be with 16 6-seat districts under generic boundaries (I didn’t have data for the actual boundaries). I got 44 Labour, 28 Conservative, 23 Plaid, 1 Lib Dem.
That corresponded in a drop in the Gallagher Index for 2021 from 11.0 to 9.2.
Michael Gallagher’s sheet has the Gallagher Index at 9.36 for 2021 though, not 11.0, and I’m not sure why my results differ from his so take that with a grain of salt. Maybe a couple grains of salt given the boundaries issue I mentioned and not taking into account changes in voting behaviour…
I wonder if a more candidate-centred system would introduce more regional variance and help smaller parties win seats with strong candidates?
LikeLiked by 1 person
Sorry that should be 16 districts, not 32.
[I fixed it–Your Orchardist]
LikeLike
Pingback: What does the Seat Product Model say about the Welsh electoral reform? | Fruits and Votes