PE2023 voting patterns across Singapore

by Ngiam Shih Tung

Former Senior Minister Tharman Shanmugaratnam was elected President of Singapore in a landslide election on 2 September 2023. Mr Tharman was the clear favorite going into the election and was widely expected to do well, especially in the Western part of Singapore, where he had received 75%-80% of the vote in his Jurong Group Representation Constituency (GRC) in two consecutive general elections in 2015 and 2020. Mr Tharman eventually received 70.4% of the nation-wide vote but following its usual practice, the Elections Department did not release any breakdown of the results by constituency or region within Singapore.

Counting of votes has been decentralized since the first Presidential Election in 1993. In last year’s election, there were 214 counting centres, each of which had between three to seven counting places, and with each counting place roughly corresponding to a polling district. As candidates are permitted to appoint counting agents at each counting place, it is possible for them to observe the voting patterns in different parts of Singapore. Unfortunately, the Elections Department does not officially publish the breakdown and up till now, no candidates in Parliamentary or Presidential Elections have released the breakdowns themselves.

In the 2023 Presidential Election, however, Mr Tan Kin Lian recruited over 500 volunteer polling agents and counting agents, and graciously aceded to a request from Gutzy.Asia to allow the release of detailed counting results collected by his counting agents. This stand is in stark contrast to the caginess shown by other political candidates towards revealing their party’s performance in different areas within their constituencies.

Some of this reticence could be due to the misperception that disclosing precinct-level results is prohibited by law. As early as 1997, however, then-Prime Minister Goh Chok Tong told journalists,

“Everyone knows which area voted for Mr Ong Teng Cheong, which area had support for Chua Kim Yeow, because counting agents were there. Everybody knows…

“Each party – Workers’ Party, People’s Action Party – will have counting agents at the counting station. It is not something secret, which the PAP knows and the other side does not. It is transparent, it is clear as daylight. If a counting agent is alert when votes are counted, they will know how each precinct votes”

Chua Mui Hoong, PM: Precincts with greater support get upgraded first, The Straits Times, 1 January 1997

It is unfortunate that thirty years after vote counting was decentralized in Singapore elections, the Elections Department still does not publish polling district-level results as a matter of course, and that all political parties do not reveal this information either.

Mr Tan’s willingness to share the vote counts collected by his team is a big step forward for political transparency in Singapore, and the volunteers who stepped forward to serve under Mr Tan’s banner have also done a great service for Singapore by contributing to demystifying the electoral process and to breaking down the taboos against political participation.

Mr Tan’s volunteers were spread out across 137 counting centres and managed to obtain the voting results at 563 counting places (each counting centre had between three to seven counting places). This constituted 61% of the total number of counting centres (214) and 64% of the number of counting places (921).

Despite the lack of complete coverage, the results collected by Mr Tan’s counting agents were very close to the actual results for the whole of Singapore. Mr Tharman’s vote share was 70.3% in the sample compared to 70.4% in the official results.

Anecdotally, Mr Tharman was believed to have done well in the western part of Singapore, and this was proven in the sample collected by Mr Tan’s counting agents. Mr Tharman received as much as 89% of the vote in Taman Jurong (JR23 and JR30 polling districts). The figure below shows the vote share received by Mr Tharman, with red indicating a high vote percentage and yellow a lower percentage. Even in Hougang, where he received his lowest vote share, he still received close to 60% of the vote.

Mr Tharman received close to 75% of the vote in Jurong GRC in the 2020 general election, so it is not surprising that he is popular there, but it appears that his popularity extends to other areas in the West as well, including Choa Chu Kang and West Coast GRC. Conversely, he did relatively poorly in Aljunied GRC and Hougang, which the PAP was unable to wrest back from the Workers’ Party in the last election. Of course, even in those Worker’s Party strongholds, Mr Tharman still received an average of 65% of the vote, demonstrating his appeal to a broad swathe of the electorate.

Given Tharman’s dominance in this election, Mr Ng and Mr Tan were effectively left to fight over the 30% of the voters who did not vote for Mr Tharman. In the end, Mr Ng came out ahead of Mr Tan by just 1.8%. No clear pattern was observed in the percentage vote shares for Mr Ng and Mr Tan (Figures 2 and 3, respectively).

Similarly, there was no clear pattern in the distribution of spoilt votes (rejected ballots) even though the actual percentage of spoilt ballots at each counting place varied from 0.6% to 3.7%.

Data Sources

The key goal for the volunteer effort was to promote transparency in the electoral process in Singapore and all the counting data collected is being released to the public domain in support of this goal.

The file “Polling Data v1.0.xslx” contains the vote counts by polling district collected by the counting agents. In some cases, only the counting centre but not the polling district was reported. In those cases, the total votes for the counting centre would be divided equally among all the polling stations which were being counted at the counting centre. Note that there were insufficient counting agents to cover all counting places, so we only have results for 563 out of 921 counting places.

Generally, each counting place corresponds to one polling district. Instead of plotting the results by polling district for this article, results were plotted by the location of the polling stations because polling stations are generally sited close to residents’ homes and this gives a better representation of the spatial distribution of population in Singapore. Some polling districts have two polling stations. In such cases, the vote counts for the polling district were simply split equally between the two polling stations.

The Elections Department published a list of polling station addresses shortly before the election. These addresses were geocoded, and vote shares were plotted using Felt. The file “Polling Stations v1.0.x” contains a list of polling stations and their corresponding polling district and counting place.

Please note that these data are not official results and were collected by volunteers. Some of the results may have been misreported. Please email the author at [email protected] if you have data for any other counting centres or corrections to the data presented.

This was first published on Mr Ngiam’s blog and republished with permission.