Modifying Northern Ireland's present electoral system

David Chapman of the Democracy Design Forum has been arguing for many years that what is needed in Northern Ireland is an electoral system which produces a government that is responsive to both Protestants and Catholics.

To achieve this end he has recently come up with an innovative electoral system he entitles Consensus STV, as an alternative to his Minimum Percentage System (described in previous journals). (For this latter system, to recap, the province is divided into small electoral areas known as 'tracts', each tract having between 200 and 400 electors in it. The system gives a party seats according to its 'minimum percentage', ie its percentage of votes in that one-sixth of tracts in which it has its lowest percentages of votes. This gives it the incentive to get votes in all tracts.)

Chapman's new proposal is a modification of the single transferable vote (STV) system. Since STV has already been used in Northern Ireland, both for province-wide and for local elections, this second option would require much less of a radical change, and might therefore be politically more acceptable than the Minimum Percentage System. How it works is that candidates are divided into 'sets' of those similar in political tendency and sectarian allegiance. One candidate is then elected from each set by a balanced electorate, so that the candidates in a set have to compete for the votes of both Catholics and Protestants.

The Minimum Percentage System was focused upon the parties, being designed to induce each party to be pan-sectarian - to respond to each sectarian group, to the Protestants and to the Catholics. But it is also possible to focus on the individual candidates or MPs, and to design a system to affect their motivation. This is the approach is taken in Consensus STV, in modifying STV so as to provide each candidate with similar pan-ethnic incentives.

When ordinary STV was used in Northern Ireland in 1982 to elect an assembly, the result was, quite predictably, that nearly all seats went to monosectarian parties, a majority of seats going to Protestant parties, and a minority to Catholic parties. How then can STV be modified to avoid this outcome?

The suggestion has been made by Donald L. Horowitz (in 'A Democratic South Africa?', University of California Press, 1991) that, as a means of providing pan-sectarian or pan-ethnic motivation in such situations, the closely related alternative vote (AV) system should be used, but in multi-member constituencies, not in the single-member ones normally used with AV. In any such multi-member constituency, one candidate would be elected. The votes would then be reallocated each to the most-preferred of the remaining candidates, and another candidate would be elected, and so on, until all seats were filled.

Horowitz's expectation is that each seat would be contested by two candidates of the majority ethnic group, neither of whom would be able to get a majority on the basis of the votes of the majority ethnic group alone.

However, if this scheme were used in Northern Ireland in let us say six-member constituencies, it seems quite likely that the Protestant majority would give the same six Protestant candidates their highest preferences. If so, it would allow the Protestant majority, or still worse, the largest single Protestant party, to capture all the seats in many or most of these multi-member constituencies, thus exaggerating the Protestant domination, and giving no incentive towards pan-sectarianism.

I therefore put forward Consensus STV, which, though somewhat similar to Horowitz's suggestion, appears to be much more likely to promote pan-sectarianism. Its principle, to repeat, is this. Multi-member constituencies are used (let us say six members), and are drawn up so that each one has the same proportion of Catholics, and of Protestants, as does the province as a whole. Now in any one constituency, six sets each of two candidates are formed, the candidates in any one set being of similar political tendency and probably of the same sectarian group. One candidate is elected from each set, but to get elected, the candidate has to compete with the other member of the set for the votes of the whole constituency, and of each sectarian group in it. It is this intra-set competition between similar candidates for the votes of both Protestants and Catholics, which gives candidates the incentive to be pan-sectarian.

'The MPs as a whole once elected themselves would then proceed to elect one MP to be the prime minister'

The proposal is also that the MPs as a whole once elected themselves would then proceed to elect one MP to be the prime minister (using the Condorcet method of election). The prime minister would choose a government.

In summary, Consensus STV can be expected to produce an assembly of virtually independent MPs, each strongly responsive to his or her constituents, whether Protestant or CathoIic. The government which they support can be expected also to be pan-sectarian both in policy and membership, and in general to be responsive to the needs of all sections of the electorate.

David Chapman, Democratic Design Forum, Coles Centre, Buxhall, Stowmarket, Suffolk IP14 3EB (tel 0449 736 223).