Rama Sithanen propose de garder les objectifs du BLS mais d’en changer le mécanisme



L’ancien ministre des Finances et expert en systèmes électoraux propose d’englober les objectifs du Best Loser System (BLS) dans un mode d’élection mixte à scrutin majoritaire et à représentation proportionnelle.

Rama Sithanen a rendu public son rapport sur la réforme électorale. Lors d’une conférence de presse, ce samedi 28 janvier, à l’Hôtel Le Labourdonnais, à Port-Louis, l’ancien vice Premier ministre et ministre des Finances qui est aussi un expert en systèmes électoraux a présenté ses propositions pour un mode d’élections plus « juste et équitable ». Dans un document de 116 pages, intitulé « Initiative citoyenne pour une réforme électorale réalisable à Maurice », le chercheur  qui détient en un doctorat en sciences politiques cerne la problématique du système électoral mauricien, analyse les différentes pistes de recherche  et aligne des possibles solutions.

Lors de la présentation à la presse du travail qu’il a entrepris volontairement et bénévolement, Rama Sithanen a longuement insisté sur des réflexions autour du Best Loser System(BSL). Cet aspect du système électoral mauricien qui pourrait disparaître dans le cadre d’une réforme suscite beaucoup de débats en ce moment.

Analysant le BSL comme une ces caractéristiques du système électoral mauricien, Rama Sithanen concède que la question comporte une forte charge émotionnelle. Toutefois, il fait remarquer depuis 1967, 82% des députés appartenant à la communauté musulmane ont obtenu leurs sièges au Parlement en se faisant élire et non par le biais du BLS. Ce taux est de 78% pour la population générale. «  Vu le faible impact du BLS, on peut trouver un système alternatif de représentation qui soit moderne et crédible, mais qui maintiendra l’objectif du BSL », écrit le Dr Sithanen dans son rapport.

Ainsi, il propose d’englober les moyens d’atteindre les objectifs du BLS dans un système mixte. Celui-ci sera un mélange de scrutin majoritaire et de représentation proportionnelle avec une liste fermée.

 

 Voici le rapport dans son integralité

 

 

 

Initiative Citoyenne

pour une réforme électorale

réalisable à Maurice

 

 

 

 

 

 

Roadmap for a better balance between stability and fairness in the voting formula

 

 

                                                                                            Dr Rama Sithanen

January 2012                                                                                                                                 Quatre Bornes

 

 

                                                                                                 

                        

 

 

 

‘Most conflicts are two–sided.  The first alternative is my way, the second alternative is your way.  By synergising , we can go to a third alternative – our way, a higher and better way to resolve the conflict’.

Stephen R . Covey

 

 

 

 

 

 

 

 

 

 

Contents

                                                                                                                             Page

Preface and acknowledgements                                                                     vi

Report Methodology                                                                                     vii

List of Tables                                                                                               viii

List of Appendices                                                                                          x

List of Boxes                                                                                                  xi

About the Author                                                                                          xii

Executive Summary                                                                                     xiii

1.     Introduction : who we are, where we are and where we want to go     1

2.      Spain in the Indian Ocean :  has it worked there and will it deliver

here?                                                                                                    3

3.      Earthquake  in  paradise :  from   tranquil  efficacy  to vengeful

revolution                                                                                            6

4.      Avoiding some pitfalls : what’s mine  is  mine,  what’s  yours

is negotiable                                                                                       22

5.      Curing the defects of FPTP :  beware of unintended consequences    25

6.      Gender fairness : the discreet charm of a  much  simpler  solution    28

7.     Subsuming the Best Loser System:  promote inclusion, foster

nationhood                                                                                        34

8.    Party list : better, safer and more transparent than the BLS                    49

9.     How should  PR seats  be  allocated?  Pandemonium  in  heaven     54

10.    The ‘unreturned votes elect’ (UVE)  formula: a new kid in town      58

11.   How  much  stability,  how  much  fairness:  the  great  dilemma      60

12.    The  way  forward :  keep  FPTP  baby,   throw   away   bath

water of  its huge unfairness                                                              70

13.    Defining   some   concepts  in   the   new  electoral   landscape:

welcome to the party                                                                         75

14.    Remember some of Sachs :  forward with the past or back to

the future                                                                                           78

15.    Unreturned Votes Elect :  can the odd couple of stability and

fairness be reconciled?                                                                       82

 

16.    Between   FPTP  and  Sachs :  the   middle  is  better  than  the                 extremes     92

17.    Conclusion :  Have we reached the goal post?                                   95

Appendices                                                                                                  98             

Selected Bibliography                                                                                109             

List of persons and organisations met                                                        115

 

 

 

 

 

 

 

 

 

 

For Mauritius, forever

 

 

 

 

 

 

 

 

 

 

 

 

 

 


 

Preface and Acknowledgements

 

The genesis of this report lay in the realisation that in spite of many attempts to reform the electoral system of Mauritius, not much progress has been made. The rationale for reforming the FPTP voting system is compelling in view of the distortion that exists between votes polled and seats won at many elections and its amplification when a party or an alliance captures all seats on slightly more than 50 % of votes, thus leaving no Parliamentary representation to a huge share of the electorate. The political actors seem to have the will to implement such reforms and yet the topic continues to shift between the backburner and the centre of the agenda.

This report provides a framework for reaching an acceptable compromise among the various stakeholders on electoral reform.  My only aim is to make a modest and an informed contribution by enabling my fellow countrymen to understand and evaluate the complex subject of electoral reform in a multi ethnic society and the likely impact of changes in the voting formula. There has been very wide consultations and dialogue in order to understand the ideas and the proposals of various stakeholders. These have allowed me to acknowledge and appreciate the complexities of finding solutions to problems in a multi faith country. I have benefitted immensely from these interactions. My objective in some of these meetings was to test the findings, conclusions and recommendations of the report. The document has been reviewed by five persons ( four in Mauritius and one overseas) who collectively have a wealth of knowledge, experience and expertise on constitutional matters, on electoral issues, on the realities, history, context and circumstances of our country and on the mathematics of simulating likely election results based on different voting formulae. I owe them a huge debt of gratitude for their critical insight and suggestions. I am particularly grateful to Mr Irfan Rahman, Electoral Commissioner, for his guidance and support in providing the much needed election statistics to carry out the various simulations.  As is customary, the responsibility for errors and omissions remains mine.

 

Report Methodology

 

The following methodology has been used to produce this report:

(i)                 Review all the major reports written on the Constitutional and Electoral arrangements for Mauritius. These include the Trustram-Eve, Banwell, De Smith, Sachs, Collendavelloo and Carcassonne reports;

(ii)               Consult the archives of discussion among political actors in the run up to the choice of the electoral system in 1959 and 1967 and exchanges between various protagonists;

(iii)             Survey the comparative literature to understand how other multi faith and multi ethnic societies address the complex  problem of choosing an electoral system;

(iv)             Refine the findings and recommendations of the PhD thesis that I wrote on this subject;

(v)               Discuss with around 50 persons and organisations in order to have a good background of their views and aspirations. This ranges from political parties to women’s organisations, from constitutional experts to those who have written reports on electoral system in Mauritius, from civil society to members of the Judiciary;

(vi)             Simulate election results based on different voting formula to check the integrity and the  robustness of the electoral system being proposed;

(vii)           Test the findings, conclusions and recommendations with an informal focus group, that represents the diversity of Mauritius, through informed discussions;

(viii)         Request 5 well respected Mauritians to review the document and to incorporate their suggestions, where possible, before its release.


 

List of Tables

 

 1.      Outcome of elections in Spain 1979-2011

2.      Hypothetical results of elections held under Spain-like electoral system

3.      Gender unfairness in Mauritius: number of candidates and Members of Parliament

4.      Type of quota in countries with high percentage of women Members of Parliament

5.    Ratio of  Non  Best Loser  Members of Parliament to  Best Loser    Members  of  Parliament

6.     Difference between Parallel and Compensatory method of allocating PR  seats

7.     Stability versus fairness at a glance

8.    When Model C of Sachs poses no threat to stability

9.              Model C of Sachs is too proportional at the expense of stability

10.   The special case of the 1976 Elections

11.             The insignificance of Sachs’ Model A on fairness

12.   Votes of unreturned candidates as a percentage of total votes

13.    Results of 1987 general elections

14.   Seat allocation using votes of unreturned candidates (1987 with 20 Seats)

15.   Simulation of 1987 Elections with  UVE formula

16.   Simulation of 1982, 1995, 1991 and 2000 Elections with UVE formula

17.   Simulation of 1967, 1976, 1983, 2005, 2010 Elections with UVE formula     

18.   FPTP, Sachs C and UVE : which stability and which fairness?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

List of Appendices

 

Appendix 1:    Allocation of  PR seats in a compensatory mode using D’Hondt’s Highest Average Method

Appendix 2:    Apportionment of 20 additional UVE seats in 1967

Appendix 3     Apportionment of 20 additional UVE seats in 1976

Appendix 4:    Apportionment of 20 additional UVE seats in 1982

Appendix 5:    Apportionment of 20 additional UVE seats in 1983

Appendix 6:    Apportionment of 20 additional UVE seats in 1991

Appendix 7:    Apportionment of 20 additional UVE seats in 1995

Appendix 8:    Apportionment of 20 additional UVE seats in 2000

Appendix 9:    Apportionment of 20 additional UVE seats in 2005

Appendix 10: Apportionment of 20 additional UVE seats in 2010

Note            :    Apportionment for 1987 elections is in the text at Table 14

 

 

 

List of Boxes

 

Box 1:         A short history of electoral systems in Mauritius

Box 2:         How the BLS will be subsumed into the new mixed electoral system

 

Box 3:         Measuring unfairness in electoral systems

 

 

Box 4:         Some countries using closed Party List to return Mps

Box 5:         Some countries where double candidacy is very common

 

 

 

About the Author

 

Rama Sithanen has developed an interest for electoral systems in plural societies by accident rather than by design. Surprised by the huge disproportionality between votes polled and seats obtained in many elections held in Mauritius since Independence in 1967 under the FPTP system, he started to research constitutional arrangements and institutional designs, with an emphasis on the electoral system and the voting formula, of  plural societies, especially those that are cleaved along ethnic, racial, linguistic, cultural and religious lines, to accommodate diversity and ensure inclusion and fairness in political and parliamentary representation.

 

As part of that research endeavour, he read political science at Brunel University in the UK and obtained a PhD in Politics.  His thesis was entitled ‘Evaluating and Proposing Electoral Systems for Plural Societies: the case of Mauritius’. He has appraised the electoral systems of many deeply divided societies such as South Africa, Northern Ireland, Belgium, Sri Lanka, Lebanon, Fiji and Iraq, amongst others.

 

He has also studied Economics at the London School of Economics and Political Science and holds a BSc (Economics) with First Class Honours and an MSc (Economics) with a Mark of Distinction.

 

He was Minister of  Finance of Mauritius between 1991 and 1995 and Deputy Prime Minister and Minister of Finance and Economic Development between 2005 and 2010.

He is currently Chairman and Director of International Financial Services Ltd.

Executive Summary

 

  1. Electoral reform is back on the political agenda after the publication of the Carcassonne report which proposes a radical departure from most features of our voting system. The recommended model is a pure PR formula with variable multi-member constituencies and a closed Party List to return Mps.  It is very similar to what exists in Spain.

 

2.         Electoral systems that are successful in almost homogeneous countries might not necessarily work in plural societies. The history, political environment, context and specific circumstances of the country have a determining role in the choice of the voting formula. It also depends on what the country wants to achieve and what weights it assigns to each of the conditions for  a good voting formula. Electoral rules should not only be fashionable; they should also fit the circumstances and match the realities of the country. They cannot be bought off the peg; they must be made to measure.

 

3.       In spite of some novel ideas, the Carcassonne report contains many weaknesses :  radical changes to the current FPTP formula; absence of well defined  boundaries for the multi-member constituencies; risk that no clear majority will emerge; danger of not meeting the criteria of diversity and inclusiveness; complete disenfranchisement of the electorate; weakening of the link between elected representatives and their constituents. In attempting to cure one defect, the Carcassonne report has created too many practical problems and there are also unintended consequences.

4.         In a plural society like Mauritius, the six core values of an electoral system should be stability, fairness, inclusion, gender representation, accountability of Mps to their constituency and the shunning of essentially communal parties.  Our unique FPTP electoral rules have delivered on most of the core values of a good electoral system.  However its main drawback is the inequitable distribution of seats to votes.  The two principal tasks to be accomplished in a reform of the FPTP formula should be to remedy the high distortion between seats and votes and, in so doing, not to unduly undermine the stability of the system.

 

5.         The political representation of women is very low in Mauritius.  The principal reason for such a paltry presence of women in Parliament has more to do with the policies and practices of political parties than the electoral system itself.  There are simply too few women being selected as candidates.  To redress this deficiency, it is proposed to have no more than two candidates of the same sex in each of the twenty FPTP constituencies in Mauritius. And if there is a Party list in addition to FPTP, it should be constituted such that there is at least one person of a different gender out of every three sequential candidates . In the medium term, we should aim for gender parity  in Parliamentary representation as women have all the characteristics to be chosen as candidates and have demonstrated that they are at par with their male peers.  As is the practice in many mature democracies, the onus for ensuring such representation should rest with political parties, beefed up by strong guidelines from the Electoral Commission.


 

6.         The Mauritius electoral system contains four special features to guarantee the political representation of all the main segments of the population.  These are specially drawn constituencies, unequal distribution of population across electoral districts, three-member constituencies and the Best Loser System (BLS). While there is a historical and an emotional  character to the BLS, a compelling analysis shows that its impact is not substantive compared to the other three features that are used to secure adequate Parliamentary presence.  82 % of Mps of one particular ethnic group has entered Parliament over the last 10 elections since 1967 through the normal, non-BLS pathway. The corresponding figure for another ethnic group is 78 %.  For a third community, the number is a remarkable 100 %. In view of its low impact on representation, an alternative, credible and modern technique can easily be found to maintain the objective of the BLS while replacing its operating mechanism.

 

7.         The objective of the BLS can be subsumed into a new mixed FPTP and Party List electoral system through two sets of measures. First is the retention of three-member electoral constituencies, coupled with the maintenance of electoral boundaries and the acceptance of unequal distribution of votersn across constituencies.  Second is an additional tier of 20 Mps combined with a closed, ranked-based Party list to return these Mps and the provision for double candidacies.  Together, these six measures will inherently provide for socio-demographic inclusion while removing the communal masonry from our Constitution.  It is a win- win situation.

 

8.         The remaining challenge is how to strike the right balance between government stability and party fairness, especially as these two criteria very often move in opposite direction. The maximization of proportionality is not always a desirable end.

If we want to preserve stability which is a fundamental feature of an electoral system, we should avoid pure PR and choose broad proportionality instead.  There will still be some divergence between seats and votes but it will tend to be small and tolerable compared to the unfairness of the current voting mode.

 

9.        A uniquely designed electoral system based on our context, history and realities is recommended for a better balance between stability and fairness. It is a mixed system with a two tier electoral formula.  The first tier of 62 Mps will allocate seats nominally through a FPTP mode while the second tier will distribute seats by party lists. It will use, in respect of each party, the aggregated votes of all unreturned candidates from all constituencies to apportion the newly created 20 Party List seats, subject to a threshold for eligibility.  As a result, there will be a Parliament of 82 Mps.

 

10.       This mixed system of 62 FPTP Mps and 20 Party list Mps has the great merit of bringing down significantly the level of disproportionality between seats and votes while affording a working majority to the winner.  It will also compensate the second party in case of an alliance taking all 60 seats in Mauritius.

 

11.       Often a country cannot choose a perfect system but has to settle for one that is significantly less imperfect than other formulae.  The best voting system for a country is not one that satisfies only one criterion completely, but one that provides a fair balance among the different attributes.  A prudent approach is to design an electoral system that avoids serious shortcomings.

 

12.       The proposed electoral system is not a perfect one but it has the features to provide the most satisfactory overall balance between them in a multi-ethnic society, especially between stability and fairness.  It meets all the requirements of a good electoral system in a plural society.  Not all of them perfectly but very satisfactorily. It produces stability and effectiveness as it gives a working majority to the winning party even in hotly contested elections. It is fair and equitable as it narrows considerably the unacceptable distortion between seats and votes, especially for the second party.  It achieves diversity and inclusion in political representation by using an intelligent combination of the specific features of the FPTP and some well-crafted characteristics of an additional tier electoral mode.  It is significantly inclusive of women and it ensures accountability by keeping the essential link between constituents and their elected representatives.  It will also shun single issue parties, promote harmony and foster nationhood.

 

13.       Never has the country been so close to reaching an agreement on reforming the electoral system.  The necessary consensus is there and we need to build on that momentum to accelerate the process.  Reformers have engaged in a very long and patient exercise and the status quo does not appear to be a durable option anymore.  Today, the balance of probabilities appears to be on the side of reformers.  It is only a question of timing when reform will actually happen.  However, it crucially hinges on the willingness and the desire of political actors and the population at large to make some historic compromises in order to reach an acceptable and a sustainable consensus.

 

Initiative citoyenne pour une réforme électorale réalisable à Maurice

 

Roadmap for a better balance between stability   and fairness in the voting formula

 

 

1.     Introduction :  who we are, where we are and where we want to go

 

Electoral reform has raised its head again and has moved from the backburner to a high place on the political agenda after the publication of the Carcassonne report which proposes a radical change to most features of our electoral system. That Report alters the voting formula from a First Past The Post mode ( FPTP ) to a variable multi-member pure Proportional Representation one; it reduces the number of votes in each constituency from three candidates to one for a party. All Mps will be chosen by political parties in a closed, rank-based Party list as opposed to being elected through open choice by voters; the number of constituencies will be both reduced and redrawn. As to add spice to the debate, there are proposals to appoint up to 33 % of Ministers from outside Parliament and to have more Mps in case of defection by members.

Often attitudes on electoral systems in plural societies depend on who you are, where you are and where you want to go. We are a multi-ethnic, multi-faith and multi-cultural country that values fairness of representation and inclusion. However we acknowledge that modern society requires decisive and effective government to achieve social and economic progress. We also cherish the vital links and accountability between elected representatives and their constituents. We know where we are in terms of how the current electoral system has served the country since Independence, what are its strengths and its weaknesses. Above all we are at a crossroad where we want to consolidate the democratic foundations of our country and move resolutely towards enhanced nationhood. In carrying out reforms to our voting formula, there is a wish to avoid the dangers of unintended consequences while purporting to cure existing ones. While a broad measure of agreement exists to remedy the anomalies of the FPTP electoral system, there is divergence on what should constitute the ideal formula to replace the existing one. Both the Sachs and the Carcassonne reports must be assessed within the framework of a search for an electoral system best suited to the realities of a plural society.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2.     Spain in the Indian Ocean : Has it worked there and will it deliver here?

 

 

The  Carcassonne report comes almost ten years after the submission of the Sachs proposals. While there was near consensus on the adoption of a mixed system – a combination of FPTP and a dose of PR to correct the inequity of the FPTP voting formula- we now face the prospects of  a complete change of almost every characteristic of our electoral system. It is my considered view that the deficiency to be remedied in our voting system – essentially the inordinate disparity between seats and votes – does not warrant such a comprehensive rejection of the current system and the adoption of a totally different formula. There is absolutely no need to throw away the baby with the bath water. FPTP has significant intrinsic qualities that we should retain. What is required is a mechanism to mitigate its key drawback in certain circumstances in addition to finding an acceptable alternative for broad representation and inclusiveness.

 

The model proposed by Carcassonne is very similar to what exists in Spain. It is essentially a PR formula with variable multi-member constituencies and a closed Party list to return Mps. The objective, like in Spain, is to have a PR formula with some plurality features that would facilitate the creation of a majority to govern the country. It is important to acknowledge that a given voting formula will not necessarily work in the same way in different countries. Electoral systems that are successful in almost homogeneous societies like Spain might not necessarily work in plural countries like Mauritius. The history, political environment, context and specific circumstances of the country should have a determining role in the choice of the voting formula. An electoral system should not only be fashionable; it should also adequately respond to the circumstances and match the realities of any particular  country. Such a system cannot be bought off the peg; it must be made to measure.

 

Undoubtedly there are some interesting proposals in the report, even if many have  already been suggested by others before, such as the absolute necessity to forge nationhood, the enhanced representation of women, the subsuming of the Best Loser System into a new electoral dispensation and the proof of identity to cast one’s vote so as to avoid impersonation. Carcassonne and his two distinguished co-authors add their considerable weight and voice to an illustrious list of forward looking people who militate for fairness of representation, gender inclusion, diversity of Mps, a constitution and an electoral process that are devoid of communal stain and a curious mix of  ethnic and religious foundations.

 

However, the report contains many deficiencies, some due to an acute lack of understanding of Mauritian realities and others attributable to the unintended consequences of some basic recommendations. Unlike Trustram-Eve, de Smith, Banwell, Sachs and Collendavelloo who held a series of public meetings, received many written submissions and engaged in protracted debate and discussion with many political and civil society actors, there was hardly any consultation and dialogue this time. Not even a focus group research which is very common in such a sensitive exercise and which could have thrown light on many issues.  It must be pointed out that it is not the first time that a PR voting formula based on multi-member constituencies and a party list system is proposed.  It was suggested as far back as 1957 during the discussion that led to the London agreement.  The objective then was to achieve a compromise between governability and representation via a balanced ticket in multi-member constituencies. Some 55 years on, we still have to find an acceptable solution to reconcile stability with fairness of representation.  Interestingly, we have also held elections in Mauritius in variable multi-member constituencies between 1948 and 1958.  There were 19 elected legislators in 5 multi-member electoral districts with three returning 3 members each, one choosing 4 representatives and one electing 6 Mps.  It was however a restricted franchise as we did not have universal suffrage of one person one vote until the general elections of 1959.  Furthermore, S A de Smith who was appointed as Constitutional Commissioner in 1961 made an appraisal of some forms of PR;  however after intense debate and negotiations among the different political stakeholders, he made two proposals with respect to the voting formula. Both were mixed systems.  The first was a combination of FPTP and Limited Vote while the second was a mix of FPTP and a form of Proportional Representation in multi-member constituencies. In 1967, Banwell considered, amongst other electoral systems, some form of PR with 20 three-member constituencies using the D’Hondt highest average method to allot seats before the country finally adopted the current FPTP voting formula. No stone was left unturned in the search for the best electoral system for Mauritius.  Even as far back as 1956, many of our compatriots were extremely unhappy about the prospects of party elites alone deciding on who would be on the list and in which rank.  And half a century on, we face the same challenges.

 

 

 

 

3.     Earthquake in paradise : from tranquil efficacy to vengeful revolution

 

The greatest drawback of the Carcassonne report lies in the fact that in trying to cure one anomaly of the FPTP system, it creates many other problems. It is crucially important, while considering reforms, to acknowledge the disadvantages of alternative systems in addition to their presumed virtues.  Equally we must assess whether it is possible to avoid the drawbacks of the current system without introducing undesirable features and consequences in the new formula

 

First is the complete overhaul of an electoral system that has become familiar in Mauritius over decades and that has served the country reasonably well even if it has some defects.  The Carcassonne formula constitutes too radical a departure from what the Mauritian electorate and political actors know and what they are all used to.  There is hardly any country that has thrown away an existing system altogether and started with a new one, except probably in post conflict societies or else after a revolution.  Most electoral system changes that have occurred in the last two decades have built on what existed rather than jettison everything and take a leap in the dark. In 1993, New Zealand moved from FPTP to a mixed electoral system with both FPTP and PR.  Lesotho shifted, in 2001, from FPTP to a combination of FPTP and PR. In 1994 Japan changed from a semi proportional system to a mixed model with both FPTP and PR.  In 1998, the Jenkins Commission recommended a mixed member system with AV (which is a majoritarian mode of election) in replacement of FPTP in the UK and a dose of PR.

 

 

          

Mauritius has always held elections under FPTP system which it inherited from Britain, even if different versions have been adopted since the first elections of 1885 to the last one of 2010.   One can distinguish four electoral formulae that have been used , all based on the FPTP plurality system

From 1885 to 1947,the Council comprised 10 elected representatives out of a total of 28 (the remaining members were either nominated or ex officio and the Governor). There were nine constituencies, coinciding with the nine districts, with eight electing one member and one (Port Louis) returning two representatives.  It was however a very limited suffrage.  The overwhelming majority of the people of Mauritius did not have the right to vote.

From 1948 to 1958, we had a variable multi-member constituency mode with 19 elected representatives out of a total of 35 members.  The 19 legislators came from 5 multi-member electoral districts with  three  returning  3  members  each, one choosing 4 representatives (Port Louis) and one selecting 6 members (Plaines Wilhems and Black River).  The suffrage was extended but it was still not universal as embodied by the concept of ‘ one man one vote’;

From 1959 to 1966, the country was divided into 40 one member constituencies and election was again FPTP- based. Besides the 40 elected representatives, there were three ex officio members, the Speaker and a maximum of 12 nominated members appointed by the Governor.  Universal suffrage was introduced at the General Elections of 1959;

As from 1967 Mauritius has a Legislative Assembly of 70 members made up of 62 directly elected MPs and 8 Best Losers.  The Best Losers are intended to mitigate for relative ethnic underrepresentation.  The elections are held under the FPTP system with the 62 directly elected representatives coming from 20 three member constituencies in Mauritius and one 2 member district in Rodrigues.

 

 

Box 1:  A short history of electoral systems in Mauritius

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

In 1986, France switched from its traditional Two Round System to PR only to revert to the old system in the following election. Mauritius has always used FPTP to return Mps, albeit in different forms.  It would be difficult to win the hearts and minds of the people and politicians alike with a radical change.  Box 1 gives a short history of the electoral rules from 1888 to today.

 

Second is the absence of constituency boundaries. In elections that are based on geographical districts, the boundary delimitations are probably more important than the voting formula itself.  It explains why in 1958, the Trustram-Eve commission proposed both a new voting formula in 40 single members constituencies characterized by a detailed boundary of each of these forty constituencies.  The two must go hand in hand. Equally in 1966, when the Banwell Commission recommended a new voting formula based on 20 three-member and one two member constituencies, it also came up with the existing boundaries of the previous 40 constituencies now combined in the proposed 20 constituencies.  The Sachs Commission did not advocate any change in the number of constituencies even if it called for an adjustment in the low number of electors in some electoral districts. There was therefore no need for new boundaries across the country. Carcassonne is proposing between 10 to 12 constituencies compared to the existing 21 (including Rodrigues).  Yet we are left guessing on the boundaries of these new constituencies in mainland Mauritius. He assigns this extremely daunting and controversial exercise to an Electoral Commission which should be ‘independent, neutral and objective’ in undertaking this mission. This would be a thankless task as there would be deep division and bitter disagreement. On several occasions, the country has been unable to agree on small adjustments to be made to the existing 20 constituencies as per the provisions of the Constitution.  Equally there was an uproar recently when boundaries for Municipal elections were redrawn.  One then wonders how consensus could possibly be reached on four extremely sensitive and complex subjects: the reduction in the number of constituencies from 20 to between 10 and 12, the  new boundaries of each of these newly created constituencies, the size and make up of each of them and the number of Mps. Will the Electoral Commission follow the principles and practices adopted by Trustram-Eve, thus keeping a fine balance between urban and rural areas and the composition of these constituencies?  Will it also pursue the objectives of the Banwell Commission when it brought down 40 single member constituencies to 20 three member constituencies while maintaining a particular balance?  Or will it start on a new base altogether?

Third, there is the risk that no clear majority will emerge from the proposed formula.  While the choice of an odd number of Mps in most, if not all, variable member constituencies and the adoption of relatively small electoral districts (between 4 and 7 Mps) should help the two largest parties, it is not at all clear that a majority will be readily forthcoming. Both the theory and the evidence of elections demonstrate very clearly that the dynamics of elections, the behavior of parties and the expectations of voters vary considerably as the voting formula changes.  One of the most fundamental ideas in the study of electoral systems is that plurality modes favour the emergence of a two party (or two coalition) system while PR formulae encourage the formation of several political parties that generally reflect policy, ideology, or leadership differences within society.

 

There is usually a higher number of parties that compete in elections under pure PR than under FPTP.  The campaign itself is different as some small parties would concentrate their effort and resources in few constituencies where they have a stronghold. For instance in Belgium where elections are held in multi-member constituencies with PR, there are at least

10 parties represented in Parliament. When New Zealand switched from a FPTP to a mixed system with a strong dose of PR in 1996, there were more parties that entered Parliament.  In France and in the UK, it is easier for smaller parties to win seats for the European Parliament that use PR than for National elections that return Mps with a majoritarian voting formula. Paradoxically in France, the National Front usually wins seat for the European Parliament but not for the National Assembly.  A similar situation applies for the Nationalist UK Independence Party. France adopted a PR formula in one general election in 1986 before reverting to the Two Round System and it was the only time when the National Front secured Parliamentary representation.

 

An analysis of the political consequences of the electoral system of Spain shows that, while it was designed to constrain the number of parties (identical to what Carcassone is proposing), it has not altogether prevented the fragmentation of votes.  There are of course two major national parties (PSOE and PP), similar to Mauritius. However there are few small formations such as the United Left that have some support across the country.  And more importantly, there are some strong regional parties that compete in specific constituencies and are rewarded by the electoral formula as their votes are concentrated in some few electoral districts (the Catalan and Basque nationalists amongst others).  In some elections they wield considerable influence as they hold the balance of power.  For instance in the 2004 general elections, the third national party, United Left, obtained just two seats with over 800,000 votes which were dispersed across the country while the Catalan nationalist won ten seats with about the same number of votes which were concentrated in few districts.  The Socialists took 46.8  % of the seats with 43.2  % of the votes.  However as they had no majority of seats, they needed the support of the small regional parties to govern.

 


 

In essence, the Carcassone formula favours not only the large national parties but also the most successful regional parties while it penalizes small political parties that have some level of support across the country.  Eleven elections have been held in  Spain   since   the restoration of democracy in 1977.  Six out of these contests produced a minority government that necessitated the support of many small parties to govern, one was a tie at 175 seats (out of 350 Mps), three delivered a small majority of around 5 % to 6 % of seats (in Mauritius this would stand at around 4 to 5 seats in a Parliament of 80 members).  Only the election of 1982 yielded a relatively large majority of 54 seats.  These are illustrated at Table 1.

 

Table  1:  Outcome of elections in Spain 1979-2011

 

Year

Winner

Second Party

Outcome

 

 

Seats advantage for winner

% votes

% seats

% votes

% seats

1977

34.4

47.4

29.3

33.7

Minority Government

Negative

1979

34.8

48

30.4

34.6

Minority Government

Negative

1982

48.1

57.7

26.4

30.6

Large majority Government

54 seats out of 350

1986

44.1

52.6

26

30

Small majority Government

18 seats out of 350

1989

39.6

50

25.8

30.6

Tie

won 175 seats out of 350

1993

38.8

45.4

34.8

40.3

Minority Government

Negative

1996

38.8

44.6

37.6

40.3

Minority Government

Negative

2000

44.5

52.3

34.1

35.7

Small majority Government

16 seats out of 350

2004

42.6

46.9

37.7

42.3

Minority Government

Negative

2008

43.8

48.3

39.9

44

Minority Government

Negative

2011

44.6

53.1

28.7

31.4

Small majority Government

22 seats out of 350

 

The figures show that the winning party requires around 44 % of votes to be almost certain of   having a small majority to govern.  This depends crucially on the distribution of votes across all constituencies.  It is  quite  likely  that  this  level  of  vote  would  not  produce  a  majority  in   Mauritius  because  of   the concentration of votes of the two major parties in specific electoral districts and the performance of some small parties with relatively moderate  regional electoral enclaves.

The same phenomenon  could very well happen in Mauritius.   It may or may not occur immediately but it will almost certainly take shape in the medium to the long term after the new voting formula would have stabilized.  The ingredients for such eventuality already exist in our electoral system.   A detailed study of the 10 elections since 1967 shows that attempt at fragmenting the system has been, to a large extent, thwarted by the FPTP formula.  However with PR in multi- member constituency, there will be more parties in the starting block.  Of course, the two large parties will have an advantage.  However small regional parties could fragment the system and prevent the creation of a stable majority. Such examples abound in our electoral landscape.

 

In the North, Dulloo of the MMSM attracted 29 % of votes in Grand Baie/Poudre D’ Or in 1995 while Busgeeth took 11.2 % of votes in Piton/Rivière Du Rempart in 2010.  In the East, Dayal of MDN captured  38 % of the votes in Montagne-Blanche/Grand River South East in 2000, while Hawaldar of PSP earned 14.8 % of votes in Flacq/Bon Accueil in 1976.  In the South, Bissondoyal of the IFB won 22 % of votes in Vieux Grand Port/Rose Belle while Molaye had 12 % of votes in Rivière des Anguilles/Souillac in 1976, AhYan of FCL polled 10 % of votes in Mahebourg/Plaine Magnien in 2005 and 2010, Ramsahok of PAL took 12.7 % of votes in that same constituency in 1995 while Lutchmeenariadoo of MDL garnered 16 % of votes in Rivière des Anguilles/Souillac in that same elections.  Gaetan Duval formed a party on the eve of the 1995 elections and obtained a relatively good share of votes in 5 urban constituencies, including 4 in Plaines Wilhems. Gaetan Duval of PGD obtained 38 % of the votes in Curepipe/Midlands in 1995.  The PMSD usually does well in many urban ridings where historically it has a stronghold. And the most recent phenomenon is the rise of the FSM of Meeah in Port-Louis Maritime/Port-Louis East and Port Louis South/Port Louis Central.  All these small parties could win one seat in relatively large multi-member constituencies. This would deprive the winning party of the odd seat required to have an edge in the constituency.  For instance, if constituencies 2 and 3 were merged and there were 5 Mps, it is clear that FSM would take at least one seat with a vote tally of around 15 % under a PR format.  And neither the Labour Party nor the MMM would have a majority in that constituency, even with an odd number of seats.

The same pattern could happen in other constituencies. Better these small parties could enter into a tactical alliance asking their supporters to vote for some specific parties in constituencies where they do not field candidates. Of course there could be other possibilities.  Two small parties that complement each other (say the MSM in rural areas and the PMSD in urban districts) could join forces and create a third formation that would deprive either the LP or the MMM of a majority in most, if not all, the newly created constituencies. This would be very similar to what happened in the 1976 general elections. In addition to these regional parties, there is also the MSM which has some support across most, if not all, constituencies. The party landscape is almost a carbon copy of what exists in Spain. The MSM, on its own, could potentially win few seats in some specific constituencies with around 11 % of the votes, even if its national average is much lower.  It attracted 19 % of votes in 1995.  It depends on the distribution of votes across the country.  If it is low and evenly spread across the constituencies, its chances are low. However if it is concentrated in few electoral districts (as appear to be the case), then it could win few seats.  The same logic would apply to the PMSD with relatively moderate support in some urban districts.   It did win some  moderate support in the 1976 and 1982 elections.

 

The fragmentation of votes among many political parties will make it difficult for the electorate to have a clear choice for a team to govern the country.  Carcassonne implicitly recognizes this problem and attempts to mitigate it by recommending a raft of measures to insulate the system against such weaknesses.  Mps could vote even when they are not physically present in the Assembly, new members would be appointed when an MP defects from his/her party, the procedure to dislodge a sitting Prime Minister becomes more difficult with a ‘constructive vote of no confidence’ and the PM keeps the threat of early dissolution as a strong deterrent.  However all these may be redundant if, after the elections, there is no clear majority that emerges because of vote fragmentation.  This may destabilise the political system, resulting in weak, ineffective and unstable government.

 

An extremely bold attempt is made to test the Carcassone model using some realistic assumptions about constituency boundaries, geographical support of the two main parties and the likelihood of a third political formation fragmenting votes in few constituencies. Following the rationale adopted by the Trustram-Eve and Banwell reports, eleven constituencies are constituted from the existing twenty (plus one for Rodrigues).

 

The main highly realistic assumptions underpinning the hypothetical outcome are as follows:

(i)                 There are 11 variable member constituencies.  Six have seven Mps, four choose five members and one elects six Mps and there are two elected representatives in Rodrigues, making a total of 70 seats.  The size of  the constituency varies according   to the population ratio;

 

(ii)               The existing 20 constituencies are collapsed into 11 new ones, following a particular pattern.  Some combinations are very obvious (Grand River North West/Port Louis West and Beau Basin/Petite Rivière, Port Louis South/Port Louis Central and Port Louis Maritime and Port Louis East, Grand Bay/Poudre D’Or and Piton/Rivière du Rempart, Vieux Grand Port/Rose Belle and Mahebourg/Plaine Magnien, Vacoas/Floreal   and    Curepipe/Midlands   and    Belle Rose/Quatre Bornes   and   La Caverne/Phoenix) while others will require some subtle geographical and demographic engineering;

 

(iii)             Party A has a majority of votes in five constituencies while Party B has an edge in three electoral districts.  The contest in the remaining three constituencies is close.

 

(iv)             Party C is an aggregation of residual parties and encompasses many possibilities.  From a combination of many regionally-based parties to one national party with some support across constituencies. For instance, it could represent FSM in Port Louis,  MMSM of Dulloo in the North, MDN of Dayal or IFB in the East, Ah Yan type civil society movement or IFB in the South and PMSD/PGD in Plaines Wilhems.  All have some regional strongholds.  Or alternatively, it could stand for a party like the MSM.  Or both at the same time.  The determining factor is the high probability of some degree of fragmentation away from the hold of the two major parties on the system.  Not necessarily immediately but certainly over time. In Spain, a party that polls around 44 % of the votes is likely to have a small working majority.  This is quite unlikely in Mauritius because of the dynamics of the elections in the 11 constituencies.  At 44 %, the leading party will probably have a plurality of seats (around 32 to 34 out of 70) but will not have an absolute majority, especially with the two seats from Rodrigues not contested by the two largest national parties.  Under these circumstances, the combined forces of the Opposition (including the two Mps from Rodrigues) may be higher than that of the party coming first in the electoral contest.

 

Table 2 illustrates the difficulties of having a stable and a strong government with the Spanish-like model proposed by Carcassonne, adapted to the local context.

Table 2:    Hypothetical  results  of  elections  held  under  Spain-like electoral system

 

 

Constituency

Seats

A

B

C

% votes

Seats

% votes

seats

% votes

Seats

1

7

41

3

56

4

2

0

2

5

40

2

39

2

18

1

3

7

54

4

39

3

4

0

4

7

53

4

33

2

12

1

5

7

55

4

32

2

12

1

6

5

53

3

43

2

3

0

7

6

54

4

40

2

5

0

8

5

49

3

47

2

2

0

9

5

42

2

53

3

3

0

10

7

47

3

51

4

1

0

11

7

45

3

42

3

12

1

Rodrigues

2

2

Total seats

70

% vote

49

43

7

Seats

35

29

6

% seats

50

41.5

8.5

 

Party A attracts above 50 % of votes in five constituencies and between 40 % and 49 % in the other electoral districts.  Party B captures over 50 % of votes in three ridings, between 40 % and 47 % in four districts and between 32 % and 39 % in another four constituencies.   Party C which combines many parties has some relatively moderate support in 4 constituencies only.  Using the D’Hondt formula to apportion seats at constituency level, the overall results show Party A with 35 seats, Party B with 29 Mps and Party C with 4 elected representatives. If the two Mps from Rodrigues (likely to be from two different parties because of the proportional system) are not counted as being on the side of Party A, there is no majority, even with 49 % of national votes.  Of course small variation in terms of votes could change the results slightly. But the impact would be marginal in terms of the overall outcome. And the above simulation is most probably a best case scenario. With less than 45 % of votes, the system becomes more fragmented. While it is highly proportional (Party A has 50 % of seats with 49 % of votes while Party B captures 41.5 % of seats with 43 % of votes), it would not produce a working majority to govern the country.   Post-election negotiation will be required to form a coalition government.

 

Representative government must not only represent, it must also govern. Simulations show that the number of seats is probably too low to have a clear winner from the 11 new constituencies on mainland Mauritius.  In a constituency where there are 6 elected MPs, a party with 57 % of votes and another one with 43 % may each win 3 seats.  While the proposal to have an odd number of seats may help, overall it could be nullified by the existence of party strongholds.  For instance, the one seat advantage of the Labour Party in the North could be wiped out by one additional seat for the MMM in Lower or Upper Plaines Wilhems.  Even at 45 % or more of the votes, the majority is extremely thin while in some cases, government formation will depend on the two Mps from Rodrigues.   The proposal to have a PR mode of election with two seats only for Rodrigues is difficult to understand.   A party with 66 % of the vote will win one seat and so will another with only 34 % of votes.  On the other hand, the larger the constituency, the higher the likelihood of party fragmentation!  It is a difficult balance to maintain.

 

Fourth, there is the danger of not meeting the diversity/plurality criteria at the constituency level. Under PR it is easier for some small groups to obtain political representation.  However the low level of seats in many constituencies, the tendency of the two largest parties to field candidates with more or less the same profile in many districts and the likelihood that the rank of a candidate will vary directly with the share of voters with similar characteristics to him/her will combine to deny representation to some components of the population.  There is a positive correlation between large-district PR electoral systems and the success of small groups in securing Parliamentary representation.  In large multi-member PR districts, it is possible for small sections of society to form part of a balanced ticket.  This is similar to the rationale behind three member constituencies as they are more likely to return members of small groups than single member districts.  At times, the constituencies also are ‘gerrymandered’ to produce a Parliament that is a microcosm of society.  Simulations show that with a low number of seats, it would be extremely difficult to have a diverse and balanced slate from each of the newly proposed constituencies. While the system could attain the objective of equity in terms of party representation at the national level, it may not be inclusive of the rainbowness of the nation.

 

Fifth, the electorate is disenfranchised as it is party elites that decide who goes on the list, in what order, who is elected, and who forms the Government.  While it is possible to agree to a percentage of Mps being returned through a closed Party List (this will be important to subsume the BLS into the new system), it is highly questionable when 100 % of Mps are chosen in that manner.  It means that however people vote, it is the party’s choice that is likely to prevail.  Political power is thus removed from voters and entrusted to the party apparatus.  As such it is very ineffective in disciplining politicians as voters do not have a say in the composition of Parliament.  Leaders become even more powerful as they have enormous influence in choosing candidates for the PR list. Internal party debate and discussion become much more difficult.  MPs lose their independence as their fortunes lie in the hands of party elites.  Their role is thus restricted to the whims of party leadership and party machinery, with little freedom of action for legislative work and constituency services.

Sixth is a weakening of the important link between elected representatives and their constituents as it exists in a FPTP system.  As Mps will depend on the party and not on the electorate to be elected, they may lose interest in constituency work and not respond to the specific needs and wishes of voters. This may pose a major problem of political responsiveness and accountability.  In a closed Party List system, the voter cannot express his/her dissatisfaction through any means but to refuse to vote for the party. Important local and community issues cannot be well articulated as in a plurality mode of election.  Elected representatives know very well that electors can punish non-performers by voting them out of office at the next electoral contest.  In marginal seats, sitting MPs tend to perform more constituency services as these could improve their prospects for re-election.  This link will weaken in the new PR mode and this is not good for the consolidation of democracy.  While a degree of choice by Party elites may be necessary to reassure some components of our society about inclusiveness of representation, it is a major concern when all Mps are returned on that basis.

 

Seventh, is the relatively high probability of small parties wielding power that is disproportionate to their electoral strength.  A fragmentation of the political system requires many parties to form a government.  For example, if one large party captures 42 % of the seats and a second one takes another 40 %, a small party with 11 % of the seats could hold the balance of power, thus becoming the kingmaker.  Under these circumstances, it is the third party and not the voters which decides who will govern the country.  It may even exert undue power and influence in the governing coalition by seeking major concessions from their senior partner on many policies. And ‘this tail wagging the dog’ situation may not be supported by a large majority of voters. The small parties can even bring down Governments by shifting political allegiance between elections, as the FDP did in 1982 in Germany.  PR and coalition government could allow small parties much greater representation and input in the political process than the electorate has expressed. In some instances, the second and third parties could come together in a coalition and deprive the most popular party of a place in Government.  It could also encourage corrupt practices to form Government and to stay in office.

 

Eighth is the temptation to balkanize the political system with the proliferation of parties. When there are multiple social cleavages, and social groups are structured into distinct organisations, the use of a pure PR in multi member constituencies may aggravate the fragmentation and political differences will be magnified. It will thus further entrench divisions and encourage the emergence and legitimizing of single issue parties.  Each party believes it can win some representation by appealing to its group only.  This could lead to polarized pluralism where parties stick to their social groups and compromise becomes very difficult.  If voting continues to be on ethnic considerations, PR can entrench political commitment along communal lines. To some extent, this is what happened after the general elections in Belgium in 2010 when the country became bitterly divided along linguistic and cultural cleavages.

 

Ninth, is the high probability of a gap between the election result and the composition of the government?. Government formation often takes a long time after the results of elections under PR.  In FPTP, on the very night of the results, the population usually  knows who will be in power and who will lead the Opposition. Under PR, Government formation depends very much on post-election discussions and negotiations in secret meetings between party leaders. It may not be the direct outcome of a popular mandate.  Belgium was without a Government for almost 16 months after the 2010 elections.  It uses a pure PR voting formula and returns Mps in variable multi-member constituencies.

Tenth, is the bias of the formula against small and even medium parties that have broad but shallow national support compared to those with strong regional support.  A small party with 10 % of national vote may not win a single seat if its support is evenly spread across the country.  However one party that has a stronghold in one or two constituencies (of around 15 % in one district) but no overall following across the country could gain Parliamentary representation.  The fortunes of small and midsized political parties will hinge on the performance of the big parties.   In a constituency with 4 seats, the threshold for eligibility to one seat could be as high as 18  %.   It may come down to around 14 to 15 % for five seats and 10 to 11 % for seven seats.  In a hotly contested election with  no party winning more than 35  % of the votes, it could be possible for a small party to squeak through with one seat with 10 % of the votes.

 

Eleventh, the proposal to handle defection of PR Mps is pregnant with difficulties.  In FPTP system, voters elect individuals while in closed PR mode, it is the party that is chosen.  The right of the individual Mp returned on a Party List to stay in Parliament after leaving his/her party is extremely tenuous as the vote is specifically for the party and not for the individual.  This is why the law, in many countries where there is a closed Party list, requires the member to vacate his/her position and to be replaced by someone else on the list.  However the rules and procedure for such replacement must be very clear to avoid any abuse by Party elites, especially with respect to independent minded Mps.  The suggestion to replace the defector by someone else could potentially lead to a huge increase in the number of MPs between elections.  Also this provision could encourage abuse and corrupt practices.  Thirty five years on, one still does not know why a duly elected Mp in 1976 resigned a couple of days after election to be replaced by someone who had lost his seat.  In the new system, the temptation is higher as the defector does not even sacrifice his seat!

 


 

4.    Avoiding some pitfalls: what’s mine is mine, what’s yours  is negotiable

 

Right from the outset, it should be clear that electoral rules are not neutral in their translation of votes into seats. There is nothing automatic about the way in which votes are converted into seats and different electoral systems will perform this function in different ways. Often the electoral formula will influence the political colour of a country’s Government, the relative strength of the various parties in Parliament and very likely the identity of the Prime Minister. The electoral system may even influence the way people vote.

 

Proposals for reforms often suffer from three main drawbacks. First, parties often favor the status quo from which they have benefitted and/or are likely to benefit .They could also introduce changes either to increase their chances of success or to prevent their opponents from winning. Second, political actors, however well intentioned, may not always have all the knowledge and the information to fully grasp all the consequences and ramifications of different electoral systems. Unfortunately, some may even have the tendency to believe what they know best is what is best for the country or what is best for their party is also best for the country. Third, electoral reforms are rarely made in a political vacuum. It depends on contextual and temporal factors.

 

Political parties often embrace strategies that would maximize their chances of converting votes into seats in Parliament, especially in the immediate term. Some often change their position on electoral reforms depending on how the system has treated their electoral fortunes. This is true in many countries and the  most glaring example is probably from the UK. When the Liberal Party was in power at the turn of the 20th century, it was opposed to PR as it benefitted from the FPTP process; today as it is severely penalised by the current FPTP electoral system, it is the most articulate proponent of PR. When the Conservative Party lost the General Elections of  1974, it proposed a review of the system. However it quickly changed its position as the FPTP kept it continuously in power for 18 years from 1979 to 1997. During that period it was the Labour Party that fought for electoral reform as it was disadvantaged by the formula, while between 1945 and 1979 they were quite hostile to reforms when they were often in office. As a consequence, the Jenkins report, which was commissioned by the Labour Government that came to power in 1997 to make recommendations on electoral reform in the UK, stated that this attitude has prompted

‘the cynical thought that there has been an element of the devil was sick, the devil a monk would be, the devil was well, the devil a devil he’d be…. about the attitude of all parties to electoral reform’.

 

It is all too often about realpolitik and party partisanship. A good reform should avoid these pitfalls. Some politicians are often tempted to use their knowledge to design electoral systems to promote formulae which they think will give them a partisan advantage. This could lead to choices that are not the best ones for the long-term political development of the country with potentially far reaching consequences for the democratic process. As the choice of electoral systems is one of the most important institutional decisions for any democracy, it is important to understand the need for reforms, grasp the consequences of each proposal from a normative standpoint, draw on the experience of other countries and build the necessary consensus for implementation. The worst that could happen is an attempt to satisfy the short term objectives of some political parties. This occured in France when the Socialists changed the electoral system for the 1986 Legislative Elections from a Two Round System to a PR mode, some months before the election in order to gain an unfair advantage. In the process they let in the right wing extremist National Front with 35 Mps, as the purpose of the new electoral formula was to divide the votes of the Right to ensure victory for the Left. Not only did the Socialists lose, but the “partisan” formula was changed very quickly and the following election in 1988 was held under the old Two Round Formula. Worse, that political shenanigan of the Left started the rise of the Extreme Right, which culminated in the defeat of the Socialist candidate, Lionel Jospin, in the first round of the 2002 Presidential election. What an irony!. This is a  classical example of perceived short term gain for long term atrocious pain. In fact it was not even an immediate gain as the Left lost the 1986 Elections. Hoist with its own petard!

 

.A partisan approach would lead to frequent changes in the electoral system as new governments attempt to concoct recipes to their advantages. This occurred in Mexico where the five elections between 1985 and 1997 were held under five different electoral systems! The one-sided approach led to constant changes in the electoral formula until the adoption of a consensual mixed member model in 1997. It should be abundantly plain that nobody could contemplate

 

 “tribal politics being allowed to choose a new voting system”

 

as eloquently stated by a Labour MP in the UK. As the choice of a particular electoral system has far reaching consequences on the future political life of the country and as, once chosen, it often stays in force for a very long period, we cannot afford a half-baked or a partisan solution. We need to be alert to the unintended consequences of reforms as, if conducted for short term partisan benefit, a country may fare worse than what it bargained for. It is of utmost importance for the stability and the progress of the country that we build consensus, seek legitimacy and secure wide acceptance for a change in the electoral formula.  The new system should also be robust and sustainable.

5.     Curing the defects of FPTP: beware of unintended consequences

 

Which electoral system is best depends on just what we want it to do.  In many countries the electoral system itself is a major political issue.  Some would like it changed while others vigorously defend the status quo.  It all depends on the criteria employed to evaluate an electoral system.

 

It is my considered view that in a plural society like Mauritius we need to focus on six core values as follows:

 

(i)                 government stability: the electoral system must provide for stable and strong  government that is capable of taking decisive action;

 

(ii)               party fairness: there must be a fair distribution of seats to votes to political parties even if it does not have to be a scrupulous mirror image;

 

(iii)             broad based socio demographic inclusion: various segments of the population must secure adequate political representation;

 

(iv)             gender representation: the system should encourage the involvement of women in the political process and their presence in Parliament;

 

(v)               accountability: the electoral system should maintain and indeed strengthen the linkages between MPs and their constituents and encourage  political responsiveness;

 

(vi)             avoid communal parties: the voting formula should not exacerbate divisions in a multi-ethnic society.

 

It is almost impossible for one electoral formula to satisfy all these attributes.  In fact some of them are mutually exclusive and often it is possible to achieve one particular objective only at the expense of another.  As a result an electoral system must balance various objectives and values. Specialists differ on which criteria matters most when choosing an appropriate electoral system as the exercise involves many trade-offs and a careful balancing act. That is why there are so many electoral systems around the world.  Each country has its own sets of values and applies them in designing its electoral formula.

 

Our unique FPTP electoral rules (three member constituency with BLS) have delivered on most of the core values of a good electoral system. They have performed well or very satisfactorily on stability, inclusion and accountability. We have also largely avoided communal parties.  As shall be shown later, the low level of female representation is not an inbuilt characteristic of the voting formula.  It is due to other factors.  The main drawback of FPTP is the inequitable distribution of seats to votes in certain elections.

 

It seems to me that the two principal tasks to be accomplished in a reform of the FPTP system ought to be as follows:

 

(i)                 to cure  the very high disproportionality between seats won and votes polled which leads to significant underrepresentation of some parties. And to remedy the amplification of ‘winner takes all’, where one alliance could obtain 100 % of seats with around 51 % of votes, thus leaving no elected representation at all to other parties with a relatively high share of  votes.  This can be  done by introducing a greater degree of fairness, measured by a better distribution between seats and votes, in our electoral system;

 

(ii)               to ensure, in so doing, that the stability of the system is not undermined.  It is assessed by the need to secure effective and stable government.  This is important as very often the objective of greater fairness can only be achieved at the expense of stability.

 

However in the process, two other issues have emerged and require solutions.

 

(i)                 First is gender representation: while it is an extremely important issue, one must acknowledge that it is has nothing to do with the reform process itself as the lack of women representation is not necessarily an intrinsic weakness of our three member FPTP voting formula;

 

(ii)               Second is the future of the Best Loser System: The aim of reforming the BLS is certainly not its pure and simple abolition but rather to find a viable alternative that keeps its objective of ensuring that all sections of the population are adequately represented in Parliament. However this will be achieved by using a different and a modern mechanism that will reassure the communities concerned about broad-based representation while removing its unacceptable features.

 

I shall deal with each of these two issues before turning to the crucial debate between fairness and stability.

 

6.    Gender fairness : The discreet charm of a much simpler solution

 

The FPTP electoral system of Mauritius cannot, on its own, explain the low representation of women. While it is true that a single member constituency (as in the UK, Canada and India)  makes it very difficult for women to be chosen as candidates, this need not be the case in our specifically-designed three member constituencies that encourage a balanced slate of candidates.  The principal reason for the low representation of women has more to do with the policies and practices of political parties rather than the voting formula. For the 10 elections in Mauritius between 1967 and 2010, the number of women fielded as candidates by the two main parties (three in 1976,1982 and 1995) stood at 88 out of 1380 candidates, representing only 6.4 % of total candidates.  It started extremely low in the early elections and has grown over the years, even if it stays very low compared to the size of our women population. If women are not selected as candidates, they simply cannot be elected to Parliament. Unsurprisingly, the average Parliamentary representation of women was very low at 7.4 % for these ten elections.  As shown at Table 3 below, it ranged from 0 % in 1967 to 9 % in 1995, 17 % in 2005 and 19 % in 2010. The same underrepresentation exists at cabinet level with usually only 4 % of Ministers being women (at best 12 % before the break-up of the LP/MSM/PMSD alliance in 2011 with 3 women Ministers out of a cabinet of 25).  In Rwanda women represent around 56 % of all Mps. In Sweden it is at 45 % while in Seychelles it stands at 44 %. Female representation is at 39 % in Mozambique and 36 % in Spain while Tunisia’s first democratic elections returned 26 % of women Mps.


 

Table 3: Gender Unfairness in Mauritius: Number of candidates and Mps

Year

Women Candidates

Total Candidates

% of women

Women      Mps

Number of  Mps

% of Women Mps

1967

1

120

0.83

0

70

0.00

1976

5

180

2.78

3

70

4.29

1982

4

180

2.22

2

66

3.03

1983

6

120

5.00

4

70

5.71

1987

6

120

5.00

5

70

7.14

1991

3

120

2.50

2

66

3.03

1995

14

180

7.78

6

66

9.09

2000

12

120

10.00

4

70

5.71

2005

16

120

13.33

12

70

17.14

2010

21

120

17.50

13

69

18.84

           

1967-2010

88

1380

6.38

51

687

7.42

 

 

The political system has been traditionally dominated by men , and women have considerable difficulties in being selected as Parliamentary candidates.  However when women candidates are fielded by political parties, they have absolutely no problem to be elected.   16 women were chosen as candidates in 2005 and 12 of them were returned as Mps.  Equally in 2010, 21 women were selected as candidates and 13 were elected.  There is no evidence of reluctance by the electorate to vote for female candidates on a gender balanced slate in three member constituencies.  The obvious reason accounting for the 19 % presence of women in the current Parliament is the fact that more women were selected by their parties in 2010 than in previous elections.   As simple as that!

 

Women have been elected to office across the country and at times in top position as evidenced by the 2010 results in Grand River North West/Port Louis West, in Port-Louis Nord-Montagne Longue and in Piton- Rivière du Rempart (it has happened in other constituencies in previous elections).  They are elected in the east, west, south and  north of the country, in Port Louis and in both lower and upper Plaines Wilhems. There are currently two constituencies (Port-Louis North-Montagne Longue and Vacoas-Floreal) with three women Mps each and one (Savanne/Black River) with two female representatives. The argument of reluctance and resistance from the voting public is both a nonsense and an insult to the intelligence and maturity of our population. Equally the argument of culture, history, sociology and societal values does not stand the test of empirical evidence, especially today.

 

Can we consider a political system where more than 50 % of the population is excluded from the political decision making process as a fair and an equitable one? Would the situation have been different if parties were not male dominated and women were more assertive politically ?  Especially when the same system guarantees seats to other cohorts in the population that account for far less than 50 % (in some cases as low as 5 %  based on ethnic representation) and when candidates are chosen on the basis of belonging to a smaller pool (various communal and sub communal groups).  As women, like men, share all the characteristics of the nation, political parties should have absolutely no problem to factor in appropriate balancing of elements other than gender when choosing women candidates. Even in countries with a history and culture of discrimination against women (like South Africa), significant progress has been made in gender fairness through the use of many techniques, including quotas over a relatively short period.  Reserved seats (only for women) as they exist in Rwanda, Uganda, Tanzania and Burundi are certainly not the way to proceed.  It is also not recommended to embed quotas in our constitution or to entrench it in an electoral law as these might arguably be challenged in court. The trend worldwide is to encourage parties to change their constitution and internal rules to adopt voluntary quotas for greater gender participation at elections. The Party quota could be beefed up by strongly worded guidelines from the Electoral Commission and a policy of ‘name and shame’ and moral suasion, especially by the voters. And it is working in more than 30 countries, including South Africa where women represent around 44 % of Mps and 30 % of Ministers. In its very short history of Parliamentary democracy, women have been elected as both Speaker and Deputy Speaker of the House. These are due to the policies and affirmative action mechanism embraced by the ANC since the first post–Apartheid elections in 1994. The quota is not in the Constitution, in spite of the fact that women have suffered the indignity of several inequalities during the cruel apartheid era.

 

Plural societies need policies and institutions that foster accommodation and encourage cooperation among different groups. They also require an electoral system that is inclusive from both a socio-demographic and a gender perspective. The key challenge is whether this should form part of a social compact, a consociational convention or it should be embedded in the constitution or another legislation. Mauritius has a very long history of tolerance, peaceful coexistence and mutual understanding. We are often cited as an example where different ethnic, religious, linguistic and cultural groups live well together. Why do we  need legislative and constitutional constraints to compel us to do the things that a nation should do towards all its citizens ? Why do we require a legal force to accept the simple and natural fact that women have a fundamental and a democratic right to fairness in political and parliamentary representation ? Why should we have legislation to compel us to ensure  adequate political representation to all the colours of our rainbow nation?  In all walks of life in contemporary Mauritius, we give and share and accept it as a vital condition for successful nation building in a multi-faith society. Why can’t political parties assume their responsibilities towards the nation and behave as mature parties do in well established democracies?

 

Table 4 depicts how quotas are implemented in countries where women account for more than 30 % of Mps. Countries with reserved seats are excluded from the analysis. It is interesting to note that women representation is generally higher at 58 % in the voluntary rather than in the legislated systems. It is important to highlight which countries use one method as opposed to the other one and which ones Mauritius should follow.  Most mature democracies like Mauritius use voluntary quotas by parties that often include them in their constitution and internal rules.

 

While the recommendation of Carcassonne on women is commendable (the first two candidates on the Party List to be of different gender), similar, if not better, proposals have been made before. There is a very simple solution that should be adopted. We should provide for the list of candidates in each constituency to comprise no more than 2 persons of the same sex. And on the Party list, we should ensure that neither gender represents less than 33 % of candidates and there is at least one person of a different gender out of every 3 sequential candidates. This has the merit of assuring both sexes in terms of both numbers and rank on the Party  list.

 

Unfortunately it has not been implemented. It should be pointed out that this recommendation is not related at all to electoral reform.   Even in the current FPTP system, it can easily be introduced if there is political will by parties.  I must also emphasise that in the medium to long term, we should aim for equality between men and women in Parliamentary representation.  Women have all the characteristics to be chosen as candidates.  No more but certainly no less than men. We have hesitated for too long on this issue and we have not even fulfilled  the commitment of all member states ( including Mauritius) taken at the SADC summit of 1997 in Malawi to achieve at least 30 % women’s representation in Parliament by 2005.  It is high time that we implement this decision, regardless of reforms to the voting formula.

Table 4: Type of quota in countries with high percentage of women Mps

Country

 

% of Women

in Parliament

 

Type of quota

 

Sweden

45

Voluntary

South Africa

44.5

Voluntary

Seychelles

43.8

Voluntary

Iceland

42.9

Voluntary

Finland

42.5

Voluntary

Norway

39.6

Voluntary

Belgium

39.3

Legislated

Netherlands

39.3

Voluntary

Mozambique

39.2

Voluntary

Denmark

39.1

Voluntary

Angola

38.6

Legislated

Costa Rica

38.6

Legislated

Argentina

37.4

Legislated

Spain

36

Legislated

Nepal

33.2

Legislated

Germany

32.8

Voluntary

Ecuador

32.3

Legislated

New Zealand

32.2

Voluntary

Macedonia

30.9

Legislated

  1. Subsuming the Best Loser System : promote inclusion, foster nationhood

 

One of the important issues facing democratic plural societies is the inclusion of diverse racial, ethnic, religious, linguistic and cultural groups in the political process, especially in a FPTP system where often it is ‘winner takes all’.  The problem is further compounded when some groups are small and dispersed geographically.  Countries have tailored different mechanisms to address the issue of inclusiveness and broad based socio-demographic representation.  Often there are difficult trade-offs involved in these choices. With some electoral systems, even free and fair elections cannot always assure the representation of all groups and interests.  The legitimacy and stability of a democratic system is challenged when it fails to include all groups in the decision making process.  It is essential, especially in plural societies, to provide representation and to avoid marginalization and exclusion.  The presence of all segments of society in Parliament strengthens representational links, encourages more positive attitude towards institutions and fosters legitimacy and acceptability.

 

The two broad electoral systems that exist- PR and FPTP- address the problem of diversity and inclusion very differently.  The focus of FPTP is effective, strong and responsive government and not the representation of all shades of opinion in society.  In such system there is no universal answer to the question of how groups that are relatively small in size and geographically spread should be represented in Parliament. Some countries ensure such political representation by providing special electoral law requirements for them.  These are ‘institutional workarounds’ to give comfort to all segments of society.  Mauritius has used various mechanisms to ensure that all sections of the population are adequately represented in Parliament while keeping the FPTP for its intrinsic qualities.

The Mauritius electoral system is very unique in four of its features.  These attributes are particularly designed to enable political inclusiveness in FPTP.  Had Mauritius chosen a PR system in 1967, there would have been no need for the BLS as the primary purpose of PR is to provide representation to all groups in society. As PR was not accepted as a voting formula and FPTP was instead retained, there was a necessity to include some special features in the system to guarantee the political representation of all the main segments of the population.

 

There are four such features in our voting formula. They are as follows:

 

(i)                 specially drawn electoral districts: this is a case of  ‘benign gerrymandering’ as the boundaries of some constituencies are drawn in such a way as to ensure to each main component of the population the maximum chance of electing representatives to Parliament in the areas where it is numerically strong.  The 40 single member constituencies were brought down to 20 three-member constituencies in such a way that the dilution of some specific groups was minimized.  Many ‘specially drawn’ constituencies were thus created to ensure broad representation.  The risks of affecting the electoral system were considerably diminished by an intelligent pairing of constituencies. More importantly it was done to allay the apprehensions of some segments of the population who were concerned about political exclusion. This was an example of electoral engineering to fit the context and the circumstances of the day. The boundaries were drawn by Banwell, following the principles initiated by Trustram-Eve in 1957 in the setting up of forty one-member constituencies. We still have this legacy of boundaries drawn to reflect some socio-demographic realities!

 

(ii)               malapportionment of constituencies: it refers to inequality in the population size of electoral districts. Here the ethnic balance of the constituencies is not unduly disturbed by either demographic changes or population movements across the country.  The consequence of maintaining such a balance is a significant disparity in the size of the 20 constituencies.  To encourage diversity and plurality, especially for groups which are dispersed across the country, our FPTP system accepts huge variation in the population/voter density of the 20 districts.

 

Such disparity in the size of constituencies explains the coexistence of two very small constituencies             with 23 000 and 25 000 electors, on the one hand (Rodrigues is another exception with still lower number of voters, albeit returning two elected representatives), and two extremely large ones with 58 000 and 60 000 electors, on the other. And each of these constituencies nevertheless returns 3 Mps. There are 20 000 electors per MP in Savanne/Riviere Noire and only around 7700 voters per elected representative in Port Louis Maritime/Port Louis East.  This is the price in a FPTP electoral system to achieve broad-based representation. Other communities also benefit from gerrymandered electoral districts. In a PR system, there is no requirement for such malapportionment as the formula works very differently.

 

(iii)             three member constituencies with a mandatory three votes per elector: the FPTP system with three members in each constituency and three compulsory votes allow a very delicate balance to be struck so that everybody is perceived to have a fair chance to gain political representation. It would have been very difficult to present a balanced slate of candidates with single member ridings. This is well articulated by Sachs in the 2001 report:

 

‘the bloc-of-three vote appears to have achieved substantial success in encouraging parties to straddle community divides when nominating candidates.  This is not simply because of any subjective desire to appear politically correct but because of the advantages which the bloc-of-three system gives to broadly based tickets.  The result is that no community is left out’.     

 

(iv)             And finally, the Best Loser System with the appointment of a maximum of eight best loser Mps. The objective is to mitigate the relative underrepresentation of some communities in Parliament after the elections of 62 Mps by FPTP mode.

While there is a historical and an emotional   dimension to the BLS, a compelling assessment of the 10 elections since 1967 unambiguously shows that the first          three features (specially drawn constituencies, unequal voters per constituency and three-member    electoral districts) have had a much greater impact on the parliamentary representation of some components of the population than the BLS.   For instance, the impact of ‘specially drawn constituencies’ and  unequal number of voters  has, in 2010, has  ensured the election of 6 Muslim candidates in Port Louis South/Port Louis Central            and Port Louis Maritime/Port Louis East. The votes polled by Soodhun alone in La        Caverne/Phoenix in the 2010 elections is more than the sum of the votes obtained by           the three elected Mps in Port-Louis Maritme/Port-Louis East. Equally Ganoo in Savanne/Black River attracted almost the same number of votes as the three elected     members in Port-Louis South/Port-Louis Central. This is the impact of ‘specially drawn constituencies, ‘malapportionment’ and three- member constituencies.  Equally the balanced ticket offered by three member constituencies can deliver additional Mps to our Muslim compatriots (and to other communities as well) across the country.  For the ten elections between 1967 and 2010,  the combined effects of ‘specially drawn constituencies, unequal distribution of voters and three

member districts  account for 82  % of all seats won by the Muslim community while the BLS represents only 18 % of such seats (around 1.9 Mps per election, on average). The corresponding figures for the General Population are 78 % (non BLS Mps) and 22 % (BLS Mps) respectively. These are illustrated at Table 5. It should be pointed out that there has never been a Sino-Mauritian best loser.

 

Table 5: Ratio of Non Best Loser  MPs to Best Loser Mps

Muslims MPs

General Population MPs

Year

Non BLS Mps

BLS Mps

Total Mps

Non BLS Mps

BLS Mps

Total Mps

1967

10

1

11

16

6

22

1976

7

3

10

15

5

20

1982

9

0

9

14

4

18

1983

9

2

11

15

6

21

1987

6

4

10

15

4

19

1991

8

2

10

18

2

20

1995

9

1

10

17

3

20

2000

9

2

11

17

4

21

2005

8

2

10

13

6

19

2010

9

2

11

16

5

21

Total

84

19

103

156

45

201

Average

8.4

1.9

10.3

15.6

4.5

20.1

%

82

18

100

78

22

100

 

While the symbolical dimension of the BLS is intense, its impact is not substantive compared to the other three features that are used to secure adequate political representation. It is precisely because of its low incidence on representation that an alternative, credible and modern mechanism can easily be found to adequately replace the BLS instrument and give reassurance to the community or communities most supportive of the Best Loser system.

As it is a sensitive issue, one rider is in order at this stage.  Even though another  community  has gained more from the BLS and other groups also have benefitted from specially drawn constituencies, the analysis focuses more on our Muslim compatriots as their size and their geographical distribution, compared to the other two main groups, could make it difficult for them to secure adequate political representation in a FPTP voting system.  This issue was comprehensively addressed by Trustram-Eve in 1958, De Smith in 1964 and Banwell in 1966 and various options were considered to ensure their fair Parliamentary representation. Equally, the same consideration of adequate presence in Parliament prevailed when it was proposed to do away with the nominations of some Mps, who represented ‘special interests’, by the Governor during the colonial era. The same principles that will operate to subsume the BLS for one community will apply for other communities while acknowledging their respective size in terms of the number of candidates fielded in constituencies and  those appearing on the Party List.

 

Conversely to FPTP, an electoral system based on Proportional Representation (PR) aims at including all components of society in Parliament. PR, with low threshold and large constituencies, maximizes the chances of small groups in electoral contests.  However it is weak on stability as there is rarely a clear majority that emerges from the elections. Countries that have diverse populations like Belgium (Flemish and Wallons) and Northern Ireland (Catholics and Protestants) invariably use the PR system to ensure diversity and inclusion.  However most of them need to have a coalition government as there is no single party with a clear majority. Many countries use special techniques to blunt the high proportionality of PR so as to produce some degree of stability.  Here the ‘workarounds’ have a completely different purpose. While in FPTP the techniques purport to have fairness of representation, in PR it is exactly the opposite. The mechanisms are intended to manufacture a majority to enable stability and ensure the effectiveness of government.  The ‘workarounds’ used are as follows:

(i)                 Introduce a threshold to make it difficult for small parties to obtain seat, thus minimizing the fragmentation of the system (10 % at the national level in Turkey);

 

(ii)               Adopt multi member constituencies instead of a country–wide PR.  Again the objective is to raise the bar to gain seats to enter Parliament. It gives a fillip to large parties (like in  Spain and what is being proposed by Carcassonne for Mauritius);

 

(iii)             Give a percentage of seats as premium to the party that comes first in a PR election. For instance, in Italy, the first party is guaranteed around 54 % of the seats even if it polls only 44 % of votes and the other seats are then allocated proportionately to the remaining parties;

 

(iv)             Allocate a surplus of seats in absolute term to the party ranked first in a PR election. For example, in Greece, the winning party has a premium of 40 seats even with less than 45 % of the votes and the remaining 260 out of 300 total seats are allotted on a PR basis among all parties.

 

FPTP and PR are completely different voting systems when it comes to the two key values of stability and fairness.  One prioritises stability and has to find techniques to provide adequate representation to minorities while the other privileges fairness and has to introduce mechanisms to ensure stability.  This is why the BLS is not seen as compatible with PR as the very essence of PR is to have broad-based socio-demographic representation. Similarly it would be inconceivable to give a premium to the winner in a FPTP election when it already has a majority of seats!  It is a matter of record that those who proposed a PR voting formula in the run up to the 1967 elections did not also ask for the Best Loser System since the PR itself would ensure broad-based representation. Such guarantee of political representation was demanded when PR was rejected and the FPTP adopted.

It is possible to provide the assurance embedded in the BLS by using a modern technique to operate its implementation while simultaneously removing its unacceptable features. Such technique is widely used in other countries to ensure socio-demographic inclusiveness. And it has been proposed by both Sachs and Carcassonne . It is wrong to state that Carcassonne has recommended the abolition of the BLS.  He has kept its objective while changing the mechanism through which it is implemented. In an interview given to L’Express Dimanche of 18th September, Carcassonne categorically stated that:

 

‘Attention, il ne s’agit pas de supprimer le Best Loser sans trouver un substitut qui garantisse à chaque communauté qu’elle sera normalement représentée.  Des substituts, on peut en imaginer. Le droit constitutionnel et la science politique ont fait, depuis 50 ans, des progrès phénoménaux. Nous avons dans la boite à outils un nombre d’instruments incomparables avec ceux de nos prédécesseurs.’

 

Equally, Sachs argued that:

 

‘the best approach would be to devise alternative and credible means of providing reassurance to the community or communities most supportive of the Best Loser system … it should be possible to achieve the original objectives of the BLS in a manner that is consistent with contemporary political reality’ .

And he emphasized that:

 

‘there was widespread support for the idea of using the list system to provide for establishing mechanisms that will subsume the BLS and further its underlying affirmative objectives without perpetuating its anachronistic and divisive aspects’

 

For instance, Carcassonne has subtly subsumed the BLS into the new electoral system with three specific proposals: the PR mode of election, a closed, rank-based party list to return Mps and the possibility of allowing a 15 % margin in the number of voters per constituency.  And of course, there could be a fourth one with the boundaries of the newly defined constituencies which he has recommended to refer to an Electoral Commission.

 

I am proposing a different set of alternative measures to subsume the BLS.  It hinges on a mix of retaining some of the existing features of our FPTP and of introducing some new elements that will give reassurance.  The current features that should be kept are as follows:

 

(i)                 maintain the three member constituencies that clearly offer a balanced slate of candidates;

 

(ii)               retain the 20 specially drawn constituencies ;

 

(iii)             keep the size and composition of constituencies as they are, until reform has settled in; this would avoid an untimely dispute on electoral boundaries. It should not be too difficult to accommodate the periodic review that is required by the Constitution;

 

In addition to the above, there is need for some additional arrangements to subsume the BLS into the new voting formula.  I submit that the most satisfactory solution for Mauritius would be a mixed electoral system.  We should keep the main features of the current FPTP system that have worked well and introduce an additional list tier of 20 Mps elected with a new formula that will considerably lower the disproportionality between seats and votes while preserving the stability of the system.  In the proposed formula, there will be two tiers of seat distribution.

 

It is a mixed system that combines the features of FPTP and a party list tier of 20 seats.  The recommendation is to subsume the BLS as follows:

 

(i)                 to introduce an additional list tier of 20 seats that will be allotted on a national basis;

 

(ii)               to have a closed, rank-based Party list to choose these 20 list tier Mps. This is similar to what is proposed by Carcassonne and Sachs; However instead of 100 % of Mps being chosen on that basis, I recommend that FPTP Mps continue to be elected in open competition as is the case today and only the 20 additional Mps be returned by the closed Party List. While the list would be an inclusive one, it is one way of providing some additional reassurance to those who feel concerned by the replacement of the BLS. The order of candidates on the closed Party list cannot be changed by voters. The system is unlikely to work in an open list as voters can alter  the rank of candidates;

 

(iii)             to provide for double candidacies.  Essentially some candidates will be fielded in FPTP constituencies and will also be included on the party list. In addition to other benefits (such as political leaders assured of presence in Parliament), it will also give a second chance if a candidate is not elected while his/her running mates are returned.  It has happened in very few cases. Candidates elected on FPTP voting mode will have their name struck from the party list as is the case in many countries with mixed system.  And other candidates will be chosen as list Mps based on their rank order.  It is a self-regulating mechanism. And  it works well in many countries with mixed system.

 

The above provisions constitute a sound trade-off between removing the unacceptable features of the Best Loser System while providing the reassurance of political representation to all segments of our population. It is consistent with the recommendation of both Sachs and Carcassonne. In this way, it will be possible to achieve the original objectives of BLS, in a new form adapted in a principled way to present-day conditions and dilemmas.  It is far from complicated. Electors will continue to cast three votes in their constituencies. Political parties will present a rank-ordered Party list of 20 candidates (some will be constituency candidates) and the 20 list tier Mps will be chosen by a specific formula. It will be more transparent than  the computation of the current BLS formula.

 

Box 2 summarises how the BLS will be subsumed into the mixed electoral system.  The maintenance of the current three-member constituency combined with the existing specially drawn constituencies and the unequal distribution of voters among constituencies will result in the election of around 8.4 Muslim Mps on average.  Three new features of the proposed voting formula  ( 20 Party list Mps, a closed, ranked-based Party list to return these Mps and the acceptance of double candidacies) will return more than 2 additional Muslim Mps. The actual number will depend on the outcome of constituency contests, on how the party list is constituted and how candidates are ranked.  The same principles will apply for the General Population, except that their number will be higher to reflect their larger population size. This is facilitated by the simple fact that there will be Muslim and General Population candidates ( and Hindus and Chinese also) on all Party Lists and the election of one member  from a community is not mutually exclusive with another one of the same community as they appear on different and distinct lists.

 

While the Sino-Mauritian community has never relied on the BLS to secure adequate representation, the case of Hindus is slightly different. Under most, if not all, circumstances they do not depend on the BLS to have adequate Parliamentary representation. This is obtained through the 62 FPTP Mps.  However there was one Hindu who was returned as Best Loser in the 1967 elections; and exceptionally two Hindus were chosen as Best Losers in 2000 and they owed their election to the Constitutional amendment introduced in 1992 to ensure that a party restores the balance in respect of the second set of four seats, irrespective of community, should candidates from the appropriate underrepresented community be unavailable. They were chosen as there was no candidate from either the Muslim or the General Population or the Chinese community  from the appropriate party to appoint with a view to restoring the balance. In the new formula, Hindus will continue to be fielded in the constituencies as is the case currently, in addition to obtaining their adequate share from the Party list that will reflect the diversity and rainbowness of Mauritius.

 

With 20 seats, it should be possible to reflect the socio-demographic diversity of our country. There is only one difference. Instead of these seats being constitutionally entrenched, it will now form part of a social compact, a consociational arrangement where the responsibility rests with political parties ,especially their leaders. And it has to work because of the self-interest of political parties. All the major parties have a very long tradition and practice of presenting a balanced slate to appeal to all segments of voters. And as shown in Table 4, the overwhelming majority of Muslim and General Population Mps are not returned by the BLS but by some social compact that compels political parties to field a balanced slate of candidates.

 

This recommendation represents a win win situation.  On the one hand, it is a great relief for those who feel that the BLS is the unacceptable face of our electoral system, that it undermines nation building, is discriminatory, archaic and anachronistic, very complex, arbitrary, erratic, irrational and paradoxical.  The reference to ethnic belonging will be removed from our Constitution, there will be no need to declare the community of a candidate contesting an election and we avoid the cumbersome post-election mathematical BLS computation based on ethnicity to return some 8 MPs.  And last but not least we will cease being open, as a nation, to stern criticism by the Supreme Court, the  Human Rights Commission and even the Law Lords on the current electoral arrangement based on ethnic classification.

 

A.  What exists today

 82 % of Muslims Mps ( an average of 8.4 Mps)  are elected  through a combination of:

i)                    three member constituency with balanced slate

 

ii)                  specially drawn constituencies

 

iii)                unequal distribution of voters across constituencies

 

18 % of Muslim Mps ( an average of 1.9 Mps) gain Parliamentary representation through  the Best Loser system

B.  What will be kept

 i)        retain three member constituency and the current 62 FPTP Mps

ii)       maintain specially drawn constituencies

iii)                allow unequal distribution of voters across constituencies

 

FPTP Impact: an average of 8.4 Muslim Mps will continue to be elected through a combination of these three features.

C.  What will change

The BLS is subsumed into the new electoral dispensation. It will not exist in its current form, but its principles are kept and absorbed as follows:

i)                    have a tier of 20 Mps in addition to 62 FPTP Mps;

 

ii)                  return the 20 Mps through a closed,rank-ordered Party List that cannot be changed by voters;

 

iii)                allow double candidacies so that some candidates will contest elections at constituency level and will also be on the Party List;

 

Party List Impact: more than 2 Muslim Mps and more than 4 General Population Mps will be returned from the Party List of the main parties, depending on what happens at constituency levels. It is a self regulating mechanism due to the existence of double candidacies. There will be Muslim and General Population ( and other communities) candidates on the Party lists of all major parties and their elections are not mutually exclusive.

Box 2: How the BLS will be subsumed into the new mixed electoral    system

 

 

 

18

 

 

 

 

 

 

 

 

C.C

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

On the other hand, it is a tremendous comfort and a reassurance for our compatriots who are concerned psychologically about adequate political representation. It keeps the objective of the BLS while it replaces its operational mechanism.  In societies that are divided on the basis of race, religion, ethnicity, languages, it is very crucial to design electoral systems that are all inclusive.  It induces shared goals and common values, encourages nation building and promotes social cohesiveness.  It creates space for groups to find some form of expression On On the other hand, it is a tremendous comfort and a reassurance for our compatriots who are  psychologically concerned about adequate political representation. It keeps the objective of the BLS while it replaces its operational mechanism.  In societies that are divided on the basis of race, religion, ethnicity, languages, it is very crucial to design electoral systems that are all inclusive.  It induces shared goals and common values, encourages nation building and promotes social cohesiveness.  It creates space for groups to find some form of expression through the political system.  However it should be done in a manner that does not create unnecessary stigmas. Constitutional engineers and electoral specialists have provided a variety of institutional arrangement to precisely achieve that objective. We know that political parties pay careful attention in choosing a balanced slate of candidates so as to appeal to voters from all communities. They will adopt the same principles for the composition and rank of the party list. As cogently argued by de Smith, Sachs and Carcassonne parties will be guided by self-interest and those that do not play by the expected rules will pay a heavy price at the poll.

 


 

8.                  Party list :better, safer, and more transparent than the BLS

 

A closed, ranked-based Party list is significantly superior than the current BLS in more ways than one. In addition to the usual criticisms that BLS institutionalizes communalism, uses archaic figures from 1972 to allot seats and is unfair against some segments of the Mauritian society, there are many practical problems with its operating mechanism. All the various  loopholes, discrepancies, paradoxes and inconsistencies of the BLS can be comprehensively addressed by adopting the safer,fairer and more transparent and predictable closed Party List.

First, the BLS, in its current form, is anachronistic. Subsequent to a Constitutional amendment in 1982, there is no need for Mauritians to reveal their ethnic identities for the purpose of population censuses. Since 1972, there is no official publication of the ethnic make up of the population. And yet Best Loser seats continue to be allocated according to the 1972 communal breakdown.  For instance, the Best Loser seats in 2010 were distributed using the population size and the communal breakdown of the 1972 census, almost 38 years ago.  The population has since increased by almost 50 per cent and it is highly likely that the ethnic composition is not the same as in 1972. Yet the system continues to use outdated statistics to allot Best Loser seats. The reliability of these figures and the fairness of the method can easily be contested.  It could also lead to anomalies. For instance, if the 1962 census (which was used to allocate the Best Loser seats in the 1967 elections) were used instead of the one of 1972, the allotment of Best Loser seats would be different in terms of communities and individuals for most elections. Second if we use religion ( based on subsidies granted by the State )  as a proxy for community in subsequent censuses and make some simple assumptions, this would give a third set of Best Losers. Finally, if there were more eligible communities than four, the allocation of Best Losers would again be different.

Second, Best Loser seats are allocated by the use of a very complicated formula using a two tiered system. Very few people understand how it operates as it has to take many factors into consideration. First the appropriate community, then the appropriate party with the appropriate community, then the sequence of appointments and finally it has to maintain a given political balance. The process can be quite arbitrary in its apportionment.  In 1995 a party that polled 20 per cent of the votes did not secure a single best loser seats. Four Best Loser seats were instead allotted to parties with a very small percentage of votes while the MSM/RMM had members of the appropriate community among its defeated candidates. In 2000 and 2005 also, two best loser seats were distributed to a very small regional party at the expense of a large national party that had many candidates of underrepresented communities in its ranks.

Third, the system is very erratic in its distribution. In the 1983 General Elections, Nawoor with only 16 % ( yes 16 %) of the vote was chosen as a Best Loser for the Muslim community while Kasenally with 47 % and belonging to the same community was not returned. Similarly in 1987, Finette was appointed with 47 % of the votes while Bérenger did not go to Parliament with 49 % of the votes even if both are from the General Population. The case of Bérenger in 1987 shows the paradox and the cruelty of the BLS. When one seat was awarded to a member of the General Population, Bérenger was in the ‘appropriate community’ but not in the ‘appropriate party’ and the seat went to Finette of the PMSD. When another seat was distributed to the MMM, then Bérenger was in the ‘appropriate party’ but not in the ‘appropriate community’ and Peerun of the MMM was chosen. This is because of the specific complexities, randomness and the two tiered basis of the allocation method. Had there been a Party list, Bérenger would have been returned to Parliament without any complication as the MMM would have been eligible for 14 Party List seats in 1987.

 

Fourth, there is no by election to replace a vacancy for a Best Loser seat.  In case of a vacancy, the seat is allotted to the most successful unreturned candidate available, belonging to the same community and party as the outgoing member. This could lead to a situation where the candidate may have given up politics or have left the country. Worse, he could have changed sides in the meantime.

Fifth, the system is also very unpredictable in its outcome. In 1982, 1991 and 1995 only four Best Loser Seats were distributed as the winning party did not have any unreturned eligible candidate. During those years, there was a Parliament of only 66 members instead of 70 MPs as provided for in the Constitution. And in 2010, only 7 seats were allotted for reasons that are not so easy to understand. As a matter of fact, the Electoral Supervisory Commission did not know whether to appoint any Best Loser in 1982 as it could upset the balance of political forces.  The matter was referred to the Supreme Court that ruled that only the first four Best Losers be distributed.

Sixth, the process can be mind-boggling in some cases. In 1983 there was an alliance among three parties (MSM, LP and PMSD). However to avoid three symbols on the ballot paper, it was agreed that in most constituencies all candidates of that alliance would stand under the joint symbol of the sun and the key (MSM/LP) while in some districts the symbol would be a cockerel only (PMSD). Notwithstanding the fact that the three formations were in a total alliance yet they were registered as two parties for the purpose of the elections (as MSM/LP and as PMSD). For tactical reasons, Malherbe, who was a very long serving member of the LP, had to stand on the PMSD platform in Grande Rivière North West /Port Louis West. He polled 44.91 % of the votes and yet was not returned as a Best loser while Candahoo, from the same alliance and belonging to the same community, got in with 43.78 % of the votes, as he contested under the MSM/LP banner!  Had Malherbe stood under the symbol of his party, he would have been chosen as a Best Loser.  In the same elections, Clifford, a PMSD candidate for the MSM/LP/PMSD alliance in Rose-Hill/Stanley, polled 46.27 % of the votes; yet he was not awarded a Best Loser seat while Michel, from the same community and the same alliance, got one such seat with 44.32 % of the votes as he fought the elections as a MSM/LP candidate.  This is how unpredictable and random the system can be. 

Seventh, the number of electors in a constituency affects the allocation of Best Loser seats. As a result, a candidate can secure a Best Loser seat with far fewer votes than another one belonging to the same community and the same party. The most known example is the case of Rima and Koenig in 1967.  Rima was allotted a Best Loser seat with 6,596 votes while Koenig, the leader of the PMSD, did not have one with 6,851 votes. Both were from the PMSD and both belonged to the same community. It was the smaller size of the constituency that favoured Rima as his percentage of votes at 45.72 % was slightly higher than that of Koenig at 44.37 %. The case of Balancy and Chaperon in the same elections confirms how constituency size can influence the allocation of Best Loser seats.  Both were from the Independence party and both were members of the General Population. Balancy became a Best Loser MP with 4,884 votes while Chaperon was not with 5,998 votes. The constituency of Chaperon was 34 % bigger than that of Balancy. In 1995 and 2000, it led to a double anomaly. In 1995, Beeharry, a Muslim, was awarded a Best Loser seat even if his party had a very insignificant percentage of the national vote compared to the 20 % of the MSM/RMM. Worse he won only 4,405 votes while Soodhun, also a Muslim, polled 7,416 votes without a Best Loser seat. In 2000, Leopold with only 7,732 votes (his party took  1 % of national vote) won a Best Loser seat while Petit, of the same community, with 14,626 (and his party had 37 % of votes overall ) did not.

Eighth, it is possible to abuse the system. After an examination of the provisions of the Constitution, Judge Seetulsing reached the conclusion that it is perfectly constitutional for Seegobin, Lallah and Renghen (all Hindus) to be classified as General Population, which is a ‘residual’ community in the Constitution. Consequently, he questioned the enforceability of the BLS and stated that this problem should be addressed within the context of electoral and constitutional reform. In essence, there is nothing that prevents a candidate of a particular community choosing, for the purpose of elections, to become a member of another community so as to enhance his chances of obtaining a best loser seat. There are candidates that select their community for election purposes by drawing lots as they castigate the communalistic aspects of the Best Loser system. Moreover candidates of the Chinese community, who are also Christians, could improve their chances of being a Best Loser MP if they chose to be classified as General Population. Such abuse could pervert the basis for which these seats are allotted. And it has happened. This is because the BLS is a curious mix of ethnicity and religion

Ninth, it leads to rivalry among candidates who belong to the same community as it is a ‘zero sum game’. If one member of a particular community is elected, it considerably diminishes the chances of another member of the same community from being appointed as they are mutually exclusive. As Party lists are distinct and separate, this problem is largely mitigated. There will be candidates from all communities on all Party list.

All these above-mentioned problems and their consequences will be addressed with the certainty and transparency of a Party List.  The system is safe, reliable and predictable. As the Party list will reflect the diversity of the country, all political leaders will find their place in Parliament without playing a game of high risks and high uncertainty with the BLS. Razack Mohamed, a leading member of the Muslim community, would have entered Parliament with a Party List in 1976. The same applies for Koenig in 1967, Duval in 1976, Ramgoolam in 1982, Berenger in 1987, A..Jugnauth in 1995 and P. Jugnauth in 2005. Yet none of these political leaders won a seat in Parliament with the Best Loser system.

9.                  How should PR seats be allocated? Pandemonium in

         heaven

 

While many people grasp the difference between FPTP and PR systems, they do not always make the fundamental distinction between a parallel PR ( as proposed in Model A of Sachs) and a compensatory PR ( as recommended in Model C of Sachs), when used in tandem with FPTP. This is currently a major source of confusion in the mind of many well-intentioned persons. Parallel PR allocates the PR seats to parties in relative proportion to their votes. It does not take into account the ( possibly disproportionate number of ) constituency seats they have won. It could thus be unfair and unrepresentative as it may lead to huge disparities between overall votes and total seats. By contrast, constituency seats won are taken into account in Compensatory PR. The method aims to align the total seats ( constituency plus PR) of each party to its votes. A party which has obtained more ( less) constituency seats than deemed by its share of votes will receive less ( more) or no PR seats. Because of its proportional features, the compensatory formula is adopted by some democracies that have a mixed model. This is the case in New Zealand, Germany, Scotland and Wales. It has also been adopted by Lesotho which moved from a FPTP to a mixed FPTP/PR system. However it makes government formation difficult as it could be too proportional. On the other hand countries which have recently embraced democracy have adopted a parallel system. Russia (until recently), Ukraine, Georgia and Armenia are such examples. Japan also has a mixed parallel PR/FPTP voting formula. While both Russia and Germany have mixed electoral system with 50 % of seats returned with FPTP and 50 % with PR, Russia uses a parallel PR while Germany applies a compensatory one. The consequences are very different, even if both are mixed systems. While both provide for representation, the degree of fairness is very different. In the Russian case, a party that receives 40 % of  the  votes  may  end up with only 20 % of the seats even if it has lost all constituency contests. In the German system it could receive, in similar circumstances, between 35 % and 38 % of the seats as the compensatory PR allotment corrects for the unfairness of the FPTP results.

Table 6  illustrates the major difference between the parallel and the compensatory method of allocating PR seats in a mixed FPTP/PR electoral model. The two elections of 1982 and 1987 in Mauritius are used as worked examples.

Table 6: Difference between Parallel and Compensatory method of    allocating PR seats

FPTP/Parallel PR

FPTP/Compensatory PR

1982

% votes

FPTP

PR

Total

% seats

FPTP

PR

Total

% seats

seats

seats

seats

seats

Seats

seats

MMM/PSM

64.16

60

14

74

90.2

60

0

60

73.2

PAN

25.78

0

6

6

7.3

0

20

20

24.4

Other

10.06

2

0

2

2.4

2

0

2

2.4

Total

100

62

20

82

100

62

20

82

100

1987

% votes

FPTP

PR

Total

% seats

FPTP

PR

Total

% seats

MSM/LP/PMSD

49.86

39

10

49

59.8

39

2

41

50

MMM & Allies

48.12

21

10

31

37.8

21

18

39

47.6

Other

2.02

2

0

2

2.4

2

0

2

2.4

Total

100

62

20

82

100

62

20

82

100

 

In 1982, the parallel mode, which disregards the fact that PAN did not receive a single seat for 25.78 % of the votes in FPTP elections, allocates 6 seats to the second party and 14 to the MMM/PSM alliance , based simply on their respective share of votes, after eliminating parties that do not meet the threshold to be eligible for these PR seats. The final outcome is 74 seats for the MMM/PSM and 6 seats for PAN. It is highly disproportional with PAN attracting only 7 % of seats for 25.78 % of votes. The compensatory method recognises the unfairness of the  FPTP elections and  compensates PAN for

its share of votes using the D’Hondt formula. PAN receives all 20 seats as it was severely penalised in the constituencies. The final result is 60 seats for MMM/PSM and 20 seats for PAN. The second party obtains 24.4 % of seats for 25.78 % of votes. It is almost a proportional outcome. The small difference between the two figures is the impact of thresholds. In 1982 the compensatory mode is fair and it also leaves a huge working majority for the winning team. The parallel formula is very insignificant in correcting the unfairness of the system.

 

The situation is completely different in closely held elections. In 1987, the parallel formula allocates 10 PR seats to each of the two main parties as their share of votes was very close. However as it does not take into consideration what happens in FPTP contest, the final tally

gives 49 seats to the MSM/LP/PMSD alliance and 31 seats to the MMM & allies. The compensation mode ,on the other hand, allots 18 of the 20 PR seats to the MMM & allies in recognition of the fact that it was penalised in FPTP elections with only 21 seats on 48.12 % of the votes. The final outcome is 41 Mps for the MSM/LP/PMSD alliance, 39 Mps for the MMM & allies and 2 Mps for Rodrigues. We are thus in a situation where a relatively large majority of 18 (maintained after the BLS) seats is transformed into a hung Parliament with no majority as the combined forces of the Opposition is equal to the numerical strength of the winning party. It is too proportional.

 

The parallel mode of apportioning PR seats leaves huge distortion between seats and votes. This is the main reason why it was rejected by Sachs and not considered by Carcassonne. However the compensatory mode of allocating PR seats converts the electoral system into a proportional one in many cases. And we are in fact in a Carcassone-like system of  very close proportionality between seats and votes. This is why the compensatory formula is also called mixed-member proportional by electoral specialists. As a matter of fact this is precisely what happens very often in New Zealand and in Germany. While on paper, it is a mixed system, in actual fact it becomes a very proportional voting formula.

 

During my consultations, I have also found that many people do not necessarily have a very good understanding of how PR seats are apportioned in a mixed FPTP/PR system using the D’Hondt formula. Consequently I am giving a worked example of the operational aspects of the formula using the 1987 election as a case study at Appendix 1.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10 .   The ‘unreturned votes elect’ formula (UVE) : a new kid in town

 

As the parallel mode is very disproportional and the compensatory formula highly  proportional, I am recommending a third possibility of allotment of PR seats on the basis of the votes of ‘unreturned candidates’ . There is in fact a fourth one which is the surplus vote. However it is complicated and less fair than an allocation by the votes of unreturned candidates. The term used in political and electoral science for such votes is ‘wasted votes’. However, I shall refer to it as the ‘unreturned votes elect’ formula ( UVE) as wasted vote may send the wrong signal as a basis to elect Mps !. These are the votes of candidates who have polled a fairly significant percentage of votes and who have not been elected (hence the appellation of ‘wasted’ in the literature as this is exactly what it is). These unreturned votes shall be used to elect the 20 list tier Mps. This is why it is called the UVE formula.

 

An electoral system which reduces the number of lost or wasted votes can be considered desirable on grounds of fairness.  For example, in the UK General Elections of 2005 which use a FPTP voting system , 52 % of votes were cast for unsuccessful candidates. This is perhaps the most fundamental criticism of FPTP, that a large majority of votes plays no part in determining the outcome. The proposal is to use the votes of unreturned candidates to have an additional tier distribution of 20 seats. It is not perfect. However it has the merit of keeping a very satisfactory balance between stability and fairness. Significantly UVE  produces a working majority while lowering the huge distortion between seats and votes of the unsuccessful party.

 

The idea behind compensating political parties on the basis of the votes of unreturned candidates is conceptually sound, especially in our system.  It aims at minimising or eliminating the number of wasted votes that often accounts for the huge disproportionality between votes and seats in FPTP voting formula.  Many candidates with a substantial percentage of votes do not get any representation.  For instance, 46 % of those who cast a ballot did not contribute to electing anybody with their votes in the general elections of 2010, as they went to unsuccessful candidates. In the 1967 elections, the Independence Party took all three seats in Vacoas/Floreal with 51 % of the votes while the PMSD got nothing with 48.6 % of the votes. The MMM captured all three seats in Rose Hill/Stanley with 53 % of the votes in the 1983 elections, while the MSM/LP/PMSD alliance did not receive anything with 46 % of the votes. The unfairness of the current formula is due to the fact a party often ‘takes all’ in one constituency (this has happened in 160 out of 200 constituencies in the last 10 elections!) and the party whose candidates come immediately after the elected representatives does not obtain anything, even if it has attracted a significant share of the votes.  The more such losers there are and if they hail from the same party, the higher will be the disproportionality between seats and votes. The UVE formula has the advantage of rewarding the overall winner also in constituencies where it has failed to win seats.  This, indeed, happens when elections are very close.  In this  system, the party that does not win the election will be better compensated by the formula; however it will not work to the extreme of making it purely proportional as is the case with Model C of Sachs. It counts only the votes of unreturned candidates for the simple reason that the votes of returned candidates have already been used to elect the three Mps. To some extent, it avoids using the same votes twice , which is the case in the other formulae.

 

 

 

 

11.  How much stability, how much fairness?: the great    dilemma

 

Of the various criteria listed by Carcassonne, the current FPTP electoral system functions very well on stability. It is effective, transparent, responsive and decisive. There is no better system than FPTP to produce a clear and distinct majority which is absolutely essential to govern the country.  PR is weak on this count and in some cases extremely so.  Gender fairness and socio-demographic inclusion have already been dealt with.  We are therefore left with the principal deficiency of FPTP which is its inability to translate votes into seats in an equitable manner.  It delivers disproportional outcomes in certain elections, magnifies majorities and could even lead to the complete elimination of opposition parties in Parliament, thus vitiating the very foundation of representative democracy.  This is the principal defect to be cured; however in so doing, one must be careful not to create problems with the criterion of stability.  Put simply, the challenge is how to strike the right balance between stable, decisive, effective, responsive and accountable government that are considered the strengths of plurality/majoritarian systems and fairness and representation which are the main attributes of proportional representation formulae.  Often, these two sets of criteria move in opposite directions and this has led many to believe that the perceived conflict between effectiveness and representation, between stable government and fairness is a very difficult one to resolve.  Many countries have attempted to reconcile this contradiction by embracing a mixed system that combines the best features of FPTP and PR electoral systems.

 

It is a fact that no electoral system could possibly fully meet all the criteria for a good voting system. It is my considered view  that a mixed electoral system that combines our current FPTP formula with an additional list tier allocation of 20 seats is best suited to address the issues of stability and fairness in our electoral landscape.  However, to avoid the high proportionality of the system in some specific cases and its lack of a working majority, I suggest that we adopt a new approach based on an apportionment formula that is different from the one used in Model C of  Sachs.  Box 3 explains how unfairness is measured in electoral systems. Most experts do not expect a fair voting formula to be completely proportional.

There are many ways of determining whether an electoral system is fair or not. Most measures of fairness concern the relationship between the share of votes captured by parties and their share of seats in Parliament. Although the concepts of proportionality appears simple, it is not that straightforward to find the best method to measure the fairness of election results. There are many indices of proportionality and each of them has the same starting point. However they differ in the manner the seat and vote deviations are aggregated

The simplest and easiest measure of the unfairness of an electoral system is the Degree of Deviation. It measures the gap between the party’s share of votes  ( V) and its proportion of seats (S).

However there are more precise formulae to compute disproportionality but they are  complex. The best statistical measure of fairness used by electoral experts is the least square index as follows:

where V represents the percentage of votes polled while S stands for the percentage of seats obtained by each party. Zero would indicate perfect proportionality while departure from zero indicates increased disproportionality

What degree of deviation constitutes a fair Parliament is a matter of judgement. Even in a PR system, nobody expects it to be a zero as it will make stable and effective government impossible. Some experts believe that in a PR system a deviation of between 4 to 8 % is acceptable.

It should be obvious that under the FPTP system this degree of deviation is significantly higher. Here, it is suggested that a disproportionality in excess of 15 % would be unfair. Some would put a 10 % deviation as a maximum to test the degree of fairness. It essentially means that a FPTP electoral system can be said to be substantially fair and representative if a party has a maximum advantage of  around 10 % of seats compared to its share of votes.

 

 

Box 3: Measuring unfairness in electoral systems

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Many were justifiably concerned that Model C of Sachs, while being very attractive in many ways, becomes too proportional in very close elections.  The maximisation of proportionality is not always a desirable end in itself.  It might produce a Parliament that is too fragmented, thus making many werit difficult to have strong and effective government.  The UVE formula is a significant

Many were justifiably concerned that Model C of Sachs, while being very attractive in many ways, becomes too proportional in close elections. The maximization of proportionality is not always a desirable end in itself . It might produce a Parliament that is too fragmented, thus making it difficult to have strong and effective government improvement. The UVE formula is a significant improvement on the current system and it is also better in terms of ‘overall balance’ to what has been proposed up to now. If we want to preserve stability which is a fundamental feature of an electoral system, we should avoid pure PR and choose broad proportionality instead.  There will still be some divergence between seats and votes but it will be small and tolerable compared to the current voting mode.  We should retain the best values of the FPTP electoral formula and design a mechanism to remedy its major drawback. Table 7 gives the results of the 10 elections held since 1967.

Table 7: Stability versus fairness at a glance

Winning Party

Second Party

Others

 

Majority

Year

% votes

seats

% seats

% votes

seats

% seats

% votes

seats

% seats

In Seats

1967

54.13

43

61.4

43.99

27

38.6

1.88

0

0

16

1976

38.64

34

48.6

37.9

28

40.0

23.5

8

11.4

Hung

1982

64.16

60

90.9

25.78

2

3.0

10.1

4

6.1

54

1983

52.22

46

65.7

46.4

22

31.4

1.38

2

2.9

22

1987

49.86

44

62.9

48.12

24

34.3

2.02

2

2.9

18

1991

56.3

57

86.4

40

7

10.6

3.7

2

3.0

48

1995

65.2

60

90.9

19.7

0

0.0

15.1

6

9.1

54

2000

51.5

58

82.9

36.7

8

11.4

11.8

4

5.7

46

2005

49.4

42

60.0

43.3

24

34.3

7.31

4

5.7

14

2010

50.7

45

65.2

42.9

20

29.0

6.4

4

5.8

21

 

It shows that FPTP has delivered on stability but has been weak on fair representation on several occasions. It is good at producing a majority to ensure effective government even when the elections are close in terms of votes  between the two  major  contenders.   In 1987,  a difference  of only 1.74 % in national votes (49.86 % against 48.12 %)  between the LP/MSM/PMSD and the MMM produced a huge disproportionality in seats of 28.6 %  (62.9 % against 34.3 %), thus giving an absolute majority of 18 seats to the governing party.  In 1991, a  divergence  of  16.3 %  in  votes (56.3 % against 40 %)  produced a difference of 75.8 % in seats for the MSM/MMM alliance over the LP/PMSD coalition (86.4 % against 10.6 %), thus yielding a majority of 48 seats for the winner.  The same degree of  unfairness occurred in 2000 when a difference of 14.8 % in votes manufactured a seat disproportionality of 71.1 %.  And  we have the two cases of 60-0 (1982 and 1995) with one party taking all 60 seats and the second party that attracted 25.78 % and 19.7 % of votes failing to win a single seat in the FPTP contest.

 

The objective of a Party list tier seats is to narrow the difference between seats and votes.  However in so doing it could make the system weak on stability as the majority becomes extremely thin.  This is the problem with  Model C of Sachs in some elections.  Table 8 depicts the results of four elections out of ten since 1967 using the Sachs compensatory method of allocating 20 PR seats.  The formula would have lowered the seat/vote disproportionality without affecting the stability of the system.  For ease of comparison, the change in majority is shown in the last column of Table 8.  The system works well when there is a huge unfairness against one party.  It happened in 1982 and 1995 (60-0) and in 1991 and 2000 (57 seats for winner and only three seats for the second party and 54 seats to 6 seats respectively).   In  1982  all  the 20 PR seats would have been  distributed to the  party  that  polled 25.78 %   of  the  votes  and  did  not  secure  a  single  seat.  After the allotment of the 20 PR seats, the  PAN alliance would have had 24.4 % of total seats.  In fact had there been additional PR seats (say 25 instead of 20); PAN would have collected another 4 seats.

Table 8 : When Model C of Sachs poses no threat to stability

62 FPTP and 20 PR Model C Sachs,10  % threshold

% votes

FPTP MPs

PR seats

Total seats

% of seats

Majority in seats

1982

MMM/PSM

64.16

60

0

60

73.2

PAN

25.78

0

20

20

24.4

PMSD

7.79

0

0

0

0.0

Other

2.27

2

0

2

2.4

Total

100

62

20

82

100.0

from 54 to 38

1991

MSM/MMM

56.3

57

0

57

69.5

LP/PMSD

40

3

20

23

28.0

Other

3.7

2

0

2

2.4

Total

100

62

20

82

100.00

from 48 to 32

1995

LP/MMM

65.2

60

2

62

75.6

MSM/RMM

19.7

0

18

18

22.0

Other

15.1

2

0

2

2.4

Total

100

62

20

82

100.00

from 54 to 42

2000

MSM/MMM

51.5

54

0

54

65.9

LP/PMXD

36.7

6

20

26

31.7

Other

11.8

2

0

2

2.4

Total

88.2

62

20

82

100.0

from 46  to 26

 

In 1995, the second party would have taken 18 of the 20 PR seats.  It did not capture all, as in 1982, because its tally of votes was only 19.7 %. As it is an almost PR system, parties are  compensated up to their share of vote.  In 1991 and 2000, the LP/PMSD alliance would have received all 20 PR seats as they were very disadvantaged by the FPTP outcome (only 3 seats for 40 % of votes and 6 seats for 36.7 % of votes respectively).  In the four cases of  1982, 1991, 1995 and 2000, the winning  party  would  still command  a  relatively large majority of seats while the disproportionality between seats and votes is lowered  for the second party.

However the situation with respect to stability becomes tenuous when a major party does  well in terms of votes, wins some seats but is heavily penalized in terms of constituency seats.  Under these circumstances, most, if not all of the seats, would accrue to that party as it has suffered badly from the unfairness of the FPTP.  This happened in 1987 when the MMM  received only 34.3 % of the seats with 48.12 % of the votes (a massive negative distortion of 13.8 % between seats and votes) while at the same time the MSM/LP/PMSD gained 62.9 % of the seats with only 49.86 % of the votes (a huge positive difference of 13.04 % between seats and votes).  In view of the huge disproportionality of the FPTP outcome, the PR seats would have been awarded mostly to the MMM (18 seats against 2 for the winning party).  And the consequence is the disappearance of the majority of 18 that the MSM/LP/PMSD had after the FPTP results.  And if the two Mps from Rodrigues are not counted as forming part of the winning alliance, there is a deadlock with 41 seats for the Government and a combined 41 Mps against it. These results are illustrated at Table 9.

 

1983 also poses a challenge for almost the same reason as in 1987.  The vote/seat difference is quite high for both parties (positive 13.5 % for the MSM/LP/PMSD and negative 15 % for the MMM).  The compensatory allocation of the 20 PR seats has the mechanical effect of lowering the majority from 22 to 2 seats as 19 out of the 20 PR seats are allotted to the MMM as it suffered heavily from the constituency outcome.  In fact the majority of 2 could easily have vanished as there were some hotly contested seats.  Three candidates of the MMM were defeated by very few votes.  Bérenger lost by 33 votes in Belle-Rose/Quatre Bornes, Jawaheer by 112 votes in La Caverne/Phoenix and Minerve by 237 votes in Beau Bassin /Petite Rivière.  The 2005 and 2010 elections would also have posed some problem for stability after the apportionment of the compensatory seats.  The seat majority comes down from 15 to 4 in 2005 and from 21 to 4 in 2010.

The case of 1967 is quite interesting and is different from the elections of 1983, 1987, 2005 and 2010.  The PMSD did relatively well in the FPTP contest by winning 38.66 % of the seats with 43.99 % of votes (a difference of only 5.33 % between seats and votes).  As a result the mechanics of Model C of Sachs distributed some seats to the winner also.  This has the effect of reducing the majority from 16 in the FPTP mode to 8 in the mixed FPTP/PR formula. However it still leaves a working majority for the winning party.  While the formula is extremely fair, it however moves away from being stable and effective.  This is the major drawback of too proportional an electoral formula.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 9: Model C of Sachs is too proportional at the expense of stability

 

62 FPTP and 20 PR Model C Sachs,10  % threshold

% votes

FPTP MPs

PR seats

Total seats

% of seats

Majority in seats

1967

IP

54.13

39

6

45

54.9

PMSD

43.99

23

14

37

45.1

Other

1.88

0

0

0

0.0

Total

100

62

20

82

100.0

From 16 to 8

1983

MSM/LP/PMSD

52.22

41

1

42

51.2

MMM

46.4

19

19

38

46.3

Other

1.38

2

0

2

2.4

Total

100

62

20

82

100.0

From 22 to 2

1987

MSM/LP/PMSD

49.86

39

2

41

50.0

Union

48.12

21

18

39

47.6

Other

2.02

2

0

2

2.4

Total

100

62

20

82

100.0

From 18 to 0

2005

LP/PMSD

49.39

38

5

43

52.4

MSM/MMM

43.3

22

15

37

45.1

Other

7.31

2

0

2

2.4

Total

100

62

20

82

100.0

From 14 to 4

2010

LP/MSM/PMSD

50.72

41

2

43

52.4

MMM

42.88

18

18

36

43.9

Other

6.4

3

0

3

3.7

Total

100

62

20

82

100.0

From 21 to 4

 

 

 

The three-cornered electoral contest of 1976 was quite unique. The FPTP mode produced a difference of 5 seats in favour of the MMM against the Independence Party (30 seats to 25 seats). However it was a very close election with the MMM on 38.64% of the votes and the Independence Party on 37.9% of votes  . As Party list seats are allocated on percentage of votes polled after taking into account what happens at the level of constituencies, the Independence Party would have obtained 9 seats and the MMM 4 seats. Both parties would therefore end up with 34 seats and the 5 seat margin of the MMM would disappear. The PMSD would have taken 7 Party List seats in these elections. Consequently the combined seats of IP and PMSD (after entering into a coalition after the elections) would have been significantly higher than the MMM (48 against 34). This is shown at Table 10.

 

Table 10: The special case of the 1976 Elections

62 FPTP and 20 PR Model C Sachs,10  % threshold

% votes

FPTP MPs

PR seats

Total seats

% of seats

Difference in seats

1976

MMM

38.64

30

4

34

41.5

IP

37.9

25

9

34

41.5

PMSD

16.2

7

7

14

17.1

Other

7.26

0

0

0

0.0

Total

100

62

20

82

100.0

From 5 to 0

 

For the sake of setting it aside, it should be noted that Model A of Sachs (which allocates the PR seats regardless of constituency results) does not cure the unfairness of the system. A combination of a relatively high number of FPTP seats in a mixed system coupled with a parallel formula to allot the low number of PR seats would not significantly reduce the seat/vote distortion produced by the excesses of the FPTP system.  Its impact is small as it would fail to correct the underrepresentation of unsuccessful parties.  In fact, the majoritarian effect of the formula remains largely dominant, although tempered by a very small degree of proportionality.  Worse, in many cases it actually widens the majority of the winning party in terms of seats.  This is illustrated in Table 11 below.

 

Table 11: The insignificance of Sachs’ Model A on fairness (20 PR seats)

 

Year

Majority (seats)

 

Majority (seats)

FPTP/BLS

 

FPTP/Sachs A

1967

16

18

1976

2

6

1982

54

66

1983

22

22

1987

18

16

1991

48

56

1995

54

68

2000

46

50

2005

14

16

2010

21

22

 

 

In the majority of cases, the winning party which has already secured an advantage from the FPTP formula reaps the benefits of the parallel system of the 20 PR seats.  For instance in 1982, the MMM/PSM won all 60 seats while four BLS were awarded to two opposition parties.  The winning coalition had a majority of 54 seats if we include the two Rodrigues Mps as not forming part of the Government. The MMM/PSM alliance would have taken 14 out of the 20 seats as it polled 64.16 % of the votes as the parallel mode of attributing seats does not take into consideration the outcome at constituency level.  The second party which was already very disadvantaged by not winning a single constituency seat with 25.78% of the vote is punished a second time as the lion’s share of PR seats still goes to the winner of FPTP elections.  This explains why the majority in fact widens (from 54 to 66 seats) after the allocation of PR seats, even if as a percentage there is a very slight improvement in the fortunes of the second party.  The same pattern is observed in 1991, 1995 and 2000.  These were the years when the second party was hugely underrepresented and yet the absolute difference in the number of seats increases after the distribution of the 20 PR seats.  It is hardly a solution to correct the deficiency of the FPTP formula except if we are looking at token representation.

12.   The way forward : keep FPTP baby, throw away bath water of its huge unfairness

 

Some countries have electoral systems characterized by conscious design after extensive and wide ranging debate on the attractiveness of particular electoral formulae. In other cases, the electoral formula has been the result of compromise by the various political actors based on its culture, heritage and tradition rather than on an objective evaluation of the relative merits of different rules. Our electoral system is clearly the outcome of a mix of conscious design after extensive discussion and the result of a compromise by the various political actors.  Banwell had produced his report after protracted debate; yet it was not accepted as such. Subsequent negotiations led to the adoption of the 20 three-member constituencies as recommended by Banwell, the rejection of the variable correctives, the revision of the fixed correctives from 5 to 8 Mps and the addition of new conditions to ensure that the apportionment of the BLS does not change the results of the elections.  The requirement to poll at least 10 % of national votes to be eligible for BLS was also dropped.

 

It is against this background of compromise by political actors that I am suggesting a uniquely designed electoral system based on our context, our history and our realities to resolve the conflict between stability and fairness.

 

The main features of the proposal are as follows:

 

(i)                 to retain the 20 three-member constituencies and one two-member constituency (Rodrigues) using FPTP mode.  Voters will continue to elect three candidates ( two in Rodrigues) of their choice as is currently the case.  This is the nominal tier;

(ii)               to keep the boundaries of the 20 existing constituencies unchanged for some time.

This may be necessary to reassure some segments of society, to allow the new electoral system to settle down and to prevent the reform from being hijacked by other considerations;

 

(iii)             to subsume the BLS into the new FPTP/additional list tier voting formula. The objective of diverse representation remains while the mechanism to ensure its attainment is replaced by the set of measures spelt out earlier in this report;

 

(iv)             to allocate an additional 20 seats on a party list tier. It would be a very difficult practical task to cure the defect of unfairness in the system without increasing the number of seats.  The alternative is to lower the number of FPTP constituencies so as to keep the overall number of MPs constant or to change the formula radically as proposed by Carcassonne.  But then we have to redefine constituency boundaries and this would become a source of bitter controversy and deep division;

 

(v)               to have a National Assembly of 82 seats, 62 elected through FPTP system and 20 by party list.  This would imply an additional 12 seats compared to what obtains currently.  It is probably the price to remedy the deficiency of the FPTP system without causing collateral damage in terms of redrawing constituencies and reducing their number. And it will also facilitate the subsumption of the BLS;

 

(vi)             to allot the 20 party list tier seats on a national basis;

 

(vii)           to return these 20 list tier Mps through a closed, rank-based party list. Parties should constitute their list both in terms of numbers and rank to broadly reflect the diversity and plurality of the country.  In that mixed system, 62 Mps or 76 % of all Mps will be elected directly by the voters that send them to Parliament through the FPTP system while 20 Mps or 24 % of all Mps will be chosen on closed, rank-based party list.  This is necessary both for fairness and inclusiveness. There are countries with diverse population that use this technique to ensure broad based political representation;

 

(viii)         to allow double candidacies where some candidates will both appear on the Party list and will compete in FPTP constituency elections. The actual number of double candidacies should be left to parties as part of their electoral strategy of aligning candidates to maximize their chances of success. Again this exists in many mixed systems. If a candidate is not elected in the constituency, he/she has the chance of being returned from the list.  The basis for this flexibility is so obvious in our electoral system;

 

(ix)             to have no more than two candidates of the same sex in each of the 20 constituencies.  We should equally provide for party lists to ensure that neither gender represents less than 33 % of candidates. And we should have a narrow ‘zipper’ so that there is  at  least one person of a different gender out of every three sequential candidates on party lists ;

 

(x)               to introduce a minimum threshold of at least 10 % of national vote for parties to be eligible for the 20 list tier seats; if it appears too unfair to some genuine, non- communal parties, we could provide that the 10 % of national vote shall be lowered to  5 %  if a party also wins three constituency seats ;

 

(xi)             to use the d’Hondt highest average  formula to allocate the 20 additional list tier seats among eligible parties;

 

(xii)           to have one set of vote (three votes per constituency on the mainland and two in Rodrigues). The three votes (two in Rodrigues) will return the 62 FPTP Mps. These three same votes will also be considered as votes for the respective parties.  Their unreturned votes will be aggregated at the national level to apportion the 20 list tier seats, subject to meeting the minimum threshold;

 

(xiii)         to provide for the party lists to be submitted to the Electoral Commissioner and published before elections as nominations for the allocation of the 20 seats (say, on Nomination Day itself).  It will inform electors on the composition and rank of the list. This is how it is done in mature democracies as the population has a right to know for whom it is voting;

 

(xiv)       to appoint the next person on the party list to Parliament in case of a vacancy for list tier Mps.  As these Mps are returned on the basis of votes polled by candidates belonging to their respective parties, there will be no by-elections, as is the case in countries with such voting formula. Provided that the newly chosen candidate is still a member of the eligible party.  For instance in Germany, a list member is replaced by the next non-elected candidate on that party’s list, with the proviso that he or she must still be a member of the party.  By-elections would continue to be organised to replace constituency MPs as they are elected ‘nominally’ and not by Party list;

(xv)           to replace the party list Mps in case of crossing of the floor.  This is straightforward as it is the party that has appointed him or her.  Therefore he/she should be automatically replaced by another candidate from the Party if he/she defects. Of course, in case of constituency Mps, the problem is trickier as he/she has been elected ‘nominally’.  Therefore the same procedure cannot apply. He/she retains the seat until the next election.

 

We are still left with the stability/fairness issue. On the one hand, if the method of allocating the 20 list tier seats is too proportional, it will sacrifice stability and effectiveness. On the other, if the formula is too timid on reasonably attenuating the unfairness of FPTP, we are still left with its defect.

 

 

 

 

 

 

13.   Defining some concepts in the new electoral landscape : welcome to the party

 

The 20 list tier Mps will be returned from a closed Party List.  Most countries with some form of Party list system resort to closed list to choose Mps , even if others use an open or a flexible list. MPs. In this system, parties make a list of candidates to be elected and seats are awarded to each party in proportion to the number of votes it receives. Candidates are elected in the order they appear on the Party list until all seats are filled. The list is pre-determined by the party and cannot be changed by the voter.

 

Box 4 gives a list of some countries that use a closed Party list to return Mps. In many of them, this is the only way in which Mps are elected to Parliament.

 

      Box 4:  Some countries using closed Party List to return Mps

  Norway            South Africa           Argentina           Japan

  Spain                 Israel                      Peru                     Uruguay

  Portugal            Mozambique       Iraq                      Angola

  Italy                   Serbia                   Cambodia            Guyana

            Mps  for European Parliament in UK and France

 

 

 

 

 

There are many states with a mixed electoral system that allow double candidacies.  Evidence suggests that in these countries, the overwhelming majority of candidates contest elections both in constituencies and on Party list.  And there are many Mps who alternate between the two modes during their Parliamentary career. In the event of the candidate being returned in the constituency race, his name is automatically struck from the list and the seat goes to the next unelected candidate on the list.  In practice high positions on the list are taken by party members who also stand as candidates in specific districts.  This strategy protects the party in case it loses power; most of its senior members will be elected from the Party List. Double candidacies  also  solve  the  problems  of   two  categories  of   Mps  in  mixed  system   (one contesting in constituency and the other standing on the list) as many candidates on the list are competing in FPTP race.  It gives a safety net to some candidates, especially the leaders of the parties. In Mauritius it will have the added advantage of affording some reassurance to those who may have been penalised at the level of constituencies for some unfortunate reasons. In this case it would operate in a similar manner to the Best Loser System.  One does not necessarily have to be highly placed to become a party list Mp.  It depends how many successful constituency candidates are ranked higher.  In 2005,  for example, one candidate of the CDU in North-Rhine Westphalia in Germany was placed only 33rd on the party’s state list, but was nevertheless elected because 11 other CDU candidates ranked above him won their constituencies.  These winners are thus no longer considered for the seat distribution by list, and other list candidates ranked lower move up.

 

 

45 % of candidates in the German 2005 election ran both in a constituency and on a party list, 37 % was on the party list alone while only 18 % contested just in a constituency.  This practice has spread to countries that have adopted such a mixed system.  The percentage of candidates on the list who did not run in constituencies was only 8.5 % and 3.3 % respectively in the 2003 Scottish and Welsh elections, and only 5.8 % in the 2002 New Zealand election. The strategy seems to have emerged from lessons learned over many elections. With experience, parties fully understand the great advantages of double candidacies.  It is worth noting that double candidacy has allowed Chancellor Helmut Kohl of Germany to enter the Bundestag despite being defeated in his constituency in some elections. And it did not undermine his authority as Chancellor in Germany. Box 4 illustrates some countries with mixed system where double candidacy is very common

 

Box 5  : Some countries where double candidacy  is very common

 

Germany                   Japan

  New Zealand            Lesotho

  Wales                        Scotland

 

 

 

 

 

 

 

 

 

A minimum threshold is usually necessary to be eligible for list tier seats. The threshold can be implicit or explicit. Implicit thresholds are the outcome of quotas from votes cast (as proposed by Carcassonne) while explicit thresholds are predetermined percentages of votes necessary for a party to secure representation ( as recommended by Sachs). Countries that favour broad representation tend to have a low threshold (South Africa is at an extremely low 0.25 % ) while those that seek to exclude extremist or single issue parties fix a high threshold (Turkey at 10 %). In Germany it is either a minimum threshold of 5 % or a requirement to have won 3 constituency seats. It is plain that the higher the threshold, the more difficult it is for small parties to gain Parliamentary representation while the lower it is, the higher the risks of fragmentation of the party system and the polarisation on communal lines that could pose a threat to national unity. It is a very difficult choice to make especially in a multi ethnic society. On balance, I prefer to err on the side of caution and am recommending a 10 % threshold even if it appears high and could be unfair to many genuine, non-communal small parties. The risks of ethnic fragmentation and polarisation is simply too elevated. However I am also recommending the possibility of allowing parties that obtain 5 % of national votes and three FPTP Mps to become eligible for the 20 List tier Mps.

14.   Remember some of Sachs : forward with the past or    back to the future

 

After thoroughly evaluating the various options and following wide consultations and dialogue with many stakeholders and taking into account both contextual factors and the historical, social and political specificities of the country, the Sachs Commission proposed a mixed model that combines features of a plurality system with some proportional elements as contained in Model C.  It advocated the retention of the 62 FPTP Mps and recommended the election of 30 PR Mps along a compensatory PR mode that takes into account the results at constituency levels.  Parties should poll at least 10 % of the national vote to be eligible for the 30 PR seats. It has a very marked preference for subsuming the BLS into the 30 PR seats, thus giving a Parliament of 92 seats. Otherwise there would be 100 Mps.

 

Model C of Sachs is attractive for many reasons and was supported by the major political parties.  It attempts to reach a balance between the two key attributes of stability and fairness. It is familiar to voters as FPTP continues to be the most important component of the mixed system.  The 62 representatives elected via FPTP would ensure stability while fairness would be guaranteed by returning the 30 PR seats through a compensatory formula that would correct the underrepresentation of Opposition parties.  It would also satisfy the other criteria of a good electoral system such as accountability and constituency links. They would also produce a balanced list that reflects the diversity and plurality of Mauritius.  The combination of three-member constituencies, the 10 % threshold and the Party list would not exacerbate fragmentation on communal and ethnic lines.  Thus Model C scores well on the issue of national harmony and social progress as most parties would present Party lists that cut across ethnic lines and promote gender fairness.

 

However it suffers from three drawbacks.  First, there are probably too many seats especially when compared to the number of persons per Mp and this may not be in line with the expectations of the population. Second, there is also no need for 30 seats to cure the defect of the FPTP voting mode; it can be done with fewer additional seats. I recommend that the number of Mps be lowered from 92 to 82. The third drawback was amplified by the context and the circumstances.  The elections to the newly set up Regional Assembly in Rodrigues took place after the submission of the Sachs report.  The elections were held under a similar voting formula to Model C of Sachs. It is a mixed system of 12 FPTP seats elected from 6 two-member local region and 6 PR seats allocated on an island-wide basis on a compensatory basis.  Many believe that the new electoral system produced too proportional a result and could affect the stability of the Regional Government as the leading party’s majority was sharply narrowed after the allotment of the PR seats.  However, one should admit that the final results probably reflected the balance of power in the island and the limited number of seats at stake.

 

Nevertheless it is a fact that Model C of Sachs becomes very proportional when elections are  close in terms of votes.  Under these circumstances, it might be difficult to have stability.  However, this is simply a question of values and they vary considerably across the international spectrum. Should two parties that receive almost the same share of votes obtain very different level of seats ?. The question is simple while the answer is highly complex. Countries that value fairness and have been used to elections that reflect these values (Netherlands, Denmark, South Africa, Germany, New Zealand and Sweden) would strongly argue that seats allocation should be proportional to the votes of parties, even if governability would be very difficult under these circumstances. For them the essence of democracy is fairness and equitable representation.  On the other hand, states that cherish effectiveness and stability and have known this electoral system for a long time have no qualm about disproportionality between seats and votes as long it delivers effective government (United Kingdom, France, USA, Australia, India and Mauritius).  For them the most important attribute of democracy is effective and decisive government.  Even in a mixed system such as Model C of Sachs, this conflict is not resolved in all circumstances. While we value fairness, our proclivity for the strengths of FPTP and our culture of confrontation, conflict and adversary politics as opposed to accommodation, collaboration and consensus are such that many would like elections to be decisive in outcome so as to encourage the quick formation of Government after even a close election.

 

There are various practical ways of addressing the problem of too much proportionality in Model C of Sachs.  I shall mention three possibilities as follows:

 

(i)                 first, it could be tweaked to suit our specific circumstances.  Essentially its application could be limited when it becomes too proportional so as to leave an acceptable margin to the party that has come first in the elections.  We need to decide on the margin (say a maximum of 10 % of seats in a Parliament of 80 members) subject to polling a threshold level of votes.  It has existed in some countries and there are similar mechanisms at work elsewhere.   It has fundamental drawbacks, especially as it artificially creates a majority through some constitutional engineering;

 

(ii)               second, we could allow Model C of Sachs to be fully operational and adjust afterwards by giving some additional seats to the winner of FPTP to produce a majority. Here we have to decide on the premium. It is essentially what has been provided for Rodrigues after the close elections to the Regional Assembly in 2002.  It has many deficiencies especially as it artificially changes the will of the people after the application of  a voting formula;

 

(iii)             third, is a formula that constitutes a departure from the apportionment mechanism used by Model C of Sachs. It also attempts to narrow the disproportionality between seats and votes significantly, but not fully as Model C of Sachs; however it leaves a working majority for the winner of the elections even in closely contested ones.  It is designed to achieve a middle ground between the criteria of stability and the value of fairness. Basically it sits between the current FPTP and Model C of Sachs, while being closer to the latter. It is the UVE formula.


 

15.   Unreturned Votes Elect : can the odd couple of stability and fairness be reconciled?

 

The concept of unreturned votes and its rationale have already been explained . Now it needs to be operationalised to produce results that would address the stability/fairness dilemma. Table 12 gives the extent of  the votes of unsuccessful candidates for the ten elections since 1967.  They are extremely high and are at the heart of the unfairness of the FPTP electoral system.  55 % of votes were used in 1987 to elect Mps while 45 % of votes had no representation at all. The percentages of votes of unreturned candidates were very high at 48 % and 46 % in 2005 and 2010 respectively.  The share of these votes went up significantly to 57 % in 1976 because of the three way electoral contest.  Many Mps were elected on 40 % of the votes which implies that 60 % of votes had no representation in these constituencies.

Table  12: Votes of unreturned candidates as a percentage of total  votes

Year

Votes of unreturned candidates

% of total votes

1967

40

1976

57

1982

36

1983

44

1987

45

1991

44

1995

34

2000

48

2005

48

2010

46

 

 

 

 

 

 

 

 

 

 

The results of the general elections of 1987 are used to illustrate the workings of the UVE mechanism.  The outcome of that election is given at Table 13.

Table 13 :   Results of 1987 general elections

% votes

Elected MPs

% of seats

Best Losers

Total seats

 

% of seats

without BLS

with BLS

MSM/LP/PMSD

49.86

39

62.9

5

44

62.9

MMM & Allies

48.12

21

33.9

3

24

34.3

Other

2.02

2

3.2

0

2

2.9

Total

100

62

100

8

70

100

Majority

18

 

The MSM/LP/PMSD alliance captured 49.86 % of the votes and won 44 seats (39 directly elected Mps and 5 Best Losers) while the MMM and its allies polled 48.12 % of votes and obtained 24 seats (21 directly elected and 3 Best Losers).  There is thus a huge degree of disproportionality between seats and votes.  The winner took 62.9 % of the seats with only 49.86 % of the votes while the second party received only 34.3 % of the seats for 48.12 % of the votes.  The Government had a majority of 18 seats. What is required is a formula that will reward the second party for having been penalized by the FPTP formula, without making it extremely difficult for the winner to govern.

 

We proceed by computing the votes of all candidates that have not been returned in each constituency, provided their parties have polled at least 10 % of national votes ( or the 5 % plus 3 FPTP Mps formula) .  In constituencies where three members of the same party have been elected, the formula will calculate the votes of the three candidates of the other parties.  In constituencies where candidates of both parties are returned (or for that matter of three parties like in Port Louis Maritime/Port Louis East in 2010), the votes of unelected candidates of all the parties are taken into account.  This is carried out for all 21 constituencies and the sum of such votes from unreturned candidates  is aggregated to reach a national figure for both parties (or more if necessary). And the parties are allocated seats proportional to their share of votes of unreturned candidates nationally.  In cases of one party winning all FPTP seats, as in 1982 and 1995, all the 20 additional tier seats will go to other parties as there will not be any unreturned candidates of the winning party.

 

For instance, in Grand Riviere North West/Port Louis West, the MMM won all three seats in 1987 and the three candidates of the second party obtained a total of 26979 votes.  These  votes are credited to the account of the MSM/LP/PMSD alliance.  In Grand Baie/Poudre D’Or, the MSM/LP/PMSD captured all three seats. The three unreturned MMM candidates polled 34824 votes and these are given to that party.  In Rivière Des Anguilles/Souillac, the MSM/LP/PMSD won 2 seats while the MMM returned one MP. In that case, the votes of the two unsuccessful candidates of the MMM and the one of the MSM/LP/PMSD alliance are reckoned as unreturned votes. The principle is simple. There is compensation for parties whose candidates have polled a high percentage of votes but which have been totally ignored by the FPTP formula.  However this is done on a constituency by constituency basis and then aggregated nationally so as to reward parties. The benefit accruing to each party  will   depend  on the extent that it has been penalized at constituency level.

 

The simulated computation of unreturned votes for the elections of 1987 are produced at Table 14 below.  In constituencies where one party has won all three seats, it will not receive any ‘unreturned votes’ as it has already taken the three seats.  The second party will earn some credit for the purpose of the allotment of the 20 additional seats.  In cases where there are split votes in a constituency, both parties will have votes from unreturned candidates; however the party that has returned two Mps will have a lower number of such votes than the one with only one elected representative.    As the MMM was disadvantaged by the FPTP contest, it would have more votes from unreturned candidates.  However it is not a one way traffic as is the case in Model C of Sachs formula.  The party that came first but lost in some constituencies is also rewarded.  As the MMM received 492468 votes from its unreturned candidates and the MSM/LP/PMSD 224764 , the share of the additional tier seats is divided in relative proportion between the different parties, using the D’Hondt formula. Parties that receive less than 10 % of national votes ( or 5 % plus 3 FPTP Mps) are not eligible for these seats.  Thus the MMM would have received 14 additional seats while the MSM/LP/PMSD would be entitled to 6 more Mps.  The candidates are chosen from the rank-based closed Party List.  14 candidates on the MMM list are returned as list tier Mps (some will be candidates who have lost in constituencies and whose name also appear on the Party List). Similarly, 6 candidates would be elected on the MSM/LP/PMSD party list . If some who appear on the Party list are elected in constituency contests, their names are removed and those further down are chosen.

 

 

 

 

 

 

 

 

Table  14 :  Seat allocation using  votes of unreturned candidates (1987) (20 seats)

 

1987

 

Votes of unreturned candidates

Constituency

MSM/LP/PMSD

MMM

 

Total

1

26979

26979

2

24553

24553

3

10998

10998

4

28181

14342

42523

5

40895

40895

6

34824

34824

7

30748

30748

8

33014

33014

9

40330

40330

10

40473

40473

11

28185

28185

12

31591

31591

13

11638

23385

35023

14

17074

32853

49927

15

33088

17166

50254

16

14817

29196

44013

17

38161

38161

18

43600

43600

19

30985

30985

20

26451

13705

40156

Total

224764

492468

717232

% of wasted votes

31.3

68.7

100.0

Additional Tier Seats

6

14

20

 

 

The final outcome of the 1987 elections with this new formula would be as given at Table 15.

 

Table 15 : Simulation of 1987 elections with UVE formula

 

% votes

FPTP Mps

% FPTP Mps

Additional Tier Mps

Total seats

 

% Total seats

Distribution

MSM/LP/PMSD

49.86

39

62.9

6

45

54.9

MMM & Allies

48.12

21

33.9

14

35

42.7

Other

2.02

2

3.2

0

2

2.4

Total

100

62

100

20

82

100

Majority

8

 

This is a significantly better outcome in terms of fairness than the FPTP formula.  It is at the same time superior in terms of stability than Model C of Sachs.  The party with 48.12 % of the votes now has 42.7 % of the seats compared to 34.3 % in the current FPTP formula (see Table 13).  There is now a difference of  only 5.4 % between seats and votes of the second party (42.7 % of seats to 48.12 % of votes).  This is very acceptable by international standards. The winning party still has a working majority of 8 seats (equivalent to around 10 % of the number of MPs) compared with 18 under FPTP.  The majority is much higher than in Model C of Sachs (which was zero).

 

Table 16 shows the simulation for the four very disproportionate elections of 1982 and 1995 (in both cases one alliance won all 60 seats) and of 1991 and 2000 (in both elections, one party with around 50 % of votes obtained almost 90 % of seats).  The proposed model based on compensating parties on the basis of the votes of unreturned candidates performs as expected.  In the two cases of 60-0, all the 20 PR seats would accrue to the second party.   In 1982, the PAN alliance would receive around 25 % of seats, very close to its vote tally of 25.78 %.

Table 16:    Simulation of 1982, 1995, 1991 and 2000 elections with UVE formula

1982

% votes

FPTP Mps

% FPTP Mps

Additional Tier Mps

Total seats

% Total seats

MMM/PSM

64.16

60

96.8

0

60

73.2

PAN

25.78

0

0.0

20

20

24.4

Other

10.06

2

3.2

0

2

2.4

Total

100

62

100.0

20

82

100

1995

LP/MMM

65.2

60

96.8

0

60

73.2

MSM

19.7

0

0.0

20

20

24.4

Other

15.1

2

3.2

0

2

2.4

Total

100

62

100.0

20

82

100

1991

MSM/MMM

56.3

57

91.9

1

58

70.7

LP/PMSD

40

3

4.8

19

22

26.8

Other

3.7

2

3.2

0

2

2.4

Total

100

62

100.0

20

82

100

2000

MSM/MMM

51.5

54

87.1

2

56

68.3

LP/PMSD

36.7

6

9.7

18

24

29.3

Other

11.8

2

3.2

0

2

2.4

Total

100

62

100.0

20

82

100

 

Similarly in 1995 the MSM would have kept all 20 PR seats, representing around 24.5 % of seats for a vote tally of 19.7%.  However in both cases, the winning party retains a considerable majority (38 in both cases).  The country would have had a strong contingent of opposition members in Parliament.  This result is not different from what would happen under Model C of Sachs. In 1991 and 2000, the second party which was penalized in FPTP contest would have received 19 and 18 PR seats respectively. It obtained only 3 and 6 FPTP seats and the compensation is very similar to what Sachs would have provided. It is far better in terms of fairness than the FPTP outcome and the winner still has a majority to govern. Overall, the LP/PMSD alliance would have more than 20 Mps compared to the 7 and 8 that it actually obtained. This makes a major difference for Parliamentary democracy and for fairness. However there is still some disproportionality between votes and seats.

 

Table 17 gives the simulations for the five remaining general elections held since 1967 (1967, 1976, 1983, 2005 and 2010).  Again they produce a good balance between the two competing attributes of stability and fairness.  In 1967, it produced a highly proportional result but it also generated a good working majority of 10 seats.  The elections of 1976 would have necessitated a post-election coalition in all cases. The results of the 1983 elections are very fair with the MMM receiving 40 % of seats for 46 % of votes and the majority for the winning team is reasonable at 12 seats (compared to 22 under FPTP and to 2 under Model C of Sachs). The same trend is observed for the general elections of 2005 and 2010 with a degree of deviation between seats and votes for the MMM at 2 % and 4 % respectively.  This is a highly equitable result indeed as it is extremely difficult to expect a pure proportional outcome. And in both cases, the working majority of 10 Mps and 12 Mps respectively are much lower than those produced by FPTP formula (14 and 21 respectively) but higher than those created by Model C of Sachs (at 4 in both cases).

 

 

 

Table 17:    Simulation of 1967, 1976, 1983,   2005,   2010 elections with UVE formula

 

1967

% votes

FPTP MPs

% FPTP MPS

Additional Tier Mps

Total seats

% Total seats

IP

54.13

39

62.9

7

46

56.1

PMSD

43.99

23

37.1

13

36

43.9

Other

1.88

0

0.0

0

0

0.0

Total

100

62

100.0

20

82

100

Majority

10

1976

MMM

38.64

30

48.4

7

37

45.1

IP

37.9

25

40.3

8

33

40.2

PMSD

16.2

7

11.3

5

12

14.6

Other

7.26

0

0.0

0

0

0.0

Total

100

62

100.0

20

82

100

Majority

Hung

1983

MSM/LP/PMSD

52.22

41

66.1

6

47

57.3

MMM

46.4

19

30.6

14

33

40.2

Other

1.38

2

3.2

0

2

2.4

Total

100

62

100.0

20

82

100

Majority

12

2005

LP/PMSD

49.39

38

61.3

8

46

56.1

MSM/MMM

43.3

22

35.5

12

34

41.5

Other

7.31

2

3.2

0

2

2.4

Total

100

62

100.0

20

82

100

Majority

10

2010

LP/MSM/PMSD

50.72

41

66.1

6

47

57.3

MMM

42.88

18

29.0

14

32

39.0

Other

6.4

3

4.8

0

3

3.7

Total

100

62

100.0

20

82

100

Majority

12

Could there be any unintended consequence in the application of the UVE formula?. Ten elections are quite a reasonable sample to test the robustness of an electoral formula.  It behaves well in all 10 simulations with respect to the balance between fairness and stability.  Of course Sachs Model C would produce slightly more proportional results in some elections but at the expense of stability and effectiveness.  The Carcassonne model could produce equitable result at the national level with fairly disproportional outcome at the level of each constituency.  However both Sachs Model C and Carcassonne could be extremely weak on stability when results are close in terms of percentage of votes.


 

16.   Between FPTP and Sachs :  the middle is better than the extremes

 

Table 18 makes the comparison between three different electoral systems on the two dimensional plane of stability and fairness.  The three voting formulae are FPTP, Model C of Sachs and UVE . Stability is measured as the margin of seats that the winning party has in Parliament.  The lower the margin, the less stable the system is deemed to be.  The difference between the percentage of seats obtained by the second party and its share of votes is used as a proxy for fairness.  The higher the negative divergence between the two, the more unfair the formula is. As expected, the FPTP produces the largest majority in all elections. It ranges from 14 seats in 2005 to 54 seats in 1982 and 1995. Sachs Model C lowers considerably  the majority in 5 elections; it varies from no majority in 1987, to a majority of 2 seats, 4 seats, 4 seats and 8 seats in 1987, 1983, 2005, 2010 and 1967 respectively. The UVE formula sits between FPTP and Sachs Model C (closer to the latter ) in the five hotly contested elections of 1967, 1983, 1987, 2005 and 2010 with a majority varying from 10 seats in 1967 and 2005, 12 seats in 1983 and 2010 and 8 seats in 1987. Essentially it gives a working margin to the winner. In 1995, it delivers all 20 seats to the MSM while Sachs operates slightly differently. The 1976 elections delivered a hung Parliament in all three modes. It should be pointed out that in 1982, Sachs Model C and UVE produce the same majority of 38 seats while in 1991 and 2000, the UVE formula gives a few more seats to the winning party compared to the Sachs formula.

 

 

 

 

 

Table 18: FPTP, Sachs C and UVE : which stability and which fairness?

 

Stability ( Y-X )

Fairness ( % of votes and seats of second party)

Year

FPTP

 

Sachs C

UVE

 

 

FPTP

Sachs C

UVE

Margin of seats

Margin of seats

Margin of seats

% of votes of second party

% of seats of second party

% of seats of second party

% of seats of second party

1967

16

8

10

43.99

38.57

45.12

43.90

1976

Hung

Hung

Hung

37.90

40.00

41.46

40.24

1982

54

38

38

25.78

3.03

24.39

24.39

1983

22

2

12

46.40

31.43

46.34

40.24

1987

18

0

8

48.12

34.29

47.56

42.68

1991

48

32

34

40.00

10.61

28.05

26.83

1995

54

42

38

19.70

0.00

21.95

24.39

2000

46

26

30

36.70

11.43

31.71

29.27

2005

14

4

10

43.30

34.29

45.12

41.46

2010

21

4

12

42.88

28.99

43.90

39.02

Y-X  is the difference between seats of winning party and those of other parties combined

 

The FPTP is the most unfair formula as shown by the very high difference between seats and votes in many elections. In 1995, the second party polled 19.7 % of votes without a single seat while in 1982 , it attracted 3 % of seats on 25.78 % of votes. There are also large differences between votes and seats in 1991 and 2000 and some distortion in 1983, 1987, 2005 and 2010.   Model C of Sachs, which is a mixed member proportional system, produces the fairest results in most elections. In fact in some cases, the percentage of seats is higher than the share of votes (1967, 1976, 1995, 2005 and 2010). This is attributable to the threshold of 10 % ( or its alternative of 5 % and 3 FPTP Mps) which excludes parties below that level from participating in the allotment of the 20 list tier seats. If we adjust for the threshold, Model C of Sachs becomes a full PR system in 8 out of the ten elections, even if it a mixed FPTP/PR system. The only reason why it is not proportional in 1991 and in 2000 is the number of PR seats which is fixed at 20. If there were 30 seats, all 10 elections would become very proportional. The UVE formula lowers the disproportionality between seats and votes considerably ; however it does not transform the electoral system into a pure proportional model. It is very fair to the second party in all the 10 elections. In 1987 the MMM and allies would have captured 42.68 % of seats with 48.12 % of votes while in 1983 it would have received 40.24 % of seats with 46.4 % of the votes. The same pattern emerges in 2005 and 2010. There are elections where there is either no difference between the two (1982) or a very marginal one (1991).  The high difference between seats and votes in 1991 is attributable to the fact that there are not enough seats to compensate the LP/PMSD alliance for its relatively high share of votes at 40 %.  The greatest merit of the UVE formula is to bring down the level of the FPTP unfairness significantly while affording a working majority to the winner.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

17.   Conclusion: Have we reached the goal post?

 

Choosing the right electoral system in a plural society is a very difficult process as it is never neutral.  It has very clear implications on the political development of the country.  Historical, political, social, cultural, contextual and temporal factors have to be taken into account as the system and any reform do not take place in a vacuum.  A  lot of emphasis is placed  on characteristics such as stable, effective, responsive and accountable government, fairness, social and political representation and inclusiveness while features like voter’s choice, links between elected representatives and their constituents are also important.  It is almost impossible for one electoral formula to satisfy all these attributes.  In fact some of them are mutually exclusive as often it is possible to achieve one particular objective only at the expense of another.  The exercise is complicated by the fact that electoral systems are not disinterested in their translation of votes into seats.  It depends on many circumstances, on what the country wants to achieve and what weights it assigns to each of the conditions of a good system.  Finally it depends on the willingness and the desire of political actors to compromise in order to reach an acceptable and a sustainable consensus.

 

Often a country cannot choose a perfect system but has to settle for one that is significantly less imperfect than other formulae.  The best voting system for a country is not one that satisfies only one criterion completely, but one that provides a fair balance among the different attributes.  As no electoral system could possibly meet all these criteria, the objective should be a solution that is satisfactory rather than optimal.  A prudent approach is to design an electoral system that avoids serious shortcomings.

 

I have attempted, through research work, wide-ranging consultations,comparative analyses of voting formulae in plural societies, best international practices, an understanding of the historical perspectives of electoral systems of our country and an appreciation of its specific realities, to make an independent and an objective contribution in respect of some fundamental issues raised by the Trustram-Eve (1958), de Smith (1964), Banwell (1966), Sachs (2002), Collendavelloo (2004) and Carcassonne (2011) reports.  After appraising their various proposals, I am recommending a mixed electoral system with 62 FPTP Mps in three member constituencies (2 for Rodrigues) and 20 additional tier Mps elected from a closed rank- ordered party list.   The seats are allocated on a uniquely designed formula based on the votes on unreturned candidates at general elections.

 

Even if all systems have both defects and virtues, some are however better than others.  The proposed electoral system is not a perfect one but it has the features to provide the most satisfactory overall balance between them in a multi-ethnic society, especially between stability and fairness.  The ‘specially designed’ mixed electoral system of 62 FPTP in three member constituencies and 20 additional tier Mps meets all the requirements of a good electoral system in a plural society.  Not all of them perfectly but very satisfactorily.  It produces stability and effectiveness as it gives a working majority to the winning party even in hotly contested elections.  It is fair and equitable as it narrows considerably the unacceptable distortion between seats and votes, especially for the second party.  It achieves diversity and inclusion in political representation by using an intelligent combination of the specific features of the FPTP and some well-crafted characteristics of an additional tier electoral mode.  It is very inclusive of women with the proposal to have no more than two candidates of the same sex in each of the 20 constituencies and at least one person of a different gender  out of every three sequential candidates on the Party List. It ensures accountability by keeping the essential link between constituents and their elected representatives as 75 % of Mps will depend on the choice of electors and only 25 % will be chosen by the Party. With a threshold of 10 % ( or 5 % plus 3 FPTP Mps) to be eligible for national seats, it will also shun single issue parties, promote harmony and foster nationhood.

 

Never has the country been so close to reaching an agreement among the main political parties on reforming the electoral rules. All agree on the need for reform. Then there is a cooling ardour and it starts all over again.  Electoral reform has featured prominently in the electoral manifestoes of all major political parties for a long time.  Our political leaders continue to curse the deficiencies of the FPTP voting formula and extol the virtues of greater women participation and more fairness. Many of our compatriots look forward to finding an acceptable solution to the BLS while ensuring inclusion and broad-based socio-demographic representation. And they all insist on the need to have stability and decisive government. All these four crucial questions have been comprehensively addressed in this document. The necessary consensus is there and we need to build on that momentum to accelerate the process. The ball is now in the court of political actors.  Reformers have played a very long and patient game and the status quo does not appear to be a sustainable option anymore. It is only a question of timing when reform will actually happen.  The balance of probabilities appears to be on the side of reformers and the deepening and the widening  of democracy in our Motherland.

 

 

Rama Sithanen

 

Appendices

 

Appendix 1: Allocation of PR seats in a compensatory mode using D’Hondt’s Highest Average Method

 

The worked example below shows how the D’Hondt formula allocates PR seats in a compensatory mode. It is illustrated with the results of the 2007 elections.

 1987

% votes

FPTP

seats

Sequence of seats

40

41

42

43

44

45

Highest Average

1.247

1.216

1.187

1.16

1.133

1.108

MSM/LP/PMSD

49.86

39

Sequence of seats

22

23

24

25

26

27

28

29

30

Highest Average

2.187

2.092

2.005

1.925

1.851

1.782

1.719

1.659

1.604

MMM & Allies

48.12

21

Sequence of seats

31

32

33

34

35

36

37

38

39

Highest Average

1.552

1.504

1.458

1.415

1.375

1.337

1.301

1.266

1.234

Other

2.02

2

100

62

Seat allocation

MSM/LP/PMSD

2

18

20

MMM & Allies

18

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

19

Final Outcome

FPTP

PR

Total

MSM/LP/PMSD

39

2

41

MMM & Allies

21

18

39

Other

2

0

2

Total

62

20

82

 

The MSM/LP/PMSD alliance won 39 seats ( or 65 % of seats) with 49.86 % of the votes while the MMM and allies took 21 seats ( or 35 % of seats) with 48.12 % of votes. Successive quotients are calculated for each party. The formula for the quotient is the percentage of votes divided into the number of seats already won by each party plus one.  Whichever list has the highest average gets the next seat and its quotient is recalculated given its new seat total. The process is repeated until the 20 seats are allocated. It simply attempts to keep the ratio of votes polled to seats allotted as close as possible.  The MSM/LP/PMSD alliance first quotient is obtained by dividing 49.86 % into 40 ( 39 seats plus one), which gives a figure of 1.247. That of the MMM and allies is reached by dividing 48.12 % into 22 (21 seats plus one) and it yields 2.187. This is due to the fact that the MMM and allies have been penalised in FPTP contest. The MMM is entitled to the first seat as it has a higher quotient. The next quotient of the MMM is now obtained by dividing 48.12 into 23 ( as it has won one additional seat). This now stands at 2.092 which is still higher than 1.247.  The process continues until all 20 seats are allotted.  The first 17 seats accrue to the MMM and allies as its quotient remains higher than 1.247 which is the quotient at which the MSM/LP/PMSD alliance would be entitled to its first seat. The MSM/LP/PMSD receives the 18 th and 20 th seats. Of the 20 PR seats, the MMM and allies captures 18 seats while the MSM/LP/PMSD takes 2 seats. Overall, the MSM/LP/PMSD ends up with 41 seats, the MMM and allies with 39 seats and Rodrigues with 2 Mps.

 

 

 

 

 

Appendix 2:  Apportionment of 20 additional UVE seats in 1967

1967

 

 

Votes of unreturned candidates

 

Constituency

IP

PMSD

 

Total

1

17379

17379

2

14371

14371

3

12908

12908

4

19607

19607

5

12320

12320

6

13476

13476

7

5333

5333

8

10758

10758

9

11846

11846

10

15457

15457

11

10972

10972

12

14899

14899

13

15444

15444

14

19678

19678

15

20306

20306

16

21458

21458

17

13965

13965

18

18438

18438

19

17846

17846

20

18024

18024

Total

112931

191554

304485

% of  votes

37.1

62.9

100.0

Additional Tier Seats

7

13

20

 

 


 

 Appendix 3:  Apportionment of 20 additional UVE seats in 1976

1976

 

Votes of unreturned candidates

Constituency

IP

MMM

PMSD

 

Total

1

14809

17502

32311

2

15605

9048

24653

3

9260

5899

15159

4

22682

13643

36325

5

21367

4715

26082

6

17103

8833

7655

33591

7

7538

12272

2597

22407

8

8883

16186

2413

27482

9

15913

8413

548

24874

10

16833

8858

25691

11

17345

17345

12

14497

7541

3699

25737

13

13270

7772

3418

24460

14

9761

19427

6124

35312

15

28419

11275

39694

16

20908

17918

38826

17

11800

16108

27908

18

6572

13140

16884

36596

19

17529

19430

36959

20

21309

15911

8185

45405

Total

223364

222500

150953

596817

% of  votes

38.3

37.1

24.6

100.0

Additional Tier Seats

8

7

5

20

 

 


 

Appendix 4:  Apportionment of 20 additional UVE seats in 1982

1982

 

Votes of unreturned candidates

Constituency

PAN

MMM/PSM

 

Total

1

10453

10453

2

12806

12806

3

2069

2069

4

19820

19820

5

26420

26420

6

22107

22107

7

13046

13046

8

15778

15778

9

26128

26128

10

22184

22184

11

16782

16782

12

15667

15667

13

14875

14875

14

27252

27252

15

27251

27251

16

21541

21541

17

16082

16082

18

17780

17780

19

16118

16118

20

19360

19360

Total

363519

0

363519

% of votes

100.0

0.0

100.0

Additional Tier Seats

20

0

20

 

 

 


 

Appendix 5 :  Apportionment of 20 additional UVE seats in 1983

1983

 

Votes of unreturned candidates

Constituency

MSM/LP/PMSD

MMM

 

Total

1

25253

25253

2

18578

18578

3

4929

4929

4

32861

32861

5

31859

31859

6

28461

28461

7

17828

17828

8

28133

28133

9

31897

31897

10

31821

31821

11

21679

21679

12

25550

25550

13

27969

27969

14

38662

38662

15

28263

14207

42470

16

35145

35145

17

29206

29206

18

31910

31910

19

31341

31341

20

27422

14247

41669

Total

168647

408574

577221

% of votes

29.2

70.8

100.0

Additional Tier Seats

6

14

20

 

Appendix 6 :  Apportionment of 20 additional UVE seats in 1991

1991

 

Votes of unreturned candidates

Constituency

LP/PMSD

MSM/MMM

 

Total

1

20909

20909

2

18527

18527

3

8565

8565

4

30272

30272

5

29608

15012

44620

6

41904

41904

7

37102

37102

8

32651

32651

9

36536

36536

10

36659

36659

11

23793

11961

35754

12

21945

11220

33165

13

29864

29864

14

40512

40512

15

41750

41750

16

41040

41040

17

37840

37840

18

40864

40864

19

24403

24403

20

32947

32947

Total

627691

38193

665884

% of  votes

94.4

5.6

100

Additional Tier Seats

19

1

20

 

Appendix 7:  Apportionment of 20 additional UVE seats in 1995

1995

 

Votes of unreturned candidates

 

Constituency

 

LP/MMM

 

MSM/RMM

 

 

Total

1

7354

7354

2

8246

8263

3

2710

2710

4

16209

16209

5

17859

17859

6

13853

13853

7

29251

29251

8

24375

24375

9

29168

29168

10

16867

16867

11

20852

20852

12

13677

13677

13

14968

14968

14

22294

22294

15

23580

23580

16

18355

18355

17

11896

11896

18

19020

19020

19

7145

7145

20

14013

14013

Total

0

331709

331709

% of  votes

0

100

100

Additional Tier Seats

0

20

20

 

 


 

Appendix 8:  Apportionment of 20 additional UVE seats in 2000

2000

 

Votes of unreturned candidates

Constituency

LP/PMSD

MMM/MSM

 

Total

1

30371

30371

2

12590

6936

19526

3

10417

5182

15599

4

33566

33566

5

30876

16001

46877

6

25654

13922

39576

7

22177

14444

36621

8

33382

33382

9

39117

39117

10

29255

29255

11

23140

13783

36923

12

30846

30846

13

27223

27223

14

47766

47766

15

36945

36945

16

33441

33441

17

34825

34825

18

36794

36794

19

21737

21737

20

31230

31230

Total

591352

70268

661620

% of votes

89.4

10.6

100

Additional Tier Seats

18

2

20

 

Appendix 9:  Apportionment of 20 additional UVE seats in 2005

2005

 

Votes of unreturned candidates

Constituency

LP/PMSD

MMM/MSM

 

Total

1

34876

34876

2

21467

21467

3

6155

12400

18555

4

28319

14258

42577

5

45049

45049

6

41959

41959

7

30098

30098

8

14623

29070

43693

9

42520

42520

10

16321

29199

45520

11

39996

39996

12

27575

27575

13

32208

32208

14

39254

20049

59303

15

18450

35194

53644

16

30240

14858

45098

17

37598

37598

18

41949

41949

19

31801

31801

20

33660

33660

Total

291297

477849

769146

% of  votes

37.9

62.1

100

Additional Tier Seats

8

12

20

 

 


 

Appendix 10: Apportionment of 20 additional UVE seats in 2010

2010

 

Votes of unreturned candidates

Constituency

LP/MSM/PMSD

MMM

 

Total

1

34464

34464

2

7709

14057

21766

3

11119

11804

22923

4

14544

29383

43927

5

48225

48225

6

38145

38145

7

32998

32998

8

42782

42782

9

37383

37383

10

44978

44978

11

31818

31818

12

31867

31867

13

35110

35110

14

21630

40802

62432

15

53090

53090

16

16278

31422

47700

17

42592

42592

18

14662

28644

43306

19

32639

32639

20

36979

36979

Total

232616

552508

785124

% of  votes

29.6

70.4

100.0

Additional Tier Seats

6

14

20

 

Selected Bibliography

 

Ahnee R (2001)   :       Report on proposed decentralised system of Government for Rodrigues

Anckar C             :       Effects of Electoral Systems: A study of 80 countries.

Banwell (1966)    :       Report of the Commission on the Electoral system for Mauritius.

Benoit K &

Shepsle K.A       :       Electoral systems and minority representation.

 

Bird K                 :       The Political Representation of Women and Ethnic Minorities in established democracies

Blais A                 :      Criteria for assessing electoral systems

Bogdanor V &

Butler D               :     Democracy and Elections: Electoral Systems and their political

Consequences

Bowler S &

Farrell D               :     ‘So Which Electoral System Is Best? Let’s Ask the Anoraks’.

Cannon D              :     Race, redistricting and representation: the unintended consequences of black majority districts

Carcassonne

& al                      :     Report to the PM on Electoral Reform (Dec 2011)

Carroll S                :     The impact of women in public office

Collendavelloo I

(2004)                   :     Report of Select Committee on introduction of a measure of PR in our electoral system

Dahlerup D           :     From a small to a large minority: Women in Scandinavian Politics.

David M &

Dinan P (1999)     :     Réforme Electorale a Maurice

 

De Smith S.A

(1964)                  :     Report of the Constitutional Commissioner

 

De Ville J &

Steytler N              :     Choosing an electoral system

 

De Ville J              :     Proportional Representation in multi-member constituencies

Dunleavy P &

Margetts H            :     Understanding the Dynamics of Electoral Reform.

 

Electoral Task

Team Round table :     Electoral Models for South Africa

 

Elklit J                   :     What could a new South African electoral system look like?

Elklit J                   :     Lesotho 2002: Africa’s first MMP elections.

Farrell D                :     Comparing electoral systems

Farrell D &

Gallagher M          :     Submission to the Independent Commission on the Voting System.

 

Faure M &

Venter A              :     Electoral systems and accountability

 

Fick G                   :     The Constitution and the electoral system in South Africa

Fraenkel J              :     Disunity through Diversity? The contours of Fiji’s post-putsch elections

Fraenkel J &

B Grofman            :     Evaluating the impact of electoral reform on inter-ethnic accommodation

 

Gallagher M          :     Proportionality, disproportionality and electoral systems

Gallagher M          :     Comparing Proportional Representation electoral systems: quotas, threshold, paradoxes and majorities

Gallagher M          :     The political impact of electoral system change in Japan and New Zealand

Gallagher M &

Mitchell P              :     The Politics of Electoral Systems.

 

Guinier Lani          :     The tyranny of the majority: fundamental fairness in representative democracy

Hartmann C                :     The Strengths and Weaknesses of existing electoral models as instruments for mitigating ethnic conflicts.

Horowitz D                 :     Democracy in Divided Societies

Horowitz D                 :     Encouraging electoral accommodation in divided societies

IDEA                          :     Electoral System Design. The International IDEA Handbook

IDEA                          :     Women in Politics: Beyond the Numbers.

IPU                             :     Women in Politics

Ishiyama John             :     Electoral rules and party nomination strategies in ethnically leaved societies

Jenkins Lord               :     The Report of the Independent Commission on the UK  voting system

Johnson N                   :     Electoral Reform: The Risks of Unintended Consequences

Lal Brij.V                    :     Electoral Systems in Divided Societies

LeDuc L, Niemi

R.G,   Norris P            :     Voting in democracies: a study of majority and PR systems.

 

Lennox Boyd A

(1958)                         :     Despatch of Secretary of State to Sir Malcolm Trustram Eve.

 

Lennox Boyd A

(1958)                         :     Despatch of Secretary of State to the Governor of Mauritius.

 

Lijphart A                   :     Democracy in Plural Societies: A Comparative Exploration.

Lijphart A                   :     Degrees of Proportionality of Proportional Representation Formula

Lijphart A                   :     Electoral systems, Party systems and conflict management in segmented societies.

Lijphart A                   :     Constitutional choices for new democracies

Lijphart A &

Grofman Bernard

(1986)                          :     Electoral Laws and Their Political consequences

 

Lijphart A &

Grofman Bernard       :     Choosing an Electoral System: Issues and Alternatives

 

Lodge T                      :     Which Criteria Are More Important in Choosing an Electoral System in the Post-1999 South Africa

 

 

 

London Agreement

(1956)                       :     Text of the London Agreement. Conference of the Constitution of Mauritius.

 

Lovenuski J                 :     Women and Politics: minority representation or critical mass?

Lovenduski J &

Norris P                      :     Gender and Party Politics.

 

Lublin D                      :     The Paradox of Representation: Racial Gerrymandering and Minority interests

Mansbridge J               :     ‘Should Blacks represent blacks and Women represent Women?

Matland R.E               :     Enhancing Women’s political participation: legislative recruitment and electoral systems.

Matlosa K                   :     Ballots or Bullets: Elections and conflict management in Southern Africa

Mathur H                    :     Parliament in Mauritius

Meade J.E &

al (1961)                      :     The economics and social structure of Mauritius

 

Nohlen D                    :     Changes and Choices in Electoral Systems

Nohlen D                    :     Elections and Electoral Systems

Northern Ireland

Forum                         :     Committee on Electoral Reform in Northern Ireland.

 

Norris P                       :     The politics of electoral reform

Norris P                       :     Choosing electoral systems: proportional, majoritarian and mixed systems

Norris P                       :     ‘Ballots Not Bullets:         Testing Consociational Theories of Ethnic Conflict,

 

Norris &

Lovenduski                 :     Political Recruitment: ‘Gender, Race and class in the British Parliament.

 

Norton P                     :     The Case for First-Past-The-Post

Phillips A                    :     The Politics of Presence

Pinto-Duschinsky M   :     ‘Send the Rascals Packing: Defects of Proportional Representation and the Virtues of the Westminister Model’

Pitkin Hanna               :     The Concept of Representation

Plant R                        :  Criteria for electoral systems: The Labour Party and Electoral Reform’.

Rae Douglas                :     The Political Consequences of Electoral Laws

Reilly B                       :     Electoral systems choice in Post 1999 South Africa: an alternative model.

Reilly B                       :     Democracy in Divided Societies: Electoral Engineering for Conflict Management

Reilly B &

Reynolds A                 :     Electoral Systems and Conflict Management in Divided societies

 

Reynolds A &

Grofman B                  :     Choosing an electoral system for the new South Africa: the main proposals’.

 

Reynolds A & Sisk T :     Electoral systems and Conflict Management in Africa

Rule Wilma                 :     Why women don’t run. The critical contextual factors in Women’s legislative recruitment

Rule W &

Zimmerman J              :     Electoral systems in comparative perspective: their impact on women and minorities

 

Sachs Commission

(2002)                        :     Report on Constitutional and Electoral Reform.

 

Saggar S                      :     Race and Political Representation Reconsidered

Shugart MS                 :     Electoral Efficiency’ and the Move to Mixed -Member Systems

Shugart MS &

Wattenberg MP           :     Mixed Member Electoral Systems: The Best of Both Worlds?

 

Sithanen R                  :     Evaluating and proposing electoral systems for plural societies: the case of Mauritius. PhD thesis

Sithanen R                  :  Submission to the Commission on Constitutional and Electoral Reform ( 2001)

Sithanen  R                  :    A proposal for electoral reform (2001)

Sithanen  R                  :    Submission to the Select Committee on Electoral Reforms (2002)

Sithanen   R                  :    Which form of mixed system is best for Mauritius (2002)

Sithanen   R                :      A pathway for stable government, fairness and rainbow representation (2002)

Sithanen   R                 :     A critical appraisal of the Select Committee report on Electoral Reforms in Mauritius (2004)

Sithanen  R                  :     Why simplicity breeds perversion and imperils social harmony (2004)

Sithanen R                  :     Coalition politics under the tropics: A case study of Mauritius

Sithanen R                   :     Avoiding instability and communal risks in a mixed electoral system (2004)

Sisk T                          :     Choosing an Electoral System: South Africa seeks new ground rules

Stonehouse J (1966)    :     Report in Sessional Paper, No. 3.

Subramanian A &

Roy D                         :     Who can explain the Mauritian miracle: Meade, Romer, Sachs or Rodrick?

 

Swain C                       :     Black Faces, Black Interests: The Representation of African Americans in Congress

Taagepera R                 :     Designing electoral rules and waiting for an electoral system to evolve.

Taagepera R                 :     How Electoral Systems Matter for Democratisation

Taagepera R                 :     The Tailor of Marraketh: Western Electoral Systems Advice to Emerging Democracies

Taagepera R &

Shugart M S                 :     Seats and Votes: The Effects and Determinants of electoral systems

Tate K                          :     The political representation of blacks in Congress: Does Race Matter?

Tate K                          :     Black Faces in the Mirror: African Americans and their representation in the US congress.

Trustram Eve M

(1958)                         :     Report of the Mauritius Electoral Boundary Commission.

 

Whitby K                     :     The Colour of Representation: Congressional Behaviour and Black Constituents

 

List of persons and organizations met

 

Dr  Navin Ramgoolam, Prime Minister and Leader of the Labour Party

Paul  Berenger, Leader of the Opposition and Leader of the MMM

Dr Rashid  Beebeejaun, Deputy Prime Minister and Deputy Leader of the Labour Party

Xavier Luc  Duval, Vice Prime Minister and Leader of the PMSD

Kailash Purryag, Speaker of the National Assembly

Pravind Jugnauth, MP and Leader of the MSM. Former Deputy Prime Minister

Cassam Uteem, Former President of the Republic

Alan  Ganoo, MP and Deputy Leader of the MMM

Abu Kasenally, MP and Minster of Land and Housing

Cader Sayed- Hossen, MP and Minister of Commerce and Industry

Bernard Sik Yuen, Chief Justice

Rajsoomer Lallah, Former Chief Justice

Sir Victor Glover, Former Chief Justice

Robert Ahnee, Former Judge. Member of Sachs Commission on Electoral Reform

Sir Hamid Moollan, QC, Lawyer

Eddy Balancy, Judge of the Supreme Court

Irfan  Rahman, Electoral Commissioner

Nando Bodha, MP and member of the MSM. Former Minister of Tourism

Shakeel Mohamed, MP and Minister of Labour and Industrial Relations

Leela Devi Dookhun-Luchoomun, MP and member of the MSM. Former Minister

Showkatally Soodhun, MP and member of the MSM. Former Minister

Steeve Obeegadoo, MP and Secretary General of MMM ( discussions over phone )

Cabinet of Ministers: meeting with Cabinet Ministers

Geoffrey Cox, MP of the Conservative Party, United Kingdom. Lawyer

Lindsey Collen, member of Lalit

Rajni Lallah, member of Lalit

Dr Jocelyn Chan Low, Associate Professor of History, University of Mauritius

Dr Mamade Khodabaccus, Mayor of Port Louis. Member of the PMSD

Nanda Kisnen, Lawyer. Member of the PMSD

Ashok Subron, Member of Reziztans and Alternative (discussion over the phone)

Dr Singfat Chu, Associate Professor of Business Analytis,National University of Singapore (discussion over the phone and by mail)

Jean Claude de L’Estrac, Journalist, writer and historian. Former Minister

Yousouf Mohamed, Lawyer and Former Minster

Lindsay Riviere, Chairman, Le Sentinelle

Raouf Bundhun, Former Minister and Former Vice-President of the Republic

Serge Clair, Leader, OPR ( discussion over phone)

Nicholas Von Mally, Leader MR and Minister of Rodrigues (discussion over phone)

Pierre Dinan, Economist and author of a report on Electoral system in Mauritius

Women’s wing of the Labour Party

Women’s wing of the MSM

Women’s wing of the PMSD

Loga Virahsawmy, Director Gender Links

Jane Valls, Women in Politics and Women in Networking

Dr Ameenah Sorefan, Vice President,Media Watch

Bruno Woomed, Women in Politics and Women in Networking

Celh Meeah, MP and leader of FSM

 

Many individuals have phoned me to express their views on electoral reforms.

Some people have also sent proposals (by post and by e-mail)  on electoral reforms and these have been considered.

Posted by on Jan 29 2012. Filed under Actualités, Featured. You can follow any responses to this entry through the RSS 2.0. You can leave a response or trackback to this entry

Leave a Reply

Search Archive

Search by Date
Search by Category
Search with Google

Photo Gallery

Copyright © 2011-2016 Minority Voice. All rights reserved.