I shall call the first alternative Model C and the second one Model A as per the Sachs Report. Both agree on maintaining the current 62 FPTP and the 8 best loser seats. They also concur on introducing 30 PR seats with a threshold of 10% of votes to become eligible for PR seats. They differ on five issues as follows:

In view of its crucial importance in the debate, I shall deal with the fundamental difference in the allotment of PR seats between the two models. It determines how countries with a mixed system strike the balance between stability and fairness. I shall also shed some light on the Rodrigues syndrome, where a FPTP majority of 4 seats became 2 after PR allocation. I have deliberately kept the analysis and the electoral algorithm very simple so as to ease understanding for those not familiar with the theory and practice of electoral systems and the methodology used to measure stability and fairness.

Difference between 30 compensatory and 30 parallel PR seats

It should be clear that a minimum percentage of PR seats is required to mitigate the huge disproportionality between votes and seats in FPTP. After a thorough analysis, the Sachs Commission reached the conclusion that 30 PR seats are necessary if we maintain the 62 FPTP seats. It firmly rejected a proposal for 20 PR seats, as it would be too few to correct the FPTP anomalies. Worse, if these 20 seats are allotted on a parallel basis, it becomes mere token. As four of the five features of the mixed system proposed by the Select Committee favour stability and penalise fairness, it is plain that if the method of allotting the 30 PR seats also prioritises stability over fairness, then it would hardly give any consideration to fairness. Hence the specific recommendation of Sachs for 30 PR seats apportioned through a compensatory method. Sachs proposed a ratio of 67% FPTP and 33% PR seats as it suggested subsuming the 8 Best Losers in the 30 PR seats. By keeping the 8 best loser seats, the two proposals of the Select Committee disturbs the 67%-33% ratio by 3% in favour of stability. Both recommend 70% FPTP and 30% PR seats.

The table explains the huge difference between the two modes with different levels of PR seats. It also underscores the reasoning used by electoral experts to provide for 30 PR seats along a compensatory mode in cases where 70 seats are returned by FPTP. A lower figure than 30 and the use of a parallel system would be largely insufficient to ensure fairness. I have used the case of 60-0 as illustration as this is what we want to avoid. In all examples, Party A takes all FPTP seats with 55.1% of votes while Party gets no seat with 44.9%. Consider the compensatory mode (Model C) with three levels of seats. As Party B has been severely penalised in FPTP outcome, it receives all PR seats in the three cases. In normal elections (1967,1976, 1983 and 1987) both parties would obtain such PR seats. With the current FPTP formula, the unfairness, as measured by the difference between votes and seats is inordinate at 45% (100% of seats for 55% of votes and 0 seats for 45% of votes). A compensatory PR with 30 seats would lower that inequity to 12%. In fact when one adjusts for Rodrigues and Best Loser seats, the 12% rises to 14% (31% seats for 45% votes). As we do not want a full PR system, there is no proportionate relationship between votes and seats. We need to preserve a margin to guarantee stability. A maximum of 14 % is considered high in many countries (it is zero in South Africa and around 5% to 10% in most mature democracies). With 20 compensatory PR seats, the degree of inequality rises to 22% (23% of seats for 45% of votes), as there are not enough seats to mitigate FPTP perversions. Unsurprisingly the unfairness skyrockets to 32% (13% of seats for 45% of votes) with only 10 PR seats.

If these same seats are apportioned using a parallel mode (Model A), it becomes mere token, purely symbolic and very inequitable. With 30 seats, the unfairness is a huge 32% (13% of seats for 45% of votes). At 20 PR seats, it grows to 35% (10 % of seats for 45% of votes) while at 10; it escalates to 40% (5% of seats for 45 % of votes). While a disproportionality of 40% between votes and seats is better than the 45 % of FPTP, it certainly cannot be considered fair. As aptly put by Dr Chu yesterday in l?express, ?it highlights the impotence of parallel PR to tone down the excess of FPTP attribution? and ?it is akin to a dose of panadol to cure cancer.?

Continue: Finding the right mix between introducing fairness and keeping stability (2nd Part)