Mauritius General Election 2019:

A data-centric analysis

To start with, I would like to state that while I do follow our local politics, personally I do not identify with any one of the big parties in our local landscape. The three main parties on the local scene, mostly resort to populist politics (usually offering people something for ‘free’) which in turns means that they are rarely, if ever, driven by any ideologies. The results of such a situation means that there is, in my humble opinion, little to no distinguishing features between all of them.

This analysis is purely my own and is based on the data that was published by the electoral commission available online in an agnostic fashion. In the two weeks preceding the vote on November 7th, I was under the impression that the PTr (L’Alliance Nationale) would win the elections narrowly in front of the MMM followed by the MSM (L’Alliance Morisien). Why did I think so — by merely following our local newspapers. Being abroad, all I could do was to read the papers, and in hindsight it was a mistake. The main online papers are biased and pushing for their own agendas most of the time which, like in 2014, gave myself and quite a few people on social media a warped impression on how the results would unfold.

For the purpose of the current analysis, Rodrigues Island will not be considered and we will mainly focus on the voting patterns in the Island of Mauritius. This is because Rodrigues has a political history of its own that is quite different from the main island’s historical three party division.

The results nationally

Without further due, lets check the overall results nationally:

Overall in fact, the votes were quite split between the three main forces as seen in the pie chart above. This was expected as it was a ‘lutte à trois’ where we expected a good split of the votes. The winning alliance (L’Alliance Morisien) indeed won the bigger share of votes (38%) but, purely from the percentage of votes, a coalition of MMM and L’Alliance Nationale (33.3 + 20.9 = 54.2%) should have mathematically gather a majority to lead the National Assembly which is not the case (more on this in the next part).

Other notable fact is the 7.4% of votes that independents and smaller parties have been able to gather nationally. This is a very weak score — although to be fair, Rezistans ek Alternativ, who would have had a decent share of votes, did not compete in this election. A lot of Mauritians complain all year long that they do not trust the mainstream parties and that the latter are all the same. But for all this talk, this never translate to any actual endorsement of alternative choices in general elections, when it actually matters — like the saying goes, talk is cheap. For the younger aspiring candidates that are hoping to bring change, studying the current and past political landscape of Mauritius, they will only come to the conclusion of either not getting involved in politics to keep their hands clean or if ever they have to get into the arena, then the only choice is to join a mainstream party. Mauritians can only blame themselves to have perpetuated such a cycle.

We take a quick look at the demographic of voters for the current election. To do so we use the estimated census of 2018 from Statistic Mauritius. The current list of registered voters should have a very similar distribution for each age group shown above. The under 40-year old account for 40% of the population which is a huge portion that can on its own win the election! While it is expected that the under 40s are less attached to the mainstream parties, the results of the election tells a different story. Since 92% of the votes went to the three traditional parties, the younger generation must have backed them in quite similar way than the older generations did.

The unfair distribution of seats in parliament

In a first-past-the-post voting system, the national assembly looks unfortunately very different to how the public has voted. Below is a pie-chart comparison, between the actual vote percentage on the left and the division of allocated seats in the national assembly (in terms of percentages) on the right, that illustrates this issue.

It is quite telling how votes of the public does not translate properly to the allocation of seats in the national assembly. What originally was a slight majority of the winning party as compared to each of the other two parties, became an overwhelming majority. Best loser allocation does not correct this situation as well as can be seen in the figure below.

The situation becomes even worse for the MMM, which scored 20.9% of the votes nationally and after the best losers are allocated, ends up with a mere 11.9% of the seats in parliament. On the other side, the winning alliance score of 38% of votes, gets boosted by more than 1.5 times to land with 42 seats which amounts to 62.7%.

A caveat to this above point though, is that, all those taking part in the election knows how the system works with its strengths and weaknesses. Their campaigns should in effect be tailored to try to extract the maximum out of the system — hence in a sense it is fair game, as they are all playing by the same rules. It is a bit analogous to the electoral college in the U.S, where candidates of the two major blocks mostly campaign in swing states, leading to the presidential election, where slight difference in votes matters nationally as compared to campaigning in fully Democrat states like California, New York or fully Republican states like Texas where the battles are already won/lost for each party. But in that particular case though, the U.S citizens are electing a single person — which makes it hard to be representative, while in our case we are electing an entire assembly of 60 persons. So the system should aspire to be as representative as possible of the voice of the public.

Abstention

I have read and heard many people commenting that the L’Alliance Morisien won with around ~30% votes as can be seen from the pie chart below when abstention is taken into account.

Indeed, from this viewpoint, only 30% of Mauritians voted for who will lead the country but there is a big red flag here. The large majority of the 22.7% of people who abstained from voting, chose not to do so, and in so doing, voided their votes. They clearly could have swayed the election in any chosen direction i.e a very large majority of L’Alliance Morisien or MMM or L’Alliance Nationale, but did not bother to choose — so we cannot and should not speculate what would have happened. At the end of the day, they refused to make their voices heard on election day and we should not include them in any analysis of the results. There are still a few people, among those who abstained, that could not vote due to work, illness, some major event in their lives and but those cases should be a minority fraction rather than the bulk.

This abstention rate though, does not include those who, unfortunately, were removed from the electoral list by mistake. The recorded 6800 cases (or probably more) are a shame to the electoral commission, because of the doubt they cast in the minds of many people and the fact that for the past few elections, the organisation and set up were much more professional that the latest one.

Constituencies imbalance

A quick look at the number of registered voters per constituency will tell anyone how imbalanced the contours of each one is. This has been a known fact for a long time, yet no one has tried to address this issue in a long time.

In essence, if we look at the above figure, a person voting in #3 (Port-Louis Maritime …) has 3 times the power of a person staying in #5 (Pamplemousses and Triolet) as both constituency has the same allocation of three MPs (Member of Parliament) in the national assembly.

Similarly, the number people found in the three largest constituencies, namely: Pamplemousses and Triolet (#5), Savanne and Black River (#14), and La Caverne and Phoenix (#15) adds up to 185,871. This figure is very close to the 200,721 voters found in the six smallest constituencies namely: Port Louis Maritime and Port Louis East (#3), Port Louis South and Port Louis Central (#2), Riviere des Anguilles and Souillac (#13), Stanley and Rose Hill (#19), Mahebourg and Plaine Magnien (#12) and Grand River North West and Port Louis West (#1). The votes of people per MP elected from the six smallest regions are worth twice that of the voters from the three largest areas. If our island wants to be a true democracy, such large discrepancies which privilege (or undermine depending on whichever angle one sees it) people purely for geographical reasons should not be allowed to prevail.

The bar-chart above shows a comparison between the number of electors in each constituency in 1983 and in 2019. The black dashed lines illustrate the minimum and maximum number of electors for all twenty constituencies for the specific year. As one can clearly see from the plot, the separation between the black lines has widened over the years and our lawmakers have not done anything to remedy the situation. Since the delimitation of the different constituencies has not changed to account for the migration of inhabitants and the overall growth of the population over the years, the imbalance has only worsened with time. Lack of political will from our leaders means that the group leading the national assembly will be unwilling to change a system which just made them won an election — as has been the case over the past 20+ years.

Urban vs Rural voting patterns

In this section we analyse the voting patterns between Urban and Rural regions. To do so, constituency numbers 1, 2, 3, 15, 16, 17, 18, 19 and 20 are classified as Urban regions. The rest namely constituency numbers 4, 5, 6, 7, 8, 9, 10, 11, 12, 13 and 14 are assigned as Rural areas. In practice, this means that urban region accounts for nine constituencies, i.e 27 MPs as compared to 33 MPs coming from the eleven rural regions. Right away one can see why rural regions are the key to winning the elections nationally.

After this separation, we find that in terms of the voters turnouts (in the diagram below), people in rural areas are slightly more encouraged to vote in comparison to their urban neighbours.

Urban regions had very evenly split voting patterns (shown with the bar-chart below). A slight advantage was found for the L’Alliance Nationale which was quite surprising as people expected the MMM to be more dominant in the urban regions. This translated into 11 MPs for L’Alliance Nationale, 8 for MMM and 8 for L’Alliance Morisien which in terms of percentage is 40.7%, 29.6% and 29.6% respectively. The 8% of votes to independents gets somehow added to the 32% actual score of the winning team (winner takes it all). In general though, people going to the polls in these regions seemed to have split their votes according to their favourite candidates of the three major groups.

In contrast, the rural voting pattern went mostly for a single party and this is where the general election was won. A clear support for L’Alliance Morisien could be observed which scored almost 10% more votes than L’Alliance Nationale, while MMM had to settle for a meagre score of 15.5% of votes. What can also be noted is the lower support of independents and smaller parties in rural regions in contrast to urban regions — though the difference in itself is not too big to be fair. The clear victory of L’Alliance Morisien in the rural region meant that they managed to get 30 seats out of the 33 (91% of seats) available with the remainder 3 seats going to L’Alliance Nationale. The 44% of voting support for L’Alliance Morisien from the public in rural regions translated into almost 91% of the assembly seats after the selection of the top 3 candidates! In the case of rural regions, people seemed to have voted in ‘block’ to support their favourite parties which resulted in the battle being one-sided.

Conclusions

The main idea behind this blog post was to inform and give some insight to the general public in Mauritius. There are still more to be analysed from the available data on elections back home — which could be the subject of a future post.

To summarise, a few key things stand out for me:

• There is a big injustice with an electoral system that is purely first-past-the-post based. Elected MPs which are the voices of the public should be to a certain extent representatives to the vote counts across the country.
• Mauritians across the board, still has a low support for smaller parties and independent candidates. Votes mainly gets funnelled to the three historical parties even though the younger generation (less than 40 yrs) account for 40% of the voting population.
• The population imbalance between the different constituencies creates an undemocratic situation where not all votes are equal in terms of representation in the national assembly.
• Rural regions were once more the main deciding factor of which group would lead the national assembly. Votes in urban region was quite evenly split.

My name is Rajin Ramphul and I am an Astronomer turned into a data scientist/machine learning engineer. If you are interested with this article and want to contact me, feel free to do so via my gmail account (rajin250@gmail.com). All the calculations for this post were run using python programming on publicly available data from the electoral service commission of Mauritius and Statistics Mauritius.