analysis Past performance is no guarantee of future performance. That is the advice I’ve been given in regard to mutual funds.
But can the same be said of elections in Malaysia? Some would say yes. The BN’s amazing electoral performance of winning 91 percent of parliamentary seats in 2004 could not be repeated in 2008.
Can the results of the Kuala Terengganu parliamentary election in March 2008 tell us anything useful about what will happen on Saturday, when the same voters (more or less) cast their ballot in the by-election?
I would argue in the affirmative, but with the caveat that our information needs to be ‘updated’ to reflect present day political realities.
In this article, I will systematically look at the following aspects of voting in KT in the March 2008 election - split voting; Malay turnout and age-related voting; Chinese turnout and voting; independent candidate voting; and postal votes - and ‘update’ these aspects to account for changes in political circumstances.
At the end, I will attempt to summarise the total effect of these changes and make a prediction on the electoral outcome on Jan 17.
Split voting
Table 1 (right) shows the magnitude of split voting in KT from 1999-2008.
A positive number means that more voters voted for the BN at the parliamentary level compared to the state level.
There was minimal split voting in 1999. It increased to 1,308 in favour of the BN at the parliamentary level in 2004 and to 2,871 in favor of the BN at the parliamentary level in 2008.
Split voting favoured the BN at the parliamentary level in all three 3 PAS-controlled state seats (Table 2, left).
Split voting actually favoured PAS in Bandar at the parliamentary level, the seat where most of the Chinese voters are concentrated.
Part of the reason for split voting in favour of the BN at the parliamentary level was because of the candidacy of Mat Sabu, who was not seen as a local candidate.
Given that the PAS candidate for the by-election, Abdul Wahid Endut, has a very strong local presence (he is a five-term state assembly person and was a member of the exco from 1999-2004 when PAS controlled the state assembly), the level of split voting should be decreased significantly.
If split voting in KT can be decreased to zero, the 628-vote majority for the BN would change to a 2,243 majority for PAS in the by-election. A swing of 2,000 votes would not be unreasonable just on the basis of reversing the split voting which occurred in 2008.
Malay turnout and age-related voting
Voter turnout in KT in March 2008 was approximately 82 percent. One of the unfolding narratives in the by-election is that a lower turnout rate will hurt the BN more than PAS. This is because the BN was more dependent on the higher turnout rate in 2008 to win this seat by the small majority.
The rationale is that BN supporters are more motivated by cash incentives to travel back to vote rather than PAS supporters who are more motivated by ideology/religion.
One simple way to show the effect of turnout on the support for the BN is to investigate the link between changes in BN support and changes in turnout from the 2004 to the 2008 elections. The relationship is shown in Graph 1 (below) using polling station from the state seat level.
The graph shows that as turnout decreases, the level of support for the BN decreases as well. There is a discernible linear relationship.
A more conservative estimate of a change in Malay turnout would affect would be to say that a 1 percent change in Malay turnout results in a 0.5 percent change in the level of BN support. (There are of course more sophisticated statistical techniques which one can use to estimate this but for now, this crude method should suffice.)
To strengthen the argument that the BN is more dependent on a higher turnout than the opposition, I examine voting patterns by polling ‘streams’ using the 2008 parliamentary elections data.
Younger voters are more likely to be either outstation voters or be harder to mobilise because their turnout patterns are not firmly established yet. If the level of support for the BN is higher among the younger voters compared to those who are older, then this can be taken as some indication that a lower turnout would hurt the BN more than the opposition. Table 3 shows the results.
From Table 3 (left), the voters in the ‘middle age’ category are the ones who support BN at the lowest level - 47.7 percent compared to voters in younger streams (50.1 percent) and those in the older streams (52.1 percent).
If the BN depends more on mobilising voters in the below-35 age group compared to those between the age of 35 and 55, then a lower Malay turnout would probably fall disproportionately on the below-35 age group and hence a lower level of BN support.
Table 4 (right) shows the estimated effects of different decreases in Malay turnout on the BN majority.
The underlying assumption here, backed by some evidence from Graph 1, is that a decrease in Malay vote would affect BN more than it would PAS given the conventional wisdom that PAS voters would be more motivated to turn out to vote without the necessary monetary incentives compared to BN voters.
Table 4 also shows the effect of different decreases in the level of turnout on the BN majority. The effects are much larger here since it involves the majority of voters in this constituency - the Malay voters. A decrease in Malay turnout of 1 percentage point will decrease the BN support by slightly more than 500 votes. A 3 percent decrease results in a 1,600-vote decrease and a 5 percent decrease results in a 2,500-vote decrease for the BN.
A conservative estimate of a 1 percent decrease in Malay turnout is not unreasonable given that some Malay voters may want to only return home during the Chinese New Year holidays, though at a much smaller proportion compared to the Chinese voters in KT.
This assumes that the Malay support for the BN and opposition remains unchanged from 2008. There have been no significant incidents since March 2008 which has shifted Malay support one way or another.
It is likely that the election issues that are of concern to the Malay community in KT - which Umno and PAS are campaigning on - will cancel each other out. These include the Altantuuya murder, hudud, Israel’s invasion of Gaza, government corruption, and development projects.
Impact of turnout and voting among non-Malays
The Chinese make up about 94 percent of the non-Malay voting population in KT. Hence, when I refer to the non-Malays here, I use it interchangeably with Chinese voters. Using a statistical software package and a methodology developed by a Harvard political scientist, Gary King, I estimated the Malay turnout rate at approximately 83 percent and the non-Malay turnout rate at approximately 66 percent.
There is good reason to think that the turnout rate among the Chinese would decrease by a significant margin given that Chinese New Year falls one week after the by-election. Many Chinese voters may be expected to stay away from the by-election and only return home during the festive period.
Table 5 (right) shows the nett impact of a decrease in non-Malay turnout from 66 percent to 60 percent and 50 percent respectively, holding BN non-Malay support at 65 percent.
If turnout decreases by 6 percent from 66 percent among non-Malays, the impact on the majority, using the 2008 GE data, is relatively small, at 158 votes.
But this increases to negative 420 votes if the turnout among the non-Malays decreases to 50 percent. This is not surprising especially if we assume that the percentage of non-Malays who live outside the state of Terengganu is higher than that of the Malays because of the lack of job opportunities within the state.
But what if the change in the level of non-Malay support for the BN decreases at that same time as the non-Malay turnout rates also decreases?
In this case, the BN faces a double whammy, non-Malay turnout decreases and among the Chinese who do turn out to vote, they support the BN at lower rates compared to 2008! Table 6 (left) shows the impact of decreases in non-Malay turnout and BN support under different scenarios.
A conservative scenario would be to say that non-Malay turnout decreases by 6 percent and non-Malay support for the BN decreases by 5 percent with a negative 608 impact on the BN majority.
A larger drop in the level of BN support among the non-Malays is certainly possible if the opposition’s strategy of using the ISA arrests of DAP ‘superstar’ and Seputeh MP Teresa Kok and the Ahmad Ismail incident turns out to be effective. There has been nothing to suggest that the level of Chinese support for the BN will increase. It can only decrease.
Independent candidate
Azharudin Mamat is running as an independent candidate perhaps hoping to equal the role of ‘spoiler’ played by grandmother Maimun Yusof, who won 685 votes in 2008.
Conventional wisdom says that when there are independent candidates in a multi-cornered fight, the independent candidate will likely take away votes from the opposition candidate more than from the BN candidate. Maimun’s candidacy in the KT seat in 2008 has been interpreted as having an effect of taking away votes from PAS and contributing to the 628-vote loss by PAS.
However, an investigation into where exactly Maimun (right) obtained her votes shows that her candidacy might not necessarily have taken more votes from PAS than from Umno. It actually shows that she won more votes in places where Umno was stronger rather than in polling stations where PAS was stronger, which would have been the case if one expected her to take more votes away from PAS than Umno.
While this cannot be taken as definitive proof that Maimum took away more votes from the BN than from PAS, it should be a warning that Azharudin’s candidacy would not automatically be detrimental to the PAS candidate. (Had the Akim candidate run, he would almost certainly have taken more votes away from PAS than Umno.) Hence, I assume that the independent candidate has no nett effect on the BN majority.
Postal votes
According to the Election Commission, the number of postal votes in KT has not changed by much since 2008, numbering slightly more than 1,000 votes.
I do anticipate that the number of spoilt ballots in 2008 - 157 - will be reduced in the by-election. A conservative estimate is that this may provide a swing of up to 100 votes in BN’s favour.
Final analysis
Table 7 (right) summarises the different voting components which have been analysed above. The results show the results changing from a 628 majority for the BN in 2008 to a 3,290 majority victory of PAS.
The analysis here is not merely an intellectual exercise on my part. By breaking down voting into different components and trying to estimate the influence of the different components, I can compare the prediction to the eventual result.
And to the extent that the final results deviate from my prediction, which clearly will be the case, I can use the deviations to understand where exactly I got my predictions wrong. And hopefully, in doing so, I can understand better the underlying dynamics of future by-elections and general elections - malaysiakini.com
ONG KIAN MING is a PhD candidate in political science at Duke University. He reserves the right to update his prediction as and when he receives new information from the field.
Posts
0 comments:
Post a Comment