Botswana election, 2014

I’ve long been skeptical of Botswana’s classification as democratic/free by standard datasets we political scientists rely on. Based on a pre-election report by Amy Poteete for The Monkey Cage, skepticism seems justified. Mysterious accidents and other odd events involving political rivals, increasing partisan use of state assets…

The Botswana Democratic Party (BDP), the only ruling party Botswana has ever known, is feeling the heat, and is experiencing internal tension, including cases of losers in candidate selection running against the party’s official choice.

In the 2009 election, the BDP won over three fourths of the seats* on just 53% of votes (FPTP). That’s some serious disproportionality, but a result like that reveals electoral precariousness.

Using the dataset of the Constituency Level Electoral Archive, I took a look at just how many seats the BDP won only narrowly in 2009. In that election, its median margin over the runner-up was .188 of the district’s valid vote. It would need to lose at least 15 seats to fail to retain its majority. The 15th most marginal seat was won by .116. (Five were under .05.) There were 17 seats that it won with vote shares below that of its nationwide share of .533, including seven won with under half the vote. Thus the party looks somewhat vulnerable if there is a modest swing against it or if the defecting candidates are in close districts and take some chunk of the BDP vote with them.

Nonetheless, the BDP in 2009 did not face a single opponent in many districts; of the 17 districts where its vote was below the nationwide share, it won with a median margin of .081. Thus unless its opposition is more coordinated in 2014, the BDP will probably hang on even in the face of an adverse swing.

The election is on 24 October.

_______
* The data I am working with show 44 seats, which would be 77.2%, while Poteete says 79%.

New Zealand split-vote results released

The New Zealand Electoral Commission has released the split-voting statistics from the 2014 general election.

This is a great service provided by the Electoral Commission, showing in each electorate (district) what percentage of voters for each party list cast their vote for that party’s candidate or any other candidate in the electorate. To make it even better for those who like analyzing voting statistics, they offer CSV files.

The NZ Herald offers a summary of key electorates.

Maharashtra 2014: BJP taking post-poll support from ex-ally of Congress

The Bharatiya Janata Party (BJP) continues to show signs of seeking to break out of the post-1998 pattern of two large pre-electoral coalitions that have taken turns governing India. While the BJP-led National Democratic Alliance (NDA) remains the governing formula at the federal level despite BJP having a majority of seats, the BJP played serious hardball in the recent campaign for the Maharastra state assembly. In elections on 15 October, the BJP won 119 of the 288 seats (41.3%). It appears that it will take “outside support”–i.e. post-electoral cooperation but no governing coalition–from the Nationalist Congress Party (NCP). While the Times of India on 19 October referred to this offer of support as “unexpected“, it was pretty clearly foreshadowed by the frenetic reshaping and ultimately breaking of alliances in the week leading up to the deadline to declare candidacies.

The significance of these developments is that the BJP has an alliance at the federal level with a Maharastra-specific party, the Shiv Sena. In the Lok Sabha election earlier this year, these parties continued their pre-election alliance, in which the parties agree not to compete against each other in districts and to support one another’ candidates. It appeared as if the alliance was critical to the strong BJP performance in the state in those elections. Further, Shiv Sena sits in the federal cabinet of BJP PM Narendra Modi. Meanwhile, the NCP is (or was) an alliance partner of the Indian National Congress Party (INC), in both the 2004-2014 federal government and in the state of Maharashtra until the run-up to these elections. The INC and NCP ruled the state in alliance for 15 years.

In late September, there was a flurry of media reports of a “seat-sharing row” between the BJP and the Shiv Sena, with the former demanding the right to contest districts currently held by the latter. The BJP was explicit in saying that there had been a “Modi wave” and that it was thereby entitled to a larger share of the districts. Meanwhile, the NCP played hardball with its ally, Congress (INC), demanding additional seats and an alternation in the Chief Minister’s post.

On 25 September, days before the candidate-filing deadline, the BJP announced it was dumping its ally, Shiv Sena. Barely an hour later, the NCP broke its alliance with the INC.

Given the concurrence of the demands from the national partner (BJP) against its ally and against the BJP’s national rival by a state partner (NCP), and the quick succession of the two announcements, it is hard to believe it was not being coordinated. It was seemingly foreshadowing the formation of a BJP-NCP post-poll alliance if the BJP won the most seats, but not a majority. And, of course, that is precisely what happened.

It was all quite dramatic, and it appears to be part of a BJP strategy of supplanting its erstwhile allies in favor of single-party minority government (when a majority is not (yet) in reach). It is especially telling that it would prefer to take outside support from an erstwhile Congress ally instead of continue a relationship with its own former pre-poll partners. (The Shiv Sena, contesting alone, won 61 seats, according to preliminary results, while the NCP won 42 (and Congress 44); BJP is 27 seats short of a majority.)

In my first post-election entry on the Indian federal result, I said that I doubted the BJP majority meant a re-writing of the fundamental rules of Indian politics. Yet the pre-poll and post-poll politics in Maharashtra suggests the BJP is attempting just such a re-write. Several key state elections are coming up in the next year, and the NDA partners have been put on notice.

The KC sweep and run differentials

The Kansas City Royals have completed a sweep of the Baltimore Orioles in the American League Championship Series. This follows the sweep of the Los Angeles Angels in the Division Series and the sensational one-game Wild Card playoff against the Oakland Athletics. Both the Angels and the Orioles had far superior regular-season records and significantly better run differentials during the season. The A’s had the league’s best run differential, despite their late collapse from a runaway division title to a ten-game deficit at the hands of division winner LAA. It really is hard to fathom how they are doing it. Royal blue smoke and mirrors? Best explanation I can offer.

In the just-concluded ALCS, the Royals continue to maximize wins out of few runs. In fact, their run differential is the smallest ever in a LCS sweep, by a good margin. In the series, they scored 18 runs, and the Orioles scored 12. They +6 differential beats the previous LCS record by three runs (1988 A’s over Red Sox, +9).

The +6 differential is reminiscent of the White Sox sweep of the 2005 World Series; as I noted at the time, only one prior World Series sweep had ever featured such a low differential (1950, Yankees over Phillies, also +6). Since 2005, there have been two more sweeps, but with typically large differentials (+19 in 2007 and +10 in 2012).

Obviously, the minimum possible differential in a four-game sweep is +4. So if you win with a +6, like the Royals just did, you are being exceptionally efficient in your run distribution.

Across all sweeps of best-of-seven World Series (19 of them), the average differential is +12.5. In the two leagues’ Championship Series, which have had this format since 1985 (when, incidentally, the Royals overcame a 3-1 deficit to win in seven), there have been six sweeps before this one. The average differential in those six was also +12.5. The Royals sweep was thus historic, setting a record for LCS and tying a record for all best-of-seven series. Too bad they scored that superfluous run in the top of the ninth in Game 2. They could have beaten the all-time best-of-seven sweep “efficiency” record.

(All data calculations by me, from Baseball Reference)

New Brunswick election 2014

The Canadian province of New Brunswick held a general election on 22 September. Notwithstanding some problems with the vote-tabulation system, and several lead changes during the night, the Liberals won by a good margin: 42.7% of the votes against 34.7% for the Progressive Conservatives (PC). The seats split 27-21, giving the Liberals 55.1% of the seats and an advantage ratio (%seats/%votes) of 1.29. This is not particularly remarkable by the standards of plurality (FPTP) systems, and is nothing compared to some significant anomalies the province has experienced in the past (see previous posts).

The Green Party won a seat. It had 6.6% of the vote province-wide. Its one victory was in the riding (district) of Fredericton South, where it won 31% of the vote, against 26% for the PC, 22% for the Liberal, and 20% for the NDP. Not even a third of the vote–sometimes plurality is good for small parties. But not as good, obviously, as the proportional system (of the MMP type) that was formally proposed but for which the planned referendum was cancelled–by a party that had just won on a plurality reversal. (I said NB elections had been anomalous!)

A “populist” party known as the Peoples Alliance elected no one on its 2.14% of the vote, but it did miss in one riding by a mere 27 votes. The NDP won no seats despite 13% of the vote. I am not sure how closely it missed in key ridings. Of course, it is hardly unheard of for a third party to win no seats under plurality rules despite such a substantial vote total. Nor is it unusual for a fourth party to win a seat despite having half the votes such a third party. It is a disproportional system, especially given the small assembly size, and regional distributions of support are critical. It is hard to argue against the proposition that New Brunswick should dust off that old proposal for a new electoral system.

New Zealand 2014 election result

Preliminary results from the 20 September general election show the National Party has won 61 seats out of 121. Thus, by one seat, it has a majority in its own. Despite having a majority, it is likely to continue to govern with confidence-and-supply agreements with the same partners it has had for its previous two terms: ACT, United Future, and Maori Party. National will want to retain good working relationships with other parties, given that a majority is not likely to be a common occurrence under Mixed-Member Proportional (MMP); indeed, it is the first majority since the system was put in place in 1996. Moreover, with just one seat over the 50% mark, trying to govern alone could be precarious.*

    [

UPDATE:

    The

final results

    show the National Party has just 60 out of 121 seats, or 49.59% of seats on 47.04% of the party-list vote. The Green Party picked up a seat in exchange. No electorates changed hands, although wherever I mention specific vote totals or differentials, they could be slightly different in the final count. I will leave the rest of this post unchanged.]

It must be noted that this majority is manufactured by the electoral system. That might seem like something that “should not” happen under MMP with nationwide proportionality. But two points. First, National is very close to 50%, currently on 48.06%. Partly the reason “fully” PR systems rarely manufacture majorities is that such high vote shares are fairly unusual. Second, New Zealand’s proportionality is limited by the 5% threshold, and with one party, the Conservatives, having obtained 4.12%, there are some wasted votes. Just excluding this party’s votes, National has 49.93% of the remainder. At this moment, I suspect Prime Minister John Key and his party are very pleased with themselves for not having adopted the Electoral Commission’s recommendation to reduce the threshold to 4%. With a lowered threshold, the Conservatives likely would have won five seats (perhaps more, as they might have picked up more support had it been apparent that a vote for them was not wasted), and National would have had 2-3 fewer.

One of the other notable features of the outcome is that the Maori Party won a list seat for the first time. It was a bad result for them overall, as they won only one electorate (district) seat. In the past, the party had won 4 (2005), 5 (2008), and 3 (2011) seats, all of them electorates. This time, they easily retained the one but were not close in any other. Their 1.29% of the list vote was just enough to qualify for a second seat under the alternative threshold (what New Zealanders call coat-tailing, although I would prefer a different term).

It was a quite bad result for Labour (24.7%, 32 seats) and a disappointing one for the Greens (10%, 13 seats, a loss of one from their current high at the 2011 election). It was a very good result for New Zealand First, with 8.9% and 11 MPs.

The Internet MANA alliance failed to win a seat, probably because the anticipated backlash did indeed occur. Voters in the one supposedly safe MANA electorate heavily voted strategically to keep the MANA leader out, thereby also obviating any chance that Kim Dotcom’s lavish spending would put some Internet Party MPs in office due to the alliance. (And to think, some folks still insist that MMP is too complex for voters to figure out; this case seems to suggest such a view is quite wrong!)

A glance at MANA leader Hone Harawira’s electorate of Te Tai Tokerau, one of the Maori special seats, makes the strategic voting quite apparent. The winner was Labour candidate Kelvin Davis, who received almost two thousand more votes than the Labour list received from the electorate’s voters. Normally, I might just attribute that to the Green list voters, who numbered 1,821. However, many Green voters might actually have wanted Harawira to make it in, because Internet MANA was another potential block of left-leaning votes in parliament. New Zealand First (2,805 list votes) and National (1,659)–neither of which contests Maori electorates–certainly will want to claim credit for defeating Harawira (and Dotcom). We will have a better idea when the Electoral Commission releases its split-voting analysis. (Also of note: The Green Party contested four of the seven Maori electorates; this was one of those in which they did not enter a candidate.)

In a comment at an earlier post, Manuel offers an interesting further observation on Internet MANA:

Incidentally, of 30,363 votes polled by Internet and Mana electorate candidates – as it has been pointed out here before, the two parties ran separately in the electorates (although never against each other) – 26,521 were for eighteen Mana candidates, and the remaining 3,842 votes for the fifteen Internet Party candidates, including 1,057 for party leader Laila Harré in Hellensville (where her poor fourth-place finish with 3.6% of the vote was by far the party’s best showing).

Harawira seems to have made a pretty serious miscalculation in forging his alliance with Dotcom.

I will now offer a few semi-random observations on specific electorates and smaller parties. Continue reading