The Best Loser System (BLS) will be used exceptionally for the next general elections only and that also in the very specific circumstances where one or more returned members have decided not to disclose their community for the purpose of allocating the 8 additional seats under paragraph 5 of the First Schedule to the Constitution. Many questions have been raised on how the mechanism will work in practice. As I have proposed the formula to the Committee on electoral reform, I think it is in order that I explain the rationale underpinning it.

The transitional provision

The Electoral Commission appointed seven best losers in May 2010 following the General elections.

Clause 4 (2) (b) of the Bill states that ‘‘where a candidate has not declared his community and is returned as member, the ESC shall, for the sole purposes of determining the appropriate community and allocating additional seats, proceed on the basis of the average number of returned members belonging to each community at all general elections held since 1976’’.

Undoubtedly the formula is not an ideal one. It is a second best solution. However given the context, it is not possible to find a better alternative. The preferred answer would be to enact the full blown electoral reform itself and there would be absolutely no requirement to declare one’s community. The new dispensation contained in the electoral reform bill which the Committee is finalising will produce an inclusive Parliament using a different mechanism from the current BLS. However I understand the Bill will be introduced into the National Assembly after the next general elections.

The Committee had to address two seemingly opposing issues;

i) how to reconcile the fact that some believe there is no mandate to carry out such a comprehensive reform while providing an effective remedy to the problem raised by the disqualification of a candidate who does not declare his community, as enjoined by the UNHRC. There is also the case before the Supreme Court and the prospects of being censured by the Privy Council on this matter. The solution advocated is the transitional provision that would apply for the next general elections only;

ii) how to balance the necessity to offer an option to our citizens not to declare their  community to stand as candidate while making sure that the allocation of the Best Loser seats are not distorted by the decision of an elected candidate not to disclose his communal belonging.

It was not easy nuts to crack. Usually there is a requirement for two sets of statistics to allow the ESC to allocate the 8 Best Loser seats after the results of the 62 constituency seats are known. First, it needs the census that divides the population into four communities. We are still using the 1972 census to accomplish this task. All elections held since 1976 have had recourse to that census to allocate the BL seats. It gives the percentage of each of the four communities so as to ascertain whether any of them is underrepresented after the election of the 62 constituency candidates.

The statistics used for the four communities to allot the BL seats is as follows:

Second is the socio-demographic composition of the 62 elected candidates into the four communities. Based on these two statistics, the ESC allocates the first four BL seats to underrepresented communities irrespective of party while the second set of four seats is then distributed on the basis of both the appropriate party and the underrepresented communities with a view to redressing any imbalance caused by the allocation of the first four seats. In essence, it is impossible to distribute these 8 seats if one does not have either of these two statistics. One can immediately see the problem if an elected candidate has not declared his community. We had to find a workaround to deliver a transitional mechanism that reconciles two seeming contradictory elements. How to give candidates the option not to disclose their community while not violating the fundamental principle of alloting BL seats to ensure a fair and adequate representation of each of the four communities, which allotment itself depends on community declaration ?

The three options considered

Three options were thoroughly evaluated as follows

i) Option one is a simple and a stand alone replacement of the words ‘shall declare’ by ‘may declare’ in paragraph 3(1) of the First Schedule to the Constitution;

ii) Option two is to provide for the EC to assign a deemed community to an elected candidate who has not declared his community;

iii) Option three is to introduce an objective criteria to allow the ESC to proceed with the allocation of the BL seats in the event of an elected candidate not having declared his community.

Each of these three options were assessed according to five key parametres as follows :

a) Would it cross the legal hurdle ?

b) Can it be used by the ESC to allocate the 8 BL seats ?

c) Is the cure not worst than the disease ?

d) Is it politically correct and acceptable ?

e) Will it garner a three quarter majority in the National Assembly ?

1. Option one: a stand alone replacement of ‘‘shall declare’’ by ‘‘may declare’’

While the starting point of any solution is to replace the word ‘shall declare’ by ‘may declare’ so as to give an option to candidates not to declare their community, it is abundantly clear that , in itself and by itself, it cannot be an adequate answer to the problem. On a stand alone basis, the replacement of ‘‘shall declare’’ by ‘‘may declare’’ is fraught with unintended consequences as elected candidates who do not reveal their community will distort the allocation of the BL seats so that ‘‘the whole exercise will be stultified, thereby rendering nugatory the allocation of the eight additional seats’’ as well articulated in the landmark judgement of the Full Bench of the Supreme Court in 2005. The Full Bench also stated that ‘‘the declaration of community made by a prospective candidate at a general election is at the heart of the Best Loser System enshrined in the First Schedule since the allocation of the eight additional seats is to ensure a fair and adequate representation of each of the four communities’’.

A simple exercise would show the imperfection of this option as it may well allocate BL seats to ‘‘over represented communities’’ and deny such seats to ‘‘underrepresented communities’’, thus going against the provision for a fair and adequate representation of each of the four communities. That option was also open to legal challenges as opined by all constitutional experts consulted. It was therefore not retained.

2. Option two: The ESC will assign a community to an elected candidate

Option two attempts to address the points raised in the judgement of the Supreme Court. Inspite of its good intention in terms of attempting to ensure that the exercise of allocating the BL seats is not stultified or rendered nugatory, it creates a major perversion in giving power to an authority to determine the community of a person who has himself not made such a declaration.

That option recommended that ‘‘the Electoral Commissioner shall, in the light of such objective material as may be gathered from a public document or from such other document as he may consider reliable, determine the community to which an independent and reasonable observer would perceive the elected member who has not declared his community to belong and that elected member shall, for the sole purpose of the allocation of the Best Loser seats, be deemed to belong to the community which has been so determined’’.

While one can understand the rationale behind the proposal, it is clearly very objectionable. How can an authority decide on the community of an elected member when the member has decided by choice and conviction not to do so? I strongly objected to that proposal. Many other members also did the same.

Also, both the Prime Minister and the Leader of the Opposition categorically rejected that suggestion. It was set aside.

We still did not have a formula to balance the necessity to provide an option for the non declaration of community and the imperative to ensure that the allocation of the eight additional seats provides for a fair and adequate representation of all communities.

3. Option three: to use the average number of returned candidates of each community for the last 9 elections

This is the proposal contained in the transitional provision bill. Let me state soberly that it is not ideal but in the circumstances, it is the best of the three options contemplated.

We have to make an extremely important distinction between two completely different situations.

a) all elected candidates have declared their community

If all elected candidates have declared their communities, there will be no change in the current BL mechanism to allocate the 8 additional seats. Clause 4(2)(c) of the Bill spells out clearly that the existing formula for allocation of additional seats remains applicable if all returned candidates have declared their community. The ESC would thus have both relevant statistics to perform the statistical exercise of allocating the additional seats. It will use the population census and the sociodemographic mix of the 62 elected constituency candidates to distribute the BL seats. Exactly as has been the case since 1967.

Many candidates would be divided between their strong desire not to reveal their community and their sense of fairness not to distort the provision of our Constitution in the allocation of additional seats to underrepresented communities. This would be in addition to the political calculation of both candidates and parties not to deny themselves the right to these 8 seats that may prove crucial from an electoral perspective.

At this juncture we need to leave political shenanigans aside and be fair to acknowledge and recognize which communities have historically been underrepresented after the 62 constituency seats and have rightly benefitted from the BL seats. Even if the demographic profile of candidates fielded at elections has evolved in some specific constituencies since the first general elections of 1967, it is an established fact that the 8 BL seats have usually gone to two of the four communities in a ratio of around 70% to 30%. One particular community has never been alloted a BL seats while another obtained one BL seat in 1967 and two seats exceptionally in 2000 following the application of the 1992 amendment.

In view of the above and to avert any possible distortion, one hopes that all candidates will exceptionally and for a last time declare their community. I submit it is part of a social contract that the framers of our constitution have bequeathed to us and we have a responsibility to deliver on that expectation even if we find that declaration repugnant. Life

is always a balancing act and living peacefully in a multi-faith and multi- ethnic society comprises concession, tolerance and a sense of sharing and accommodation within a broad consociational compact.

b) one or more elected candidates have not declared their community

However we have to provide for the alternative scenario. Should an elected candidate decide not to reveal his community, we have an obligation to minimize the unintended consequences of such a decision on the allocation of the 8 BL seats. Basically to ensure that these seats are apportioned to underrepresented communities so that there is a fair and adequate representation of each of the four communities. This is where the formula comes into play.

As stated earlier on, the ESC needs both the population census and the community composition of the 62 elected members to proceed to the allocation of the 8 BL seats. If a returned candidate has no community, it represents a challenge as well stated by the Full Bench of the Supreme Court in 2005

‘‘Although an independent unreturned candidate has no claim to any one of the additional eight seats under the First Schedule, yet if he is elected, the declaration as to his community plays an important role in the determination of the eight additional seats. Indeed in order to determine as to which community a seat is to be allocated in turn, it is essential to ascertain the number of persons of that community who have been returned as successful candidates, irrespective of the party to which those persons belong as that number will be a component of the denominator by which the total number of persons comprising that community as per the 1972 census will be divided in order to ascertain the representation of that community in respect of each exercise for the purpose of allocating the additional seats.’’

Ten elections have been held since 1967 under the current FPTP formula supplemented by the BLS. We know exactly the community mix of the 62 Constituency MP’s and also to which of the four communities the BL seats have been allocated. The statistics are published after each election and are also publicly available on the web site of the Electoral Commision Office. Anybody can thus easily compute the socio-demographic mix of the 62 candidates elected at each of the general elections from 1967 to 2010.

It should be noted that in the case of BL seats, there are many elections when fewer than 8 seats were distributed. After its deliberations, the Committee has recommended an average of the community mix of elected candidates for the 9 elections contested between 1976 and 2010 to be used to allocate the BL seats only in the event that an elected candidate or more has not declared his community. It has excluded the 1967 elections as it was held under a different population census than the 1972 one.

And let us be candid. It is highly likely that the socio-demographic composition of the 62 returned candidates at the next general elections would be very close to that average. Whatever the political configuration, nobody expects the candidates to be aligned in most, if not all, constituencies to be markedly different . This formula has the merit of preventing the system from being abused with a view to distorting the distribution of the BL seats.

Five potential difficulties

It is not an ideal solution. But it is difficult to find a better alternative given the two constraints, mentioned earlier. The formula is not without problems. First, if the actual community configuration of the 62 elected Mps at the next elections is different from the one used in the Bill , there could be a slight problem in the allocation of the BL seat. This, of course, assumes that some would assign a slight community to the elected member or members who has not declared his community, as it can only happen in that specific case.

Second, my simulations and prospective analysis have lead me to conclude that an average of the last five elections would be a better fit to minimize the chances of a community imbalance in the allocation of the BL seats. I have argued in the committee that the average of the last five elections better reflects the ground realities as in some constituencies the fielding of candidates is different today from what it was in 1967,1976 and 1982. And that the deemed average of the last five elections is probably a better proxy for the outcome of the next elections. However the committee felt that there is no objective basis to decide on that average while there is a robust reason for using the last nine elections as they were all held using the 1972 population census. As the difference is marginal and could potentially affect the distribution of the 8th seat in some exceptional circumstances, I have accepted the consensus view.

Third, one unintended consequence of the provision is to create two categories of candidates. Those who declare their community have a further chance of entering the National Assembly through the BLS while those who have not done so, will be excluded. I am not happy about this unfairness. However it is inescapable as one cannot assign a community to someone who has, by choice, decided not to do so. Fourth is the possibility of the transitional provision staying on for a longer time if there is no consensus after the next elections to introduce the full blown reform bill that also requires a three quarter majority. And fifth, the much criticised 1972 population census will still be used to allocate BL seats for the next general elections. Again we hardly have any choice in the circumstances.

Concluding remarks

There is never such a thing as an ideal voting formula. The proposal of the transitional provision is certainly not perfect but it has the features to provide an overall balance between giving the choice to a candidate to declare his community and ensuring a fair and adequate representation of each of the four communities. Often one cannot choose a perfect system but has to settle for one that is significantly less imperfect than other formulae.

Of the three options considered, option three is the one that avoids serious shortcomings and offers a prudent approach. It is predictable and will likely deliver an allocation of the BL seats that is fair and adequate for each of the four communities. If all elected candidates declare their community, the BL mechanism will operate as it has since 1967. And should a returned candidate decide not to disclose his community, the formula ensures that there is hardly any distortion as possible in the allocation of the BL seats. I say hardly advisedly, because even today, the mechanism is an approximate one based on the population census of 1972. And equally, we have seen in 2010 how the current system is open to abuse. So overall, the risks are well hedged against a perversion of the mechanism. Until we have 1972 (%) the real reform !