Tag Archives: election statistics

A Brief History of Elections III: SDP

Everyone is interested and excited about how the opposition would fare in the next elections. At least coming from someone who can’t vote and is an observer. It is extremely difficult to predict how people vote, because last-minute issues can pop up and turn the tide of the supposed winner. But at the same time, assuming there’s no such Black Swan event, there is a small amount of certainty we can predict how people cast their votes, by speculating on their motivations, the context – and a little bit of history.

Since there’s no straw poll to see which party has the most amount of goodwill or support, we’ve to look at past performances to peer into the future. I’ll do these in a few ways: 1) percentage of popular votes, 2) percentage of total parliamentary seats contested, 3) percentage of votes in contested wards. I’ll elaborate later. Have a look at the first graph of this post:

Share of popular votes in %. Source: Singapore Elections

* The results of SDA in 1997 are from SPP, the party which Chiam See Tong founded after leaving SDP. The results of NSP are the same as SDA in 2001 and 2006 because NSP was a major component party in SDA.

As my post suggests, I’ll be looking at the SDP (the yellow bar). Between 1988 and 1997, they were ranked third in terms of popular vote share, or the second best-performing opposition party. In 2001, they emerged second in popular vote share, but this is due to the WP’s exceptionally poor performance. In the last elections, they were ranked bottom.

Why does popular vote share matter? Seeing how the parties score is an indicator of their support across the country (hmm, at least in contested wards). By ranking the PAP and the major opposition parties, a pattern can emerge that suggests which party S’poreans prefer, elections after elections. As you can see, the popular vote share of the SDP has been shrinking over the past 5 elections. This graph is actually flawed, as no opposition party has put up candidates in all wards. If all the opposition parties and the PAP fight in every ward i.e. each ward having a 4-way fight, the popular vote share graph would be more accurate in terms of gauging voter support.

But this is still useful – see which party is most consistent. And from the graph, SDP has been consistent in the wrong sense, that its popular vote share is on a gradual decline. The next elections seem to be bust or boom for them – either they continue their decline (below 8 percent), or they jump to the 10-odd percentage points of the popular vote share. It took the NSP (first elections 1988) four elections to increase their popular vote share beyond 10 percent, and even then, as one of the two major component parties of the SDA. The SDP, with a longer history, should have it easier to reverse their decline.

Absolute no and % of seats contested. Source: Singapore Elections

This graph shows the absolute number and percentage of seats which SDP has contested in the past elections. Why is this important? Well, obviously the opposition party’s goal is to form the next govt, and expanding their presence in Parliament is the first step to projecting their influence. Ideally, a party’s number and percentage of seats contested should increase, suggesting it is increasingly confident of challenging the ruling party. And the number of seats contested shows how many candidates it can find, which means if the party is able to retain its core members or find new ones.

For the SDP, their absolute number and percentage of seats contested has been on a descending trend. Ironically, the sharp drop in seats contested in 1991 coincided with the election of 3 candidates. As the first graph on popular vote share corroborates, the SDP is on a decline; decreasing support from voters, and smaller percentage of total seats contested.

% of votes in contested wards vs. popular vote share. Source: Singapore Elections

The red line here represents the percentage of votes in SDP-contested wards. That means, if we combine all the wards which SDP contested to form one entity, we can calculate the percentage which SDP scored only in its contested wards – unlike the first graph, where popular vote means including votes from wards which SDP did not contest.

Hence we are seeing a micro picture of PAP vs SDP (there might be 3-cornered fights, but the results of the third party should be minute). The SDP’s percentage of votes peaked in 1991, where it won a handsome 3 seats, but from then on, it has been on a decline too. In other words, it is doing quite badly against the PAP, if the wards which it contested represents S’pore, and there were only two parties, the PAP vs. SDP. As I mentioned, it seems to be going bust.

Now, for something interesting:

% of seats contested vs. % of votes in contested wards. Source: Singapore Elections

In 1991, where the percentage of seats which SDP contested fell, and in absolute number, only 9, the percentage of votes it received in contested wards was 48.6 percent. When it fielded a slightly larger number of candidates, its percentage of votes fell instead. This can be attributed to both SDP weaknesses and PAP strength. But one lesson from this statistic is that a party can still do well in its contested wards, even if it fields a small number of candidates.


If I use the above graphs to predict the SDP’s performance, they would probably do very badly. However, they are active in the new media, and recently presented their Shadow Budget. And from a glance at their website, they seem to be strong and confident. Whether voters will give them a chance to reverse their decline is another question, to be answered soon.


Leave a comment

Filed under Statistics

The gravity strategy of WP

I was reading this commentary, when I thought of the WP’s strategy to focus on constituencies around Hougang. I believe this strategy was implemented since 2001. Hence during the 2006 general elections, they contested these areas:

GE 2006 electoral map. Source: Wikipedia

As you can see, Hougang is smacked in the middle of all the wards which they contested. This is interesting! So I’ve a hypothesis: does the distance from Hougang affect electoral results of the WP?

So I’ve categorized the wards which WP contested:

0: Hougang

1: Aljunied GRC

2: East Coast GRC, AMK GRC

3: Joo Chiat, Nee Soon Central and Nee Soon East SMCs

These values will form the x-axis in the graph I’ll show later. The larger the number, the farther they are from Hougang.

WP graph. Source: Elections Department

You might need to click on the graph for a larger view. Anyway, the y-axis represents the elections’ results. So I plotted the graph, and with these points, I drew a best-fit line. The gradient is negative, which suggests an inverse relationship between election results and distance from Hougang. Hence the WP’s strategy is somewhat working. Aljunied GRC, with envelops Hougang, ranks second in electoral performance. East Coast GRC, which is Aljunied’s neighbour, ranked third. AMK, though in the same category as East Coast GRC, is second from the bottom, and the only odd one out. The worst performers are the SMCs, Joo Chiat, NSC and NSE, which are in category 3, the farthest from Hougang.

For the upcoming elections, it seems WP will contest the new wards of Seng Kang West SMC, Punggol East SMC, Nee Soon GRC and Moulmein-Kallang GRC. Let’s give each ward a value.

1: Punggol East

2: Seng Kang West

3: Nee Soon GRC

4: Moulmein-Kallang GRC

It’s necessary to give Moulmein-Kallang GRC a value of ‘4’, because the wards which WP intends to contest do not border the GRC at all. In contrast, in 2006, the places where WP contested were contiguous to one another. As WP is not contesting AMK, Nee Soon GRC seems to be ‘chopped off’ from the WP’s intended battlegrounds. From the linear graph, it seems logical to argue WP might be victorious in Punggol East, polling above 40 percent in Seng Kang West, about 30-odd percent in Nee Soon GRC, and score miserably in Moulmein-Kallang GRC.

Er-hem. Obviously my predictions are highly inaccurate. But this is all hypothetical. The graph does not take into account GRC or SMC, the voters, the candidates etc. I’ll revisit this blog after the elections to see if my predictions match the results (:

Leave a comment

Filed under Strategy

A Brief History of Elections II: 3-way fights

Since the introduction of GRCs in 1988, only one GRC has faced a fight with more than two parties – Marine Parade GRC in the 1991 by-elections. And it is a 4-cornered fight to boot.

% of votes. Source: Elections Department

As you can see, the PAP received nearly three-quarters of the vote (72.94), while the SDP, with its name recognition coming after its recent 1991 general elections success, took under a quarter (24.5). The NSP and SJP, both relatively new and unknown parties, lost their electoral deposits. But this by-elections was special. It came a year after the general elections, as then PM Goh wanted a stronger mandate, and for JBJ to contest against him. A new candidate, Teo Chee Hean, was also introduced.


What are the lessons for the opposition parties? Don’t fight the PM himself, of course! But WP came up with roughly 33 percent of the votes in Ang Mo Kio GRC in 2006. Secondly, in a 3-cornered or 4-cornered fight, one or two parties are bound to lose their electoral deposits i.e. <12.5 percent of votes polled.

% of votes. Source: Elections Department

For example, in 1991, the WP challenged the PAP in Bukit Merah. There was an Independent candidate who threw in his hat, and of course, lost spectacularly, polling 1.63 percent. Lesson for Andrew Kuan, who might run in Joo Chiat? 😛 But he seems rich enough to lose the deposit… This example uses an Independent in a SMC. What about parties in a SMC?

% of votes. Source: Elections Department

Chua Chu Kang SMC, 1997. Unfortunately for the DPP, the Independent candidate was better than them, and the DPP guy lost his electoral deposit. Even if the Independent’s and the DPP’s votes went to the NSP candidate, he still could not beat the PAP MP. There was no such nonsense of splitting the non-PAP votes: in this SMC, the PAP held sway. Learning points? New political parties in 3-cornered fights are extremely vulnerable. By 1997, NSP was in its third elections and would have gained some name recognition. If the youngest and probably smallest Reform Party fights the more established parties NSP or SDP in some SMCs, well, history has shown the Reform Party would be at a disadvantage. And that might mark the start of the end for its embattled leader. But that said, the Reform Party is different from the DPP, and 2011 is different from 1997. Just a point they might wanna bear in mind.

I can’t find an example in recent electoral history (from 1980s) of a SMC which was contested against the PAP by two established parties. So 2011 might make history: 1) a second GRC seeing more than two contestants, 2) two well-established opposition parties entering the contest, not one strong and one weak. I’m referring to Moulmein-Kallang, and there’s no example from past elections. It seems bleak for one of the two opposition parties though – one party has a very good chance of scoring under 12.5 percent. But which party, NSP or WP? While the WP is the older and more experienced of the two, the NSP has gradually increased its popular vote share, first by itself, later by being part of the SDA, where it contributed the most candidates among the component parties. And judging from its chief Goh Meng Seng’s blogs, the NSP seems to have strengthened itself for the coming elections.

Of course, the best scenario for them is for one of them to win. There are a few subsets: 1) X wins majority, by marginally – obviously Y would poll so low to lose its electoral deposit, and 2) X wins the most, followed by Y, and PAP loses its electoral deposit, 3) X wins the most, followed by PAP or Y, and no one loses electoral deposit. Go permute the possibilities yourself.

As I mentioned in a previous blog, 3-way fights are in fact beneficial for voters to choose the party and the ideology which they can best identity with. Assuming the parties got ideologies and policies in the first place lah. But analyzing from the opposition’s perspective, 3-way fights breed uncertainty and tension for them. They might appeal to ‘fringe’ views or say something outrageous or shocking during rallies. Dunno them, if they are despo enough.

As for the SMCs, the principle I teased out from past examples is the same too. 1) One party has a good chance of losing very badly and 2) Newer and smaller parties are at a disadvantage. It seems the SDA and WP might face the PAP together in some wards. I suspect WP might come up tops, given their experiences. And the bad press surrounding SDA doesn’t help them.

Besides giving voters a greater variety of choices, these 3-cornered fights can ultimately consolidate the opposition scene in S’pore. SDA, with only two remaining component parties, seems to be on the verge of certain electoral extinction: it is losing relevance. A quick scan through the elections results shows the PKMS has been consistently beaten, and the SJP is a virtual unknown besides its name. In wards where there are 3-cornered fights, the party which polls after the PAP (assuming they lose) would have the strongest say to contesting there again. That is if the ward never vanish lah. And that would force the worst-performing parties to 1) go other wards and engage in other 3-cornered fights, or 2) improve their electability by better campaigning or better policies etc. The latter is what will benefit voters.

In the next blog, I’d touch on the popular votes and number of seats contested by the opposition parties.

1 Comment

Filed under Statistics

A Brief History of Elections I

Electoral boundaries. Source: The Online Citizen


General elections take place in a context, and this context includes a bit of history. As S’pore is a city-state, in terms of geography and demography, it is highly homogeneous. There are no rural-urban, highland-lowland, coastal-inland divides, because S’pore is highly urbanized. Furthermore, the bulk of the population resides in HDB flats, so Tampines isn’t much different from Jurong, or Woodlands from Punggol. It is nearly impossible for politicians to exploit geographical differences or base politics on locations. There is no concentration of the poor or rich such that their votes can swing an election. The population is also overwhelmingly Chinese, and there is no discernible ethnic enclave, such that in each constituency, it is more or less reflective of the national average. But in some places, like Tampines, the Malays have an above-average presence. Yet there is scant evidence to suggest minority ethnic or religious groups vote as a bloc.

While S’pore has beeb criticized as a one-party state, technically such a term should have been dropped from 1981, with the Anson by-elections producing the first opposition MP since independence. Since 1981, Parliament always has at least 1 non-PAP MP. The number of political parties taking part in each elections is relatively high for a tightly-controlled country.

Number of political parties/indpts in GE. Source: Elections Department

As you can see, 2006 seems to be the exception rather than the norm of the number of political parties in the elections. Independents are lumped together as one group for convenience. In 2011, it is expected 8 parties will contest the elections (PAP, WP, SDP, SPP, SDA, NSP, DPP, Independents). Though more parties means a greater variety of choices for voters, it remains to be seen if it translates into quality. For example, the DPP contested in 1997 and 2001, and in each elections one of their candidates lost their electoral deposits i.e. vote share below 12.5 percent. Furthermore, the DPP is in reality a father-and-son team, and it is EXTREMELY surprising that according to news reports, they will be contesting Marine Parade and Tanjong Pagar GRCs. Where are they going to get 10 candidates in the first place?

At least they are lucky to be contesting against the PAP. In Moulmein-Kallang GRC, the WP and NSP might be contesting against the PAP. Most SMCs have not been settled, at least not publicly. In the next blog, I shall look at three-cornered fights in the past.

Leave a comment

Filed under Statistics