The question of dual candidacy in the Welsh Assembly mixed-member proportional (MMP)* system is being debated again. “Dual candidacy” refers to a provision permitting candidates to run simultaneously in a nominal (district) race and on their party’s list for the proportional component of the system.
Wales permitted dual candidacy in 1999 and 2003 and banned it for elections of 2007 and 2011. Now a bill is in the House of Commons (yes, this decision is taken in London) to ban dual candidacy again.
Scully mentions various other reforms that have been debated, including an increase in the size of the assembly, from 60 to 80 or 100. As he notes, such an increase would have an impact on the proportionality of the system (independent of dual candidacy).
For instance: the easiest way to change from 60 to 80 AMs would be to raise the number of list AMs in each region (from 4 to 8). But with list AMs now comprising half of the Assembly’s membership, rather than one-third, the proportionality of the electoral system would be changed substantially. An 80-seat Assembly where 40 members came each from the constituency and list ballots would be more-or-less a fully proportional system, rather than the semi-proportional system we have at present.
This is an important point. Because the compensation in the Welsh MMP is carried out in regions instead of Wales-wide, and because the number of seats per region is relatively low, the proportionality is indeed modest. Michael Gallagher‘s Election Indices shows values on the Least Squares (Gallagher) index of disproportionality in the four elections of 8.61, 10.39, 11.36, and 10.47. By contrast, New Zealand, with nationwide proportionality in its MMP system and a 5% threshold, has had index values ranging from 1.13 to 3.84. The UK, with only single-seat districts, has averaged 16.53 on the index over the elections of the same period.
Alternatively, the number of constituencies for the nominal tier could be increased. To keep the same ratio between tiers as is current practice would require 53 constituencies, which “would require the drawing of new constituency boundaries, and losing ‘co-terminosity’ between Westminster and Assembly constituencies.”**
Scully’s preference is for STV, which would resolve the dual candidacy question by reverting to a single tier, while keeping the level of proportionality about the same (potentially). A commission proposed STV a decade ago. Scully notes that there have been two main proposals: grouping the current 40 constituencies into 20 pairs that each elect 4 assembly members or using local authority boundaries as districts (which, I assume, would mean district magnitude varying by municipality population).
As for the dual-candidacy issue, many readers of this blog will know my position. Dual candidacy is an essential feature of mixed-member systems, especially MMP systems, without which many of the main benefits of the system are unrealized. Sure, it does not affect proportionality, but the system also delivers benefits on the intra-party dimension, by encouraging more constituency focus of members elected from party lists than would be the case under pure PR.*** This benefit is likely lost if parties refrain from nominating their best personnel in districts where they are unsure of victory and instead nominate them only on the list. Thus the “legitimacy” problem of list members that underlies the charge against dual candidacy (“entering through the back door,” “zombies”, etc.) is actually made worse by eliminating dual candidacy and thus severing the constituency link of list candidates. The MMP Review in New Zealand extensively commented on this issue and came clearly down in favor of retaining the right to dual candidacy. Wales should do the same–if it retains MMP.
* In Wales it is called the Additional Member System (AMS). I very much dislike this name, as it treats the list-elected members as mere add-ons, rather than an integral (in fact, the decisive) component of the system. In fact, the name would fit better for the other main category, MMM (mixed-member majoritarian).
** I think “co-terminosity” is a new word for me. I like it.
*** And also without the direct intra-party competition of STV or OLPR, or the partisan incentive for “vote management” and “friends and family” voting/clientelism concerns that STV is especially prone to.