by Ong Kian Ming
Jan 17, 09 8:01am
Malaysiakini

There is this phenomenon in the United States called Monday Morning Quarterback.

It describes a person who says that he or she always knew what the outcome of the Sunday football games was going to be and proceeds to give an explanation for why the game resulted in the eventual result. In other words, this person is passing judgment from a position of hindsight, which is always 20/20.

I could easily play this game in regard to the upcoming Terengganu by-election. It would be far too easy for me to list down, on Sunday, reasons as to why PAS won the by-election and at the same time, to have ready an alternative list of reasons as to why the BN/Umno managed to win the same by-election in case of an upset.

It is however a far harder and much riskier endeavour to predict who is going to win, and by what margin, before the fact.

In a previous analysis piece in Malaysiakini, I predicted that PAS would win by a majority of 3,300 votes based on a breakdown of different turnout and voting components. In the same article, I also reserved the right to update my prediction as and when new information was made available to me.

I came upon new information recently in the form of two Kuala Terengganu surveys conducted by the Merdeka Center, one in December 2007 before the March 2008 general election and another one conducted just last week.

I then used my own methodology to make an educated guess as to what the voting patterns will be today in the KT by-election.

Don’t expect by-election to be a close fight

Let me make it absolutely clear that the surveys done by Merdeka Center do not give conclusive or definitive results as to how the Malay and Chinese voters were/are going to vote.

The reason is because when asked about their voting intentions, about half of the voters did not indicate who they were going to vote for by responding ‘don’t know’ or ‘my vote is secret’.

In addition, there is always the possibility that some respondents will mask their true voting intentions because of the fear that the surveyor may be someone from the Special Branch who is trying to find out if this individual is leaning opposition or BN.

In other words, some voters may say that they are voting for BN to the surveyor but actually have no intention of voting for BN. This makes predicting or forecasting voting trends based on survey results much more of an art than science.

But a political analyst worth his or her salt should be willing to put his or her analytical skills to the test by trying to guess the voting intentions of those who did not indicate their voting preference using a variety of tools and techniques.

For example, one can try to compare the answers of those who have indicated their voting preference either for BN or for the opposition and ascertain if an individual who has not revealed his or her voting preference is more like the typical BN or opposition voter. Or one can try to estimate the changes in the underlying responses of those surveyed over time to guess at whether the voters are leaning more towards BN or the opposition compared to past surveys.

I have used a variety of quantitative and qualitative methods to arrive at the prediction I am about to make.

For the purpose of brevity, I shall not go into the details here. If my predictions are in the ballpark of the eventual result, I may explain it later. If my prediction is way off, it’s back to the drawing board for me.

Regardless, I am making my prediction public so that I can be held accountable for my predictions after the fact. The easier alternative would be to play it safe and to say that the race is too close to call and then try to explain the outcome after the fact – like a good Monday Morning Quarterback.

But my analytical findings tell me that the eventual outcome will NOT be close and I do not think that I would be honest if I told my friends otherwise.

How I calculate the predicted majority

Here then are the details of my prediction and the steps I use to calculate the predicted majority.

In the previous article, I said that I did not have conclusive evidence to show me that the Malay vote was trending either to the opposition or BN. I simply said that split voting that worked in BN’s favour in March 2008 would not occur again in this by-election.

But by comparing the underlying responses of those in the 2007 and 2009 surveys using my own estimating techniques, I found consistent evidence that the Malay vote has swung to PAS since the time when the 2007 survey was taken in December of that year.

My estimates show that the Malay vote will swing to PAS by approximately 4%. Given that my estimate of the Malay support for BN was 47% in 2008, this means that the overall Malay support for the BN will fall to 43% or that PAS will win 57% of the Malay vote (not including postal votes).

With an estimated Malay turnout of 81% (slightly less than 2008) and by excluding spoilt votes and votes for the independent candidate (which I estimate to total approximately 2%), this translates into a majority of 7,720 for PAS among the Malay voters.

Among the non-Malays (mostly Chinese), I estimate that the turnout will decrease by approximately 10% from 65% in 2008 to 55% in 2009, primarily because of the Chinese New Year holidays that will take place the following week.

Among the non-Malays who will turn out, I estimate a BN support level of 45%. This represents a significant 20% swing in the non-Malay vote from 2008. This is consistent with the Merdeka Center survey results as well as the feeling from the ‘ground’ that the Chinese vote has swung against the BN and in favour of the opposition. Because of the relatively lower turnout and small percentage of Chinese voters in this constituency, this works out to a 510 vote margin in favour of PAS.

The Malay and non-Malay majorities calculated above results in an 8,230 majority in favour of PAS. However, there are approximately 1,300 postal votes in this constituency. Assuming that 90% of the postal votes will go to the BN, this means that BN will have a majority of 1,170 votes from postal votes alone.

Subtracting this figure from the 8,230 majority, I arrive at the final prediction of 7,060 votes or approximately 7,000 votes. This represents a 3,700 vote increase from my initial prediction of 3,300 votes for PAS.

PAS to gain 4% swing in Malay votes

The main reason for this is because of my assumption, based on my own analysis of the Merdeka Center survey results from 2007 and 2009, that the Malay vote has swung in favour of PAS by a margin of 4%.

Some may say that for me to predict a swing in the Malay vote to PAS by a margin of 4% is foolhardy and doesn’t make sense given that most people are saying that the Malay vote is split 50/50 or that PAS has only a slight edge.

Again, I could have easily hedged my bets and echoed what everyone else is saying. But my analysis of the underlying responses in both surveys leave me with no other choice but to conclude that the Malay vote will swing in PAS favour.

Of course, the underlying sentiment may have changed among the Malays in the week since the 2009 survey was taken or that the accusations by the opposition of massive vote-buying on the part of the BN will come true or that the final push by the BN to undermine the reputation of PAS among the Malays may have worked to turn the tide in BN’s favour.

I have no way of ascertaining if any of this is true. Neither do the people on the ground. Their guess in regard to the effectiveness of these last-minute strategies is as good as mine. And my instinct tells me that while these sorts of strategies may have worked in the past, they will not work in this KT by-election.

So there you have it, my prediction is that PAS will win by a majority of 7,000 votes. Tonight, I will know if my analytical skills have served me well … or not!