Nine general elections have been held under the electoral system outlined above since 1967. During that time, there have been regular changes of government. There has been widespread acceptance of the validity of the outcomes of the elections and Mauritius has established itself as a vigorous and stable parliamentary democracy.
However, the First-Past-the-Post system, while ensuring stability, has been criticized as insufficiently proportionate to the share of votes secured by the political parties. For example, in 1982, the Labour led alliance polled over 25.78% of the national vote, yet obtain no constituency seats in the Legislative Assembly. The allocation of seats pursuant to the Best Loser System gave it only 2 seats. In 1995, the MSM–RMM alliance garnered 19.7% of the national vote but obtained no constituency seats. It received no seats under the Best Loser System. In 2000, the main opposition party won 36.7% of the national vote but only 6 of the 62 elected seats; it received 2 additional seats under the BLS. In 2010, the main opposition party took 42.9% of the national vote but obtained only 29% of the FPTP seats and was allocated 2 additional seats through the BLS. These results show that the electoral system can produce parliamentary outcomes that are, at times, so dramatically misaligned to the share of votes cast for the respective parties that they call for alleviation. They also show that the Best Loser System, which was designed for a different purpose, is inadequate to perform this task.

Pre-independence concerns about the Best Loser System
The Best Loser System was conceived in the years before Independence. However, even then there were serious concerns about it. The 1957 London Agreement set out the principles for any proposed electoral system, to include voting methods that facilitate the development of voting on grounds of political principle and party. Transforming the Governor's power of nomination into a corrective to the voting system risked the "consecration of communal considerations in cons-titutional terms". Despite any advantages it would have in terms of 58

ensuring adequate representation, in 1964 the Constitutional Commissioner, Prof. De Smith, regarded it as an "evil nonetheless, and one which may seriously inhibit the growth of national consciousness over the years". Even the 1966 Banwell Commission thought it "undesirable" to require candidates to declare their communities at the time of nomination.

The case for change
Mauritius' fortunes have dramatically improved since the Best Loser System was introduced at Independence in 1968.
The case for reform of the Best Loser System has now been made by many voices from civil society, by constitutional experts, by the Supreme Court of Mauritius, the Judicial Committee of the Privy Council and the United Nations Human Rights Committee. In 1991, the Supreme Court sounded a note of concern about the Best Loser System being based on an out of date 1972 census. There has been considerable debate about the appropriateness of the constitutional classifications. In any event, constitutional entrenchment of ethnic groups and determining individuals' "ways of life", have all proved highly contentious. Moreover, cultural differences between groups are tending to reduce with the rise in social mobility and the emergence of common patterns of consumption, of material culture and attitudes.108
108 A Jahangeer-Chojoo, “From Minority to Mainstream Politics: the Case of Mauritius Muslims” (2010) J Soc Sci 121, 126
109 Dr Sithanen, 26 June 2013, L’Express
The Best Loser System depends for its validity on an accurate and up to date census of the number in the population of each constitutionally prescribed community. As the UNHRC has now observed, the System cannot be sustained as consistent with the International Covenant on Civil and Political Rights unless that data gathering resumes, something which the main political parties agree is unacceptable. Together, the views of the UN Human Rights Committee and the lack of desire to update the 1972 community-based population census, make clear that the, "ethnic balance in Parliament cannot be underpinned by constitutional design. It has to be driven by the political parties' natural strategies and their continued imperative to play what has been called the 'balancing game’”.109 59

The process of change
Consideration of reform of the system has been around for over a decade. The Sachs Commission 2001/2002 recommended reform by creating 30 additional pro-portional representation seats to be chosen on the basis of lists provided by parties receiving more than 10% of the national vote. The 2002 Select Committee suggested a form of implementation for the new seats. In 2011 Carcassonne recommended a complete proportional representation system. In 2012 Dr Sithanen proposed the "unreturned votes elect" formula to elect 20 proportional representation MPs by using the votes of candidates who had polled well but had not been elected.
Other safeguards in the system
As commentators such as Dr Sithanen have observed the Best Loser System is not the only or even the major safeguard for ensuring diversity in the National Assembly. This has been confirmed by all the constitutional experts consulted. The mechanisms include the need for political parties to select candidates so as to produce broad electoral appeal.
Hopes and aspirations for the future
An electoral system does not simply reflect voter preferences and existing divisions in society, it should also signal the values and preferences of the future. As Mauritius develops as a nation, it cannot continue to rely on a system based on old statistics, 42 years old, that no longer reflects the changes in society.

The hopes and aspirations of our people also include those of one half of our population who have historically been under-represented in our National Assembly and whose potential contribution to our political culture and society is not being maximised – women.
The parliamentary representation of women rose from just 5.7% in 1983 to 17.1% in 2005 and to 18.8% in 2010.110 Despite an acknowledgment of the under-representation, only modest advances have been made. All parties agree we should do better.
110 R Sithanen, “Roadmap for a better balance between stability and fairness in the voting formula” (January 2012) p 31
In addition to the above judicial decisions and statements and the pronouncement of the UNHCR, the Sachs Commission, the Carcassonne Report and the Sithanen Report all expressed the view that the Best Loser System had "outlived its 60

original purpose" and must be "sub-sumed" in a reformed electoral system.
In a paper published in 1999 two-respected Mauritian personalities, the late Sir Marc David, QC and Mr Pierre Dinan wrote "Le Best Loser System est archaïque et anachronique. II réclame de tout candidat aux élections générales de déclarer son appartenance ethnique alors que depuis 1982 la Constitution empêche toute référence à celle-ci lors des recensements démographiques."
When the BLS was introduced, it was meant to be a temporary device to last for three elections. It has now survived ten elections. Besides, apart from negating nation-building, it is a system with many inherent problems, as we have seen:

(1) It is anachronistic since it is based on an outdated census of 1972.

At the next general elections in 2015, it will be 43 years old.
In the meantime, the size of the voting population has changed in the various constituencies.
The ethnicity, on which the BLS is based, has also evolved. It is never static in any country.
To illustrate this point at the general elections of 1976 – the 1972 census was used.
But if the 1972 census were to be applied to the 1967 elections, the allocation of Best Loser seats would have been different both in terms of communities and individuals.

(2) The formula for the allocation of the BLS is very complex:

― It starts with the appropriate community which is under-represented;

― Then it has to look at the appropriate party with the appropriate community; and

― Then it has to ensure that the majority of the winning party or alliance is not reduced, as far as possible.

And the final result can be arbitrary, for example a small party with a low percentage of votes may secure a Best Loser seat while a larger party with a greater percentage of votes does not.
For example In 1995, the MSM-RMM with 19.7% of the votes got no BL seat while a very small party secured two such seats.

(3) The allocation of seats can be erratic.

For example, in the 1983 elections,
Mr Nawoor (MSM/LP) with 16.2% of the votes (1,684) was allocated a seat, while Mr Kasenally (MMM) with 47% of the votes (9,695) was not. 61

Yet both belonged to the same community that was under-represented. Similarly in the 1987 elections, Mr Finette (PMSD) with 47.4% of the votes (13,541) was allocated a seat. While, Mr Bérenger (MMM/MTD/FTS) with 48.7% of the votes (15,332) was not.
What is worse, in the case of Mr Bérenger, was when the seat had to be allocated to the General population – the community under-represented, Mr Bérenger was in the appropriate community but not to the appropriate party.
However, when another seat was to go to the MMM, then Mr Bérenger was in the appropriate party (MMM) but not in the appropriate community (General population) and instead Mr Peerun (MMM/MTD/FTS) who had polled 44% of the votes (12,999) was selected.
These examples speak for themselves.

(4) It can also appear illogical and irrational –

For example In the 1983 elections – there were 2 candidates belonging to the same community and to the same alliance – Mr Malherbes and Mr Candahoo. Mr Malherbes polled 44.9% of the votes and Mr Candahoo polled 43.3% of the votes.
Yet it was Mr Candahoo who was selected.
The reason was although they were both in the same alliance (MSM/LP/PMSD) they had different electoral symbols on their ballot papers.
Mr Malherbes had the PMSD symbol (the cockerel) while Mr Candahoo had the MSM/LP symbol (the sun and the key).
This was done to avoid three symbols on the ballot paper.
(5) The Constituencies in Mauritius do not have the same number of electors – they vary considerably.
Yet a candidate belonging to the same community and the same party but with fewer votes can be allocated a seat while the one with more votes not.
For example in the 1967 elections:
Mr Koenig (PMSD) obtained 6,851 votes
and Mr Rima (PMSD) obtained 6,596 votes.
Both belonged to the PMSD and to the same community, but Mr Rima stood in a smaller Constituency and therefore, had a higher percentage of votes.
Similarly in the 2000 general elections:
Mr Leopold obtained 7,732 votes in Rodrigues and his party had 1% of the national vote. 62

While, Mr Petit polled 14,626 votes (nearly double), he was not allocated a best loser seat.
(6) The BLS can also be unpredictable as it was in 1982, 1991 and 1995 – where only 4 Best Loser seats were allotted because the winning party did not have any unreturned eligible candidate. We ended up with a Parliament with 66 Members instead of 70.
Similarly in the 2010 general elections, only 7 Best Loser seats were allocated.
(7) There is a real possibility, and it has happened, that a candidate decides to choose a different community for different reasons.
For example in the 2010 elections, Mr Yeung Sik Yuen opted to stand as a member of the General Population although he could have chosen Sino-Mauritian.
Had he done so, he would not have been allocated a Best Loser seat.
In a previous judgment, Judge Seetulsingh had found that it is perfectly Constitutional for this to be done and he was in no position to decide on the way of life of any individual.
This raises serious questions about the enforceability of the BLS.
(8) In case of a vacancy of Best Loser seat, the next unreturned candidate could well have switched sides and therefore again this could pervert the system.
It is therefore, abundantly clear that there are now compelling reasons to subsume the Best Loser System into the new electoral reforms being proposed.