Proposals to the Electoral Boundaries Commission on the delimitation of electoral boundaries   

An informed balance between voter equity and effective representation in plural Mauritius

1. Voter equalisation is important but effective representation is also key in diverse societies

1. The Commission has a tough needle to thread

The Electoral Boundaries Commission (EBC)  has an unenviable and ungrateful task in delimiting the 20 constituencies in the Island of Mauritius. The exercise is a hornet’s nest. It is plain that it will be criticised whatever it recommends , especially by some stakeholders who would like the Commission to accomplish miracles that go beyond its mandate. Some want it to equalise the number of inhabitants per constituency to putatively comply with Clause 39 (3) of the Constitution . Expectations are so unrealistic that the report could end up being either not tabled in the Assembly or rejected. However , inspite of this heightened risk, the Commission should carry out its mission without fear or favour. Above all, it should shun populism and embrace responsibility.

Equal voting power in terms of constituencies with nearly the same number of electors or population is a crucial element of parliamentary democracy. It is consistent with international conventions and it applies universally to all countries that hold free and fair elections. Hardly anybody disputes this . The challenge , constitutionally, politically and practically, is whether equal constituency is an ABSOLUTE rule. Put simply, should it be the only criteria to redraw constituencies or should countervailing factors be taken into account and derogation allowed for variances from the norm ? . This is the crux of the matter and it constitutes the overriding challenge confronting the EBC.

1. What the Constitution provides for ?

Clause 39 (3) of the Constitution tasks the EBC to make recommendations for

‘ any alterations  to the boundaries of the constituencies as appear to the Commission to be required so that the number of inhabitants of each constituency is as nearly equal as is reasonably practicable to the population quota’.

This is however subject to a derogation such

‘ that the number of inhabitants of a constituency may be greater or less than the population quota in order to take account of means of communication, geographical features, density of population and the boundaries of administrative areas”.

Geographical features, density of population and boundaries of administrative areas have been used to explain differences in the number of electors among constituencies .

1. Four key challenges in boundaries delineation

This document addresses four main issues in redrawing the boundaries of constituencies in the Island of Mauritius. My submission is that while voter equality is the most important delimitation requirement , there are key countervailing factors that must also be acknowledged in a plural society as Mauritius.    

First , I argue that the near equalisation of constituencies, while desirable in the absolute, must be contextualised and cannot be carried out in isolation of countervailing factors. The most important one is the historical trade off reached by those who negotiated BOTH the electoral system and the electoral boundaries in the 1960’s . The Commission has a responsibility to examine those factors too, as very well articulated by Supreme Court judgements in similar cases in many countries including Canada, Japan and South Korea. While section 39 (3) on near equal constituencies is key, the delimitation exercise cannot ignore other considerations including Clause 5 (1) of the First Schedule to the Constitution  which explicitly states the need ‘ to ensure a fair and adequate representation of each community’.

Second, I posit that in FPTP voting system , it is extremely rare for constituencies to be of equal size.  I shall demonstrate that it is certainly not the norm internationally even when there is a constitutional requirement for such equalisation. Many concrete examples exist, some in very robust democracies. And there are good reasons for justifying departure from such equality of electors . This is the case in India, Canada, Japan, the UK, Singapore and even the US and Australia to name but a few countries.

Third I discuss the case for a new set of considerations to delineate constituencies as the current ones are either confusing or not relevant anymore .

Fourth I propose two main factors to redraw electoral boundaries in our country . Like deeply democratic Canada, I recommend a balance between voter parity and effective representation in a plural society as Mauritius. Using this approach, I suggest a pathway to reduce the difference among constituency size while broadly recognising the countervailing factors of diversity of representation and inclusive Parliamentary presence in a multi-ethnic , multi-religious and multi-cultural society. This balance should hold until we have a full -fledged electoral reform that reshuffles the cards in a comprehensive and fair manner. Which only Parliament can accomplish. However the Commission can lead the way and make strong recommendations to that effect .

1. The landmark ruling of the Canadian Supreme Court on absolute and relative voter parity

With a FPTP electoral system, Canada faces the same problem of unequal constituencies, inspite of the paramount importance of the equality of electors between electoral districts . The discussion has always hinged on whether parity of voters is an absolute rule or other factors should be considered when reviewing electoral districts.

The members of the EBC being all of the legal profession will surely appreciate the merit and relevance of the judgement of the Canadian Supreme Court in 1991.

After a thorough examination of all factors, the Court cogently stated that

‘it is our conclusion that the purpose of the right to vote enshrined in the Constitution is not equality of voting power per se, but the right to effective representation. And the primary condition of effective representation is relative parity of voting power, modified where necessary by factors like geography, community history, community interests and minority representation’

The Court found that achieving voter parity at all costs may be undesirable as it could detract from effective representation by making it very difficult for minorities to be elected . As a result of that landmark judgement, the Boundaries Commissions in Canada are enjoined to pursue the objective of equal population size while also being required to consider non-population factors such as geography, community history, community of interests and minority representation . The outcome is that the Boundary Commission is permitted to accommodate these factors within a variance of 25 % above and below the population quota . It goes further by providing for no limit on the extent to which the population of a constituency can depart from the average in ‘extraordinary cases’, which must be justified.

1. The context and the text for boundary revision in Mauritius 

Let us be frank. While the principle of equal constituencies has been a core value of our electoral system, it has from the very beginning been tempered by other factors such as fair and adequate representation of all segments of the population.

The Trustram-Eve Commission of 1957 was both an electoral and a boundary assignment . It laid the foundation for our FPTP electoral system and the delimitation of boundaries together with the introduction of universal suffrage in our country. Its mandate was presciently very much an echo of the judgement of the Canada Supreme Court to balance parity of voters with effective representation. Its terms of reference were as follows:

i) to examine whether it is possible to divide the Colony of Mauritius into a series of single member constituencies up to a maximum of 40 of approximately equal voting strength but with a minimum electorate in any constituency of 5000 so that :

1. each main section of the population in Mauritius shall have adequate opportunity to secure representation in the Legislative Council corresponding to its own number in the community as a whole;

but providing that

1. each constituency shall have reasonable geographical boundaries;
2. the boundaries could be expected to endure for a reasonable number of years while yet remaining consistent with the primary objective at (a) above.

It was, right from the outset, a balancing act between the principle of equal constituency and effective representation . Sub section (c) that refers to the primary objective of fair representation at (a) is a crucial countervailing factor.  The same delicate balance between equal constituency size and effective representation underpinned the Banwell Commission of 1966 which recommended the current voting system and the electoral boundaries. And neither the boundaries nor the electoral system has changed since then.

Our history clearly demonstrates that ‘ equality of constituencies’ is NOT an absolute rule. It is a relative one and has always been balanced by the need for effective representation. Since the advent of universal suffrage in our country , there has been TWO major delimitations of constituency boundaries. Both were inextricably linked to the design of a new electoral system. The Trustram-Eve Commission of 1957 went further to qualify the principle of equal representation. If fair and diverse representation were not possible from the outcome of these FPTP elections, the Governor would appoint up to a maximum of 12 members to

‘ ensure representation of those who had no chance of obtaining representation through FPTP elections’.

The Commission had to draw the boundaries of the 40 single member constituencies so as to afford electable opportunities to the three main sections of the population (Hindu, General Population and Muslim) in numbers broadly corresponding to their proportion of the total electorate.  It was vintage ‘ gerrymandering’ of constituencies for the purpose of diverse Parliamentary representation.

The second major delineation of constituencies was carried out by the Banwell Commission in 1966. Again it was BOTH a boundary and an electoral commission. And one of its key terms of reference was

‘the system should give the main sections of the population an opportunity of securing fair representation of their interest ’ .

It was a compromise between those advocating a simple FPTP system with near equal constituencies and those proposing a full PR with the whole country as one single electoral district. There was a fear that many groups would fail to be represented in a FPTP voting system with well-defined electoral boundaries.

To allay these apprehensions, three ‘workarounds’  were introduced to the conventional FPTP system. First, the single member constituency was replaced with three member ridings to ensure a balanced slate of candidates. Second , the 40 constituencies were gerrymandered  into 20 new ones by an intelligent pairing to provide adequate opportunity for the main components of society to have Parliamentary representation . And a safety net was added through eight best loser seats to compensate for community underrepresentation arising from the elections of the 62 FPTP seats.

1. The countervailing factor of inclusive Parliamentary representation

It is also debatable whether the electoral boundaries can be significantly redrawn on a stand alone basis as it could nullify the safeguard provided explicitly in our Constitution at Clause 5 (1) of the First Schedule on the need

               ‘ to ensure a fair and adequate representation of each community’.

The Constitution goes on to explain how such fair and adequate representation should mathematically be achieved by comparing the share of each of the four communities in the population census to the share of the 62 FPTP elected representatives. And it provides for 8 Best Loser seats to correct any imbalance so as to reach a ‘ fair and adequate representation of each community’.

If the constituencies were substantially altered as to equate all of them , the 8 Best Loser seats would be largely insufficient to ensure a fair and adequate representation of each community. And that would pose significant representational problems. This is why the proposition by one political party to merge constituencies 2 and 3 into a single larger one  to reach the population quota and to divide constituency 14 into 2 separate ridings ( so as to retain 20 constituencies in the island of Mauritius) is preposterous while also being outside the mandate of the EBC which cannot alter the pairing of the existing 20 constituencies.

1. Equal constituencies are rarely the norm in FPTP voting system . Quite the opposite !

1. Mauritius not an exception to very unequal FPTP constituencies

Unequal constituencies remain a highly prominent feature of FPTP systems in many countries and significant levels of malapportionment persist over time. The focus on the principle of the equalisation of constituencies is not borne out by empirical evidence across countries that use FPTP voting system.  While many well intentioned people are clamouring for equal size constituencies , many countries with FPTP system do not actually implement such equality even when it is embedded in their constitution.  Often for geographical and demographic reasons to ensure diversity and plurality of representation in terms of region, states within a country , minorities and the balance of power between various groups.

Only in a full Proportional System is it possible to have each vote counted equally. The Netherlands, Israel, South Africa, Namibia and Peru have full PR electoral system with either no or very low threshold. In the recent election in South Africa, the whole country was one constituency and each vote , wherever it was cast,  carried exactly the same weight . This perfect situation cannot simply exist in FPTP electoral system.     

Both the Indian and the Canadian Constitutions enshrine the principle of representation based on the desirability of constituencies with comparable voter numbers. Nevertheless, the logical inference of voter equality is not the same as a clear endorsement of it and legislators do not pursue it with single-minded enthusiasm. Instead, they also consider accommodating countervailing factors like geography, community of interests and minority representation in delimiting boundaries. It explains why there are many derogations from equal constituencies in countries with FPTP voting system.    

1. Unequal boundaries in India to mitigate geographical dispute in Lok Sabha representation

The  best case study to illustrate the departure from equal size constituencies is probably the largest democracy in the world. There was a major disagreement between the Southern and the Northern states in terms of political representation in the Lok Sabha in the early 70’s. Some States did not want to be penalized as a result of their successful family planning programmes. The boundaries of constituencies in India were thus frozen in 1976 and have been based on the 1971 population census, even if Article 81 of the Constitution requires that each state receives seats in proportion to its population and allocates those seats to constituencies of roughly equal size and Article 82 calls for the reallocation of seats after every census based on updated population figures .

India went to the extent of prohibiting the delimitation of constituencies until 2001. For some 30 years since 1971. And in 2001, another amendment extended this freeze until the next decennial census after 2026. As the next population census after that year would be in 2031, it means that constituency boundaries would remain unchanged for almost 60 years in deeply democratic India , while the population would rise unevenly from 548 million to 1.4 billion . The great apprehension is that redrawing boundaries and distributing the existing 543 MPs in the Lok Sabha would mean that the South would lose many seats to the North.

The outcome of this persistent deferral is severely unequal representation. The population size of the largest constituency in the 1999 election had 3,101,838 voters and was more than 86 times larger than that of thesmallest constituency with 35,716 voters.

The Table below gives the massive disparity between the current five largest and the five smallest constituencies in India .

The largest constituency has over 61 times the number of electors as the smallest one. Even the ratio of the average of the five largest to the five lowest constituencies is very high at 15.4 .

In addition , 84 seats and 47 seats are reserved for the scheduled castes and the scheduled tribes respectively in the Lok Sabha  ( 24 % of the 543 seats) to provide for the descriptive representation of designated categories of disadvantaged groups. It is the demographic factor and promotes effective representation as these castes and tribes were believed not to be in a position to secure fair and adequate representation without such accommodation to the FPTP electoral system.

1. The justification to go beyond the 25 % margin in Nova Scotia, Canada

In its 2019 report , the Boundaries Commission of the province of Nova Scotia stated that it has balanced the two key concepts of voter parity and effective representation in its redrawing of the  constituencies. It explained that

‘Voter parity is achieved by creating electoral districts that contain roughly the same number of voters.  However voter parity may limit the voices of minority voters. This allows for the creation of electoral districts that contain fewer voters to allow for those minorities to be represented effectively in the legislative assembly’.

The Table below shows how the Boundaries Commission has balanced effective representation with voter parity.

1. The diversion of all constituencies from the population quota has come down with the new delimitation. This is in favour of voter equalisation;

1. However it had to protect 4 seats that return minority candidates to satisfy the requirement of effective representation. And the dispersion is higher than 25 % from the quota. The smallest constituency represents only 56 % of the quota with 7,458 electors ( instead of 75 % if the 25 % margin were respected) . As a result the ratio of the largest to the smallest district has increased from 2.14 to 2.27 . And it is constitutional.

1. USA : the most powerful democracy has extremely unequal constituencies

The US senate is arguably the most powerful legislative institution in the world . Yet it has one of most unequal distribution of electors as each state has two seats in the Senate irrespective of its population size. The State of Vermont with  625,000 residents have two senators, and so do New York’s 19 million. It means that a Vermonter has 30 times the voting power in the Senate of a New Yorker who lives next door over the state line. Worst, the 563,626 residents of Wyoming have the same representation as the 37,253,959 residents of California. It signifies that California has 66 times the number of electors as Wyoming. And the disparity is only increasing. The same malapportionment exists at the level of the House of Representatives , albeit on a smaller scale. The objective is to protect small states from being ‘swamped’ by the large ones.

1. The Singapore version of effective representation

To strengthen the sense of inclusivity in a multicultural country , Singapore has introduced some accommodating measures to its FPTP voting system .

1. it allows a tolerance of 30 % from either side of the quota in the size of its constituencies. The smallest single member constituency has 17,389 electors while the largest has 34,299 . A ratio of 1.97 .

1. it has variable member constituencies.  There are 13 one-member, 6 four- member, 8 five- member and 2 six-member constituencies for a total of 89 seats. Each multi member constituency must field at least one candidate from the official minority communities ( Indian, Malay, Eurasian and other minorities);

1. to ensure that minority candidates are not penalised by voters from the majority community, a single vote elects ALL Mps in each multi-member constituency. There is therefore no split vote among candidates from the same party as is the case in Mauritius;

1. over a period of time , the Presidency must rotate to a candidate from a minority community.

The point is not to judge these measures but to highlight that one cannot embrace a very simplistic solution of equal constituencies without an understanding on the history and context of the country and the significance of countervailing factors. Singapore strongly argues that this is absolutely essential as peace and prosperity also depends on preserving harmony and ensuring representation to all groups in a diverse country.

1. Rules and case law that permit departure from the norm

Some countries specifically provide for dispersion from the average quota either in their Constitution or at the legislative level. Malta and New Zealand provide for a variance of 5 % , Australia has a 10 % tolerance, Germany accepts a 15 % deviation , Lithuania and Hungary permit a 20 % divergence, Canada allows a 25 % departure while Singapore has a 30 % diversion from the average quotient. In some countries it is decided by the Court. In South Korea, the Constitutional Court held in 2014 that a 33.3 % variance was constitutional while in Japan the highest court approved a ratio of 2 to 1 between the largest and the smallest constituency. 

The question for Mauritius is what degree of tolerance should be allowed to strike a balance between the requirement for  ‘

‘the number of inhabitants of each constituency to be as nearly equal as is reasonably practicable to the population quota’.

and the

‘consideration of effective representation to ensure fairness , equity and inclusion of all segments of our multi-ethnic and multi-cultural society ’ .

1. Time to introduce ‘fit for purpose’ criteria for boundaries delimitation 

1. The current statutory principles for determining boundaries : lacking in clarity or not relevant at all

Clause 39 (3) of the Constitution requires the Electoral Boundaries Commission ( EBC) to make recommendations for

‘ any alterations  to the boundaries of the constituencies as appear to the Commission to be required so that the number of inhabitants of each constituency is as nearly equal as is reasonably practicable to the population quota’.

This is however subject to a derogation such

‘ that the number of inhabitants of a constituency may be greater or less than the population quota in order to take account of means of communication, geographical features, density of population and the boundaries of administrative areas”.

First there is the key determinant of the population quota. Then it speaks of the number of inhabitants of each constituency . And then there is the interpretation of what constitutes ‘ as nearly equal as is reasonably practicable ’. And finally there are permissible derogations from the population quota based on the four factors of the Clause .   

The Constitution mentions ‘number of inhabitants ‘ and ‘ the population quota’ when as a matter of fact, elections are about ‘registered electors’. To confuse matter further, the population census counts the number of  ‘residents’ rather than ‘inhabitants’.  It is high time to tidy up the Constitution to avoid such confusion in terminology as there is not always a close relationship among population, residents and registered electors. For instance some constituencies have more residents but fewer electors than others ( number 3 against 2 ; number 5 compared to 15 and number 20 versus 18 ) . Equally their variance from the population quota is very different to their dispersion from the electors quota ( constituencies 3, 5 and 16  amongst others). Another problem is the fact that while the average population has risen by 4.6 % between the last two censuses ( 2000 and 2011) , Constituency 5 grew by 14 % while number 14  went up by 13 % . These are regions with a higher level of foreign citizens. In the current context , the population quota may not be the best standard to measure voter parity, especially with many resident foreign citizens without voting rights in some constituencies. The number of registered electors looks more accurate and I recommend its use to delimit electoral boundaries .

1. How much flexibility is there in ‘ nearly equal’ and how do we measure inequality?

It is impossible in FPTP system to have equal number of electors in each riding. Nearly equal sends the message that there is no requirement for absolute parity and that some deviation from the quota would be tolerated . However there is no legislation or guideline as it exists in many countries to enable the EBC to ascertain those departures from the population quota . Would the deviation be          5 % , 10 % ,15 % or higher ? . Would they depend on countervailing factors ?

There are three indices that specialists adopt to measure voter disparity . Most people use the ratio between the largest and the smallest constituency to underscore  the inequality in the number of electors. While it is intuitive and easy to calculate , all experts dismiss it as  misleading as it computes only outliers even when the rest of the constituencies have near equal number of voters and fails to assess the whole picture. Electoral scholars prefer two other indicators. First the Loosemore–Hanby index which aggregates the value of the difference between the seat share ( in our case three seats or 5 % of the total of 60 seats) and the population in each district . If there is no difference between the two , it shows that the principle of “one person, one vote” applies in practice. The third method is similar to the Gini coefficient that measures inequality of income or wealth in society where 0 means perfect equality and 1 perfect inequality.

I suggest that we use deviation from the electors quota to measure inequality as it considers all constituencies in the country instead of looking at the smallest and the largest districts only. Similar to what the Canada, the UK, France and India use .

1. Are the four considerations for derogations still relevant ?

Using ‘ means of communication ‘ to account for dispersion from the population quota would be totally unrealistic as we can move from one place in the country to another with great ease. Geographical features such as mountain ranges and rivers are very unlikely to explain large variances in the size of constituencies as distances are very close. The argument of density of population can be used to justify fewer electors in sparsely populated and geographically isolated areas but certainly not in very densely populated regions like Constituencies 2,3, and 19. The only relevant factor that could possibly warrant inequality of voters among constituencies are the boundaries of administrative areas such as district councils and municipal lines . For instance , it would make sense for some electors who vote in number 18 to be transferred to 19 as in municipal elections they cast their vote in the town of Beau Bassin/Rose Hill. Similarly for the electors of Bassin, Palma and Residence Kennedy who vote in the Town of Quatre Bornes for municipal elections but in Constituency 14 instead of 18 in the General Elections. Likewise for the electors who reside at ‘ l’Autre Bout du Monde’ who should logically vote in constituency 19 instead of 8 .

1. Other constraints imposed on the EBC

.

The EBC must operate within its mandate which can be restrictive. Besides clause 39 (3), the EBC should work within the ambit of Section 7 (1) of the Mauritius Independence Order of 1968 that sets the number of constituencies in the Island of Mauritius at 20 following the pairing of the former 40 single member districts. The EBC cannot interfere with the pairing of these constituencies , nor can it amend the number of elected representatives from the current 60 and nor modify the number of Mps per constituency. This limits the ability of the EBC to make significant changes as it cannot mechanically operationalise the population quota

Only Parliament can make such profound alterations.

1. Retain three criteria for delimitation : electors parity, geography and demography

It is high time to revisit the basis for the delimitation of constituencies.  I submit that the dominant consideration for redistricting must be the near equality of the number of  electors in constituencies. The EBC should  be required to apply the principle of representation by electors to the greatest extent possible as it is fundamental to the exercise of democracy. However there should be no condition for absolute equality of voting power. The EBC should also recognise geographic and demographic considerations. The most relevant geographical factor should be the boundaries of administrative areas.  And the countervailing demographic factor is the necessity to ensure fair representation in a plural society. Some countries specify that community interests and minority representation must be taken into account so that legislatures effectively represent the society that elects them. As a result, deviations from voter parity can be justified, as long as they better serve the goal of effective representation. However, such dispersion must be as limited as possible and for special circumstances.

1. How much deviation is permissible for effective representation ?

By its socio-demographic make up, Mauritius is probably a more diverse and plural country than  Canada or Singapore . It seems to me that we should embrace a formula for redrawing electoral boundaries that combines the elements of equality of votes with the features of effective representation to echo the diversity and plurality of our society to ensure inclusive Parliamentary representation. And it should be done in a well-informed , transparent, and accountable manner .

I submit that something similar to what exists in British Columbia in Canada should inspire our EBC.

It runs as follows

‘ In determining the area to be included in and in fixing the boundaries of proposed electoral districts, the commission must be governed by the following principles:

(a) that the principle of representation by population be achieved, recognizing the imperatives imposed by geographical and demographic realities, the legacy of our history and the need to

balance the community interests of the people of British Columbia;

(b) to achieve that principle, the commission be permitted to deviate from a common statistical  electoral quota by no more than 25%, plus or minus;

(c) the commission be permitted to exceed the 25% deviation principle where it considers that very special circumstances exist.

The adjustment in our context could take the form of a different tolerance ratio to mirror our specific circumstances. Should the derogation from the quota to provide for non-population factors be 25 % as in Canada or 30 % as in Singapore or a higher one ?. As our case for effective representation is stronger , I would recommend a 35 % deviation . It means that a constituency may have either 135 % or 65 % of the quota. This should be the EXCEPTION and not the rule and should apply to very few constituencies as in Canada and Singapore .

1. An Electoral Boundaries Bill to propose new criteria

We should enact an Electoral Boundaries Bill to make changes to the legislation governing the boundary redistribution process along the following lines .

1. the dominant criteria should be voter equality to the greatest extent possible;
2. allowance should be made for the boundaries of municipalities and district councils;
3. the countervailing consideration of fairness of representation and inclusion must be recognised and departure from voter equality allowed ;
4. most constituencies must fall within a maximum margin of 15 % below/above the electors quota;
5. a deviation of 35 % to be permitted to ensure effective representation . It should apply to very few constituencies  ;
6. to replace the current population quota with an electors quota;   
7. the delimitation exercise must be carried out every ten years ;
8. to use the Loosemore-Hanby index that evaluates all 20 constituencies rather than the ratio of the largest to the smallest riding that considers only two outliers.

This should hold until we have a thorough review of our entire electoral system as it happened in 1959 and 1967.

1. A granular approach to balance voter parity with effective representation 

1. By how much to alter existing boundaries : to  balance the two criteria

I have suggested a new delimitation of boundaries using a combination of equality of votes and effective representation and a dispersion of 35 % in few cases to ensure inclusive Parliamentary presence. The delimitation exercise should pursue three objectives. First is to increase the number of electors in the two smallest constituencies ( 2 and 3 ). Second is to lower it in the three largest constituencies ( 5,14 and 15 ). And third is to decrease the variance of the other constituencies from the electors’ quota ( 6,9,10,13 and 19)

Table 1 depicts the number of electors per constituency ( as opposed to population) , the quota and the dispersion of each constituency from the quota in 2018.

                   Table 1 : The variance from voter parity in the 20 constituencies

The two lowest constituencies ( 3 and 2 ) are 52 % and 46 % lower than the average number of electors of 44,676 while the two largest ones ( 5 and 14) are 42 % and 39 % higher than the quotient. Besides these exceptions, there are significant deviations for constituencies such as number 15 with 25 % positive dispersion and number 13 with 22 % negative variance.  Constituencies 6,9,10,13 and 19 also have significant variances from the quota.

1. The intuitive and easy review of boundaries

Based on the constraints imposed on the EBC, the temptation would be great to embrace a simple redrawing of the existing constituencies motivated by a strict adherence of its mandate. It is anchored on the necessity to lower the ratio of the largest to the smallest constituency so as to make the boundaries less unacceptable. It would run as follows

1. to raise the number of electors in constituencies 2 and 3 to around 29000 each ;
2. to lower the number of electors in constituencies 5 and 14 to bring both of them to around 54000/56000;
3. the ratio of the largest to the smallest constituency would decline from 3 to 1 to 2 to 1 ;
4. Constituency 4 will receive the surplus of electors from 5 ;
5. Number 18 is likely to witness a major transfer from Constituency 14;
6. Constituency 1 will shift electors to Constituency 2 while number 4 will transfer electors to number 3;
7. There will be some small changes in few constituencies  and no change in many ridings .

The review would have two major drawbacks

1. While the ratio of the two largest ( 5 and 14 ) to the two smallest constituencies ( 2 and 3 ) would come down , there would still be large deviations well in excess of 15 %  for many other constituencies such as 6,9,10,13,15 and 19 ;

1. It would be very difficult to justify the still huge disparities  in constituency size on the basis of ‘ means of communication, geographical features, density of population and the boundaries of administrative areas’.

It would lack credibility .

1. A review based on new considerations and a better measure of inequality

The EBC could explain its constraints and proceed to recommended a different course of action . The end result would be to lower the deviation from the quotient while also bringing down the ratio of the largest to the smallest constituency . In some cases, the EBC  may have to choose  between the criteria of near equal constituencies and the boundaries of administrative areas. I submit that it should favour voter equality which is a much more important consideration.

The exercise could unfold as follows .

1. 9 constituencies are within a 10 % diversion from the quota of 44,676  electors. These are constituencies 1,4,7,8,11,16,17,18 and 20. These dispersions look reasonable and they should, as far as practicable, be subject to little changes even if some will be affected by the  reallocation of electors from other constituencies;

1. 4 constituencies – numbers 6,10,12 and 19 – are between 10 % to 20 % deviation from the average number of electors.. A review of the five constituencies of 6,7,8,9 and 10   (and may be 5 also) with variances of + 18.29 %, -3.15 %, -0.74 %, +24.46 % and + 15.82 % from the quota, could address the over- representation of numbers 6, 9 and 10 and the under-representation of 7 and 8 and lower the dispersion from the quota for all five constituencies to within 15 % of the quota. We could also address the under representation of Constituency 19  by a transfer of some electors from 18 ( those who vote in Rose-Hill for the Municipal elections)  and may be some from 20 ;

1. Changes to constituency 12 with a negative variance of 13.83 % looks tricky as it is surrounded by 11 and 13 which are both small constituencies compared to the mean . Unless we proceed by double transfers of electors , first from number 14 which is way above the average to 13 and then from 13 to 12 . Or from 10 to 11 and then from 11 to 12 ;

1. Two constituencies – number 9 and 13 – have variances of between 20 % and 25 % . One has an excess of  24.46 % while the other has a deficit of 22.22 % . Both dispersions could be lowered with proposals contained at (ii) and (iii) above and bring them within a margin of 15 % of the quotient ;

1. The two very small constituencies – number 2  and 3 with negative variances of 46.37 % and 51.81 % – should have an increase in the number of electors . Constituency 2 should receive electors currently voting in number 1 and some from number 3 while Constituency 3 will draw from constituency 4 and some from 2 also . However they would still remain the two smallest constituencies as it would not be possible to be within a margin of 15 % of the average quota of 44,676 electors . For the sake of effective representation, equity and participation, we should accept around 28000/29000 electors in both constituencies. It is the rationale for the 35 % margin from the mean in ‘special circumstances’ ;

1. The two largest constituencies will see a distraction of electors that will move to other ridings. There should be a relatively large transfer from number 5 to 4 and a similarly one from constituency 14 to either 13, 18 or 20 .  Constituency 5 would end up with around 55000/56000 electors which means a margin of 25 % above the quota. Constituency 14 is largely dependent on whether electors from its southern part is transferred to number 13 or electors from Palma, Bassin and Residence Kennedy are shifted to 18 ( to coincide with Municipal elections) or those from Bambous are moved to Constituency 20 and then some electors from  20 are transferred to number 19 .  Anyway, constituency 14 could fall within a 20 %  margin above the quotient with around 52000/53000 electors;

1. There may be a case to better balance the two adjacent constituencies of 15 and 16 with around 50000 electors each compared to the current 56000 and 44000 respectively ;

1. If electors from Palma, Bassin and Residence Kennedy  currently voting in Constituency 14 are transferred to number 18 , then the latter would have a significant rise in its number of electors ( to more than 50,000) even if some voters in 18 are shifted to Constituency 19  ;

1. As a result of the above additions and distractions, there would be consequential changes in  some remaining constituencies. Number 1 will be smaller after the removal of electors who will vote in number 2. Constituency 4 will see both additions ( from 5 ) and substraction ( to 3 ) . If number 20 receives electors from 14 ,then it will have to transfer some to constituency 19 . Otherwise Constituency 19 and 20 will witness small changes only ;

1. The ratio of the largest to the smallest constituency would come down from around 3 to 1 currently to about 2 to 1 . Many constituencies will be within a margin of 10 % of the quota while some will have a dispersion of 15 % . There will be very few outliers driven by  demographic considerations. The Loosemore-Hanby index which measures inequality among constituencies would decrease , indicating an improvement in voter parity.

1. It is undoubtedly a challenging and complex exercise . However it will provide a much better balance between voter equality and effective representation with the priority given to near equal constituencies in most cases and very few exceptions made to ensure fairness of representation;

1. This special accommodation could be revisited in the context of a genuine electoral reform . If a dose of PR were injected in the current FPTP formula, the impact of unequal constituencies  would decrease as ridings with more voters will have a greater weight in the allocation of the PR seats.   

1. Concluding Note

The argument that constituencies should have more or less the same population/electors number  has impeccable democratic credentials. However there are countervailing factors that must be acknowledged. Many countries, especially multi-cultural and multi-ethnic ones, with similar electoral systems to our’s , accept the consideration of effective representation and allow for deviations from the population or electors quota as they contribute to the better government of the people as a whole. Landmark court judgements have supported such derogations even when the Constitution enshrines the principle of equality of voters. Our own history shows that such countervailing factors have been recognised and there have been accommodation to the FPTP formula for the sake of equity, inclusiveness, and participation .

Ultimately, the EBC faces a difficult choice . Either it operates within the given constraints and proposes changes to lower the ratio between the two largest and the two smallest constituencies while keeping the deviation still very large in many ridings.  In which case , it could prove very difficult for the EBC to justify such huge remaining variances from the population quota on the basis of  ‘means of communication, geographical features, density of population and the boundaries of administrative areas’ only . Or it takes the bull by the horn and recommends a more informed redistricting based on the twin principles of voter equality and effective representation as embraced by some deeply democratic countries as Canada . The innovative proposals have the benefits of being open, transparent and objective . They could be the solution until we have an electoral reform that reshuffles the cards in a fair and inclusive manner.

Dr Rama Sithanen

                                Annex 1 : Unequal constituencies across the world