‘Landslide’ election in Cape Verde and the effects of district-magnitude variance and separate elections

Cape Verde‘s presidential election was on 12 February. It was won by incumbent Pedro Verona Rodrigues Pires of the Partido Africano da IndependÃÂȘncia de Cabo Verde (PAICV), who defeated Carlos Alberto Wahnon de Carvalho Veiga of the Movimento para a Democracia (MPD). The margin was 50.97% – 49.03%, or 3,282 votes.

That was a landslide compared to five years ago when Pires defeated the same challenger in a runoff 50.006% – 49.994%, or by 17 (yes, seventeen) votes. In 2001 the candidates had been much, much more separated in the first round: 46.5% – 45.8%. (These two politicians have been dominant political figures since independence in 1975.)

Talk about a country with closely divided politics!

Its legislative elections produce rather less sharply divided results, however. Cape Verde is one of those separated-powers systems in which not only the powers, but also the election dates, of executive and legislative branches are separated, albeit not by much. Legislative elections occur a few weeks before the first round of the presidential election. And in both 2001 and 2006 the PAICV has peformed better (and the MPD worse) in legislative than in presidential elections.

The PAICV won 49.5% of the legislative vote in 2001 (but, as already noted, only 46.5% for president in the multicandidate field and barely over 50% in the two-candidate runoff), while the MPD won only 40.5% (or more than five points behind what its presidential candidate would do less than a month later). In 2006, the pattern was much the same, with the PAICV managing 52.3% of the legislative vote and the MPD almost ten points behind (42.7).

The electoral system is also somewhat disproportional, giving the PAICV about 57% of the seats in 2006 and 56% in 2001, while underrepresenting the MPD.

Cape Verde’s legislative electoral system is PR (d’Hondt method) in districts of varying size. The two major parties are differentially affected by varying district magnitudes, because one party (obviously the PAICV) is able to win: (a) most of the seats in outlying districts with low magnitude where it is the dominant party, and (b) some seats in the higher magnitude and thus more proportional districts in urban areas where it is weaker.

Districted PR systems with divergent magnitudes have this characteristic by their very nature: The smaller party in urban districts wins a share proportionate to its minority status there, and the larger party in the rural districts wins an above-proportional share there. If the “small urban” and “large rural” parties are the same party, the electoral system exhibits a partisan bias in favor of that party. This effect is known as the Monroe-Rose ‘variance effect’ after the two political scientists who discovered it (and they indeed used Cape Verde as one of several illustrative cases).

So, the electoral institutions of Cape Verde are biased towards the PAICV in two respects: the variable magnitude, and the separate timing of the presidential and legislative elections. The variable magnitude gives us the Monroe-Rose effect (which, the authors’ simulations show, may have cost the MPD a majority of seats in 1995, in an election in which it had to settle for a plurality). The separate election dates mean that even though the MPD has a presidential candidate who is obviously popular enough to win almost exactly half the votes nationwide, he has no opportunity to have coattails to the benefit of his party in the legislature.

The net effect is that the MPD is obtaining both fewer votes cast for its legislative slate than for its presidential candidate and a lower ratio of seats to votes cast than is the PAICV.

Cape Verde, like its “mother” country, Portugal, has a semi-presidential system. Had the MPD won either of these last two presidential elections that it lost so narrowly weeks after the election of a PAICV parliamentary majority, the country would have seen “cohabitation” between a president and prime minister from opposing parties.


Electoral data from Adam Carr.

The Monroe-Rose effect is from: Burt L. Monroe and Amanda G. Rose, “Electoral Systems and Unimagined Consequences: Partisan Effects of Districted Proportional Representation,” American Journal of Political Science 46, 1. (Jan., 2002), pp. 67-89. JSTOR access (restricted). This is one of my favorite articles of all time in the field of electoral-system research.

0 thoughts on “‘Landslide’ election in Cape Verde and the effects of district-magnitude variance and separate elections

  1. In districted PR systems with divergent magnitudes the smaller party in urban districts, you note, wins a share proportionate to its minority status there, and the larger party in the rural districts wins an above-proportional share there.

    No doubt it is preferable to have German-style linked districts.

    However, if one wishes the virtue of simplicity, I question the emphasis on uniform district magnitudes. Scotland managed that, but it suited their geography.

    To quote from Prof. Henry Milner’s brief to the Quebec National Assembly Special Committee [PDF] last November (for which he kindly gave me more credit than I deserved)

    “Portugal has 230 deputies in 22 districts, an average magnitude of only 10, although four of them are fairly large — 48, 38, 18 & 17 — giving the Greens 2 seats and a small left bloc, with only 2.8% of the vote, three seats. Switzerland has 194 deputies in 20 multi-member cantons, again an average of only 10, although four are large enough — 34, 26, 18 & 15 — to give small parties representation. Finland has 199 deputies in 14 districts, with four large enough — 33, 21, 18 & 17 — to give a degree of proportionality. Belgium has 150 deputies from 11 provinces, ranging in magnitude from 4 and 6 seats to 22 and 24 seats. Norway has 165 MPs in 20 districts — an average of only eight per district — four with magnitudes as small as 4 or 5, but three with magnitudes of 17, 16 and 15. Hungary has 334 local and regional seats in 20 districts, four with magnitudes as small as 8 or 9, but five with magnitudes of 60, 30, 24, 19 & 18.

    As Louis Massicotte says, his 16-region model, and our 14-region model, produces a two-speed proportional system: voters living in a metropolitan area get a more complete proportionality, whereas the residents of the nonmetropolitan areas must be satisfied with less proportionality. This is the same outcome as in the countries listed above, which are nonetheless models of democracy. If rural residents want to be represented in small regions to ensure that their particularity is respected, a quite legitimate wish, they must be prepared to accept less proportionality.”

  2. Cape verde really is growing ! [OK, so this is a weed. But at least it relates to the planting under which it sprouted! URL deleted from the entry–MSS]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s