The Kingdom of Lesotho is an enclave, entirely surrounded by the Republic of South Africa. This geography has fundamentally influenced Lesotho and its foreign and domestic politics. As with most African countries, its post-independence history is filled with military governments, elections, coups d’état, and juntas. Throughout it all, Lesotho kept friendly relations with Britain as insurance against South Africa and apartheid.
The 2017 general election came about because Prime Minister Mosisili did not survive a no-confidence vote. Did Lesotho hold a free and fair election?
During the colonial era, Lesotho was known as Basutoland, named after the primary ethnicity in the country. It is interesting, however, that the Basotho are scattered across southern Africa, with twice as many outside Lesotho as inside its borders. This is a results of two history and economics. Historically, the Basotho were originally a confederation of clans spread throughout southern Africa. When the British finally set the borders of Basutoland, it excluded many Basotho. Economically, many Basotho found employment in the South African mines, even during the apartheid era when South Africa tried to force them into black homelands.
The Kingdom of Lesotho is a constitutional monarchy in which the monarch, currently King Letsie III, has no official political power. The legislative branch is broken into the Senate and the National Assembly. The Senate consists of 33 members. Twenty-two of these are traditional chiefs, and 11 are directly appointed by the king.
The National Assembly has 120 members elected through a mixed electoral system. Two-thirds of the members are elected first-past-the-post in each of the 80 constituencies. The remaining 40 are elected (on the same ballot) using proportional representation (PR). The effect of those 40 is to ensure that the composition of the National Assembly more closely matches the composition of the official ballot counts.
To illustrate this, the following table gives the composition of the official ballots and the final composition of the National Assembly. Note that the PR ballots improve representation.
|Party||Vote Percent||Constituency [%]||Seats [%]|
|All Basotho Convention (ABC)||37.75||50.00||38.22|
|Lesotho Congress for Democracy (LCD)||9.91||2.50||6.67|
|Basotho National Party||5.53||1.25||5.83|
|Popular Front for Democracy||1.73||0.00||1.67|
|Reformed Congress of Lesotho||1.18||0.00||1.67|
|National Independent Party (IND-NP)||0.95||0.00||0.83|
|Marematlou Freedom Party||0.60||0.00||0.83|
|Basutoland Congress Party||0.48||0.00||0.83|
|Lesotho People’s Congress||0.34||0.00||0.83|
In the table above, note that the difference between the proportion of seats awarded is closer when the proportional-representation seats are included. This is the rationale for Lesotho’s hybrid electoral system.
The 2017 general election came about because Prime Minister Pakalitha Mosisili lost a no-confidence vote in early March. This signaled a dissatisfaction with the Democratic Congress in the National Assembly. That dissatisfaction with the DC extended to the people. As Africa news reported on March3, 2017:
Lesotho’s Prime Minister on Wednesday, deepening political uncertainty in the southern Africa kingdom.
Prime minister Pakalitha Mosisili of the Demoractic Congress (DC) who lost a no-confidence vote in parliament early this month has headed a coalition government since the ouster of former prime minister Thomas Thabane’s All Basotho Congress (ABC) two years ago.
The law makers voted in favour of replacing Mosisili with Monyane Moleleki whose Alliance of Democrats party split from the Democratic Congress last year.
From the above table, we can see that the All Basotho Convention (ABC) received a plurality of the votes cast (albeit not enough for a majority). While the did not have the consistent support across Lesotho, they did quite well in the more-populated areas. This, we can see from the map below: the higher the level of yellow, the higher the support for the ABC.
Here is a video from Al Jazeera regarding the expectations and results of the election.
The party of the incumbent, the Democratic Congress, also failed to show consistent support across the entire country. They did not even field a candidate in 26 constituencies. This lack of support severely hurt them in the National Assembly.
While the support levels are interesting for the political analysts trying to help their party take control of the legislature, it is the combination of those figures and the invalidation rate that interests the electoral forensic scientist. First, note the map of the invalidation rate at the district level.
At the district level, there does not appear to be a pattern to the invalidations. In fact, comparing this map to the support maps for the two main parties does not provide any apparent relationship.
However, that relationship becomes apparent when examining the next lower level: the constituency level. While we have no maps at that level (if you have access to those GADM files, please let us know), the invalidation plot at that level is very clear in the relationship.
The curve is the regression curve resulting from the Binomial generalized linear model using a logit link. As the other typical link functions produced similar substantive results, and as the logit link is the canonical link, we only report the logit results. [As a side note, we estimated the model using maximum quasi-likelihood to account for the natural overdispersion in election data.]
The curve’s slope is negative. This indicates that higher levels of support for the ABC party correspond to lower levels of invalidation. That is, higher levels of support for the ABC party correspond to higher levels of having the vote count. This is differential invalidation in favor of the All Basotho Convention. [The p-value, the support for the hypothesis that there is no differential invalidation, is 7 × 10-7.]
What does this result mean? It means that there is significant evidence that the rules for invalidating votes in favor of the Democratic Congress were very different from those for the All Basotho Convention — with the All Basotho Convention benefitting for this rule change.
For those who have been following the Center, you may realize that this conclusion is frequently used as evidence that the election was not democratic, was not “free and fair.” From reading through the post-election results, including the assassination of the premier’s wife, there seems to be no reason for the differential invalidation. Perhaps the incoming prime minister can explain this significant — and damning — result.
|Official electoral body||Independent Election Commission (IEC)|
|Election||June 3, 2017|