Supporters staying home cost BN the election
We previously discussed how Malaysia’s voter turnout rate (defined as the percentage of eligible voters who actually voted) compares favorably with the rest of the world. Voter turnout matters, not only because participation in the democratic process attests to its legitimacy, but also because it can shape election outcomes. In this post, I analyze how lower voter turnout in some constituencies affected the results.
Who wants higher turnouts?
In a democracy, high voter turnout rates are considered an ideal to strive for. “Get out the vote” campaigns are a common part of elections – each candidate/party tries to increase the election day turnout amongst supporters who are eligible to vote. On the surface, this is good for the democratic process. However, candidates/parties truly care about increasing turnout only amongst their supporters because it can win them the election.
When a party thinks higher turnouts will favor their opponents, they won’t bother getting out the vote. Instead, they might engage in voter suppression, which involves discouraging or making it harder for some groups to vote (usually supporters of the other party). According to some, that’s what happened in the lead up to GE14. In particular, the election was held on a Wednesday, creating a barrier to voting that means only sufficiently motivated voters are going to turn up.
The data is pretty clear about whether Barisan Nasional (BN) or Pakatan Harapan (PH) benefited from lower vs. higher turnouts. The following figure shows each party’s margins and outcomes for all parliamentary-level contests depending on the voter turnout for that constituency.
As turnout increased, performance for BN got dramatically worse, and led to them losing more seats (points below 0%). On the other hand, PH won more votes and more seats as turnout increased. This suggests PH benefited from more motivated supporters. PAS performance was unrelated to turnout, because most of their contests were three-cornered fights (with more complex relationships between turnout and outcomes).
GE14 vs. GE13
Nationwide voter turnout in GE14 (82%) dipped slightly from GE13 (85%). This may have been due to the midweek election day, which would make it harder for some to vote in their home constituencies, so this drop probably does not indicate a drop in motivations to vote amongst eligible voters.
How did this drop in turnout affect results? The following graphs show the change in BN and PH vote margins based on changes in voter turnout. I have divided the data into:
- Safe BN seats, won by BN in GE13 with a margin > 15%
- Safe PH seats, won by Pakatan in GE13 with a margin > 15%
- Marginal seats, won by either coalition with a margin < 15%
Let’s take a look at PH’s performance. First, notice that most constituencies have changes in turnouts that are negative (a drop in GE14 relative to GE13), so we are looking at which constituencies had larger decreases vs. minimal decreases. As turnout decreased by more, PH’s average margin decreased. However, their performance in most seats showed increases in their margins (most points are above 0%), which allowed them to gain seats formerly held by others.
In marginal and safe seats, based on the trend lines above, PH support was maintained or increased as long as turnout did not drop by more than 3% from GE13. However, in almost all BN strongholds, PH support was increased regardless of turnout – this suggests something unique is going on amongst voters in BN safe seats.
Let’s switch over to BN’s performance. As voter turnout dropped more relative to GE13, BN tended to perform better. However, even with much lower turnouts, they still experienced drops in their margins from GE13 in almost all constituencies (most points in all types of seats fall below 0%). Consequently, BN lost a ton seats, despite the drop in turnout that should have favored them.
Who stayed home?
Let’s assume the midweek election day made it inconvenient enough that only motivated voters would show up. The relationship between turnout and increased support for PH suggests that PH supporters were more motivated to turn up and vote. In contrast, the midweek election may have proved an obstacle large enough to deter more BN supporters, who had lower motivations to show up and vote.
The following figure demonstrates this. It shows how changes in turnout from GE13 to GE14 are distributed for all seats of different types. (If you’re interested, it’s a plot of the probability density functions for each seat type).
This figure shows that while the majority of seats had decreases in turnout, there are subtle but significant differences between the types of seats. On average, PH safe seats had decreases in turnout of 1.8% (the red distribution), marginal seats had decreases of 2.6%, and BN safe seats had the largest decreases of 3.8% (the blue distribution is shifted furthest to the left). If you’re familiar with statistics, the total variance explained was 14.5%, which implies these differences in turnout changes by seat type are huge (by social science standards).
In summary, PH supporters maintained their turnout levels the best, which implies they were the most motivated to vote. This helped PH maintain most of their safe seats and win some marginal seats from BN.
In contrast, voters in traditional BN strongholds had decreased motivation to turn up and vote. BN took voters in these places for granted, assuming their support would stay at the same levels as in past elections. The overall decrease in support in BN strongholds suggests that many of them simply stayed home – this led to BN losing 21 of their safe seats.
Voter suppression: bad strategy
The election outcome was shaped way more by the turnout of BN than PH supporters. The expected BN supporters who stayed home this time around led to BN losing seats, while PH supporters only had to maintain their turnout levels to a reasonable level. In Malaysia’s elections, choosing not to vote is an option, but it it rarely inconsequential.
If the Election Commission employed voter suppression as a tactical move, it seems to have successfully depressed voter turnout slightly in GE14 (compared to GE13). This would have helped BN, except that voters in traditional BN strongholds were disproportionately likely to stay home. Failing to anticipate the low motivation of traditional BN supporters, the Election Commission appears to have shot BN in the foot.