Report to the Government Office for London: electing the London Mayor and the London Assembly

1. Choosing a system for the London Mayor election involves a large number of considerations, and raises some novel challenges for all existing electoral systems. Plurality rule elections could produce a winning candidate with only minority support, and detract from the success of the office. Double ballot systems are unlikely to be implementable under British conditions. A majoritarian system, either the Alternative Vote or the Supplementary Vote, would offer the best method of election. 2. There would be significant advantages in using the Supplementary Vote method for electing the Mayor - retaining X voting; a simpler and more transparent counting system; a guarantee that the winner would emerge with a visible majority; no risk of initially low-placed candidates eventually emerging as the winner; and the system would be less likely than the Alternative Vote to encourage a large number of candidates to stand. 3. Electing the London Assembly in a proportional way poses some special problems because of its small size (24 to 32 members). Plurality rule elections should be ruled out, since they could easily produce strong one-party dominance and highly unproportional results. AV and SV elections would not ameliorate this problem. The Assembly should be elected by one of three proportional systems - the Additional Member System (AMS), list proportional representation (list PR), or the Single Transferable Vote (STV). 4. The best performing electoral systems for the Assembly in this study were: • A ‘classic’ AMS system with 16 local seats (each covering two London boroughs), and either four ‘quadrant’ areas or five constituencies used for electing the top-up members. • An AMS system with 14 local seats (some covering three boroughs), and all top-up members elected on a London-wide basis. Ballot papers would have to be carefully designed under this system. • A List PR system again using either the quadrant or five-constituencies schemas to elect all Assembly members. • An STV system using the five-constituency schema. Any one of these four systems would deliver stable and consistently proportional outcomes. The detailed formula to be used in allocating seats under each system should be carefully considered, along with the detailed drawing of boundaries for constituencies to best ensure fair apportionment of seats across areas. 5. A number of other combinations of systems and constituency patterns for electing the Assembly produced less satisfactory outcomes and are not recommended: • AMS using 16 local seats and six ‘partnership’ constituencies for top-up seats could produce less proportional outcomes, depending on the allocation formula being used. • List PR in a single London constituency would create information problems for voters because of the large number of candidates. • Some STV schemes with four or six constituencies might not produce proportional outcomes consistently or in an easily predictable way. 6. It will be very important to choose voting systems for the Mayor and the Assembly which are attractive to voters and operate consistently with each other (and with the electoral systems that London voters will be using for other bodies). SV and AMS systems are the most popular choices with voters, while STV is clearly disliked. The consistent combinations of Mayor and Assembly electoral systems would be: • SV for Mayor and AMS or List PR for the Assembly • AV for Mayor and STV for the Assembly. 7. The responses of political parties to the new voting systems will greatly affect the legitimacy of the systems with the public. There is a case for government intervention to guarantee the use of ‘one member one vote’ procedures in the selection of party candidates to stand for Mayor, and in the compilation of party lists under either the AMS or List PR systems