Revised 10 Sept.

Sierra Leone is awaiting the results of its runoff election for president, as reported by All Africa. In the first round, on 11 August, incumbent Vice President Solomon Berewa, the candidate of the ruling Sierra Leone Peoples Party (SLPP), came in second with 38% votes cast. Ernest Bai Koroma of the All People’s Congress (APC) led the field with 44%.

Congressional elections were held at the same time as the first round, and the APC won a majority of seats (59 of 112).

The electoral system for congress is first-past-the-post. (IFES says that the congress is elected by “party-list proportional representation,” but that was true only for 2002. Thanks for Bancki, in the comments, for pointing this out.) Assuming that the APC did not perform markedly better in the congressional party vote than its 44% for president–which would be very unlikely¹ — Sierra Leone’s FPTP has produced a very disproportional result (advantage ratio of 1.2).²

The combination of two rounds to elect a president and a disproportional system to elect the congress greatly raises the risk of divided government. The APC might not be the consensus party choice of a majority, in which case it could lose the runoff, despite already having won a (manufactured) majority of seats in congress.

This isn’t a combination of electoral rules that any institutional engineer seeking to maximize a young democracy’s chances of survival would ever choose.

And there is yet another institutional oddity: The first round for president requires not simply a majority, but 55% of votes cast.³ I know of no other case requiring an extraordinary majority of popular votes to elect a president.⁴ While it is a laudable goal of institutional engineering to seek to elect a president with a broad consensus of support, the rule creates the potential anomaly in which a candidate who already had the support of more than half the voters was forced into a runoff, in which he or she would be pretty unlikely to lose the majority already obtained.

As best I can determine, Sierra Leone is a “pure” presidential system, with the president legally entitled to appoint a cabinet of his choosing, regardless of who has the majority in the congress.

The runoff was said to have gone smoothly with a close result expected. The winner, and thus whether Sierra Leone has majority or divided government, may not be known for 12 days.

1. The National Electoral Commission posts results, but does not include national parliamentary data. I am not even sure if there is a separate vote from that for president. [↩]
2. For 2002, when list PR was used, Adam Carr’s 2002 results show that all districts had a magnitude of 8 seats. An even-number magnitude, unless very high, tends not to give as big a boost to the largest party in any given district as it gives to trailing parties that have enough to win any representation. This may have been intentional, given that the 2002 election was held in the wake of a settlement to a civil war. A quick glance at the 2002 results, in which the SLPP was much more dominant, indeed shows the second party, usually the APC, over-represented in numerous districts. However, the results also show often rather large percentages of the vote for parties that won no seats. On the other hand, in 7 of the 13 districts in 2002 the SLPP won all 8 seats. [↩]
3. IFES does not mention this (strangely), but various news reports, including that from All Africa, linked above, do. [↩]
4. There are of, course, various “distribution” requirements elsewhere, some of which require a given minimum share in some stipulated number of territorial sub-units. These have a similar motivation, and may be more workable, inasmuch as the distribution can be combined with a plurality requirement. If it is combined with a majority requirement, it has the same potential flaw I note here for Sierra Leone’ rules. [↩]