1. Introduction

On September 15^th 2016 the Election Commission of Malaysia (Suruhanjaya Pilihanraya Malaysia) published the proposed redelineation of electoral boundaries for State and Federal constituencies. Under this proposal:

• No new Federal constituencies would be created
• 13 new State constituencies would be created in Sabah
• No new State constituencies would be created in states other than Sabah
• 12 Federal constituencies would be renamed
• 36 State constituencies would be renamed

This report provides an overview of the impact of state constituency redelineation on the Selangor State elections. Analysis was performed based on the 2016 1^st Quarter (Q1) electoral roll (before and after redelineation), State and Federal seat results from the 13^th General Election (GE13) and individual historical voting patterns from GE12 (2008) and GE13 (2013).

2. List of Acronyms

The following table shows a list of acronyms used in this document.

Acronym	Full name
PR	Pakatan Rakyat (coalition of PKR, DAP and PAS)
PH	Pakatan Harapan (coalition of PKR, DAP and AMANAH)
BN	Barisan Nasional
UMNO	United Malays National Organisation
GERAKAN	Parti Gerakan Rakyat Malaysia (also known as PGRM)
MCA	Malaysian Chinese Association
MIC	Malaysian Indian Congress
PBB	Parti Pesaka Bumiputra Bersatu Sarawak
PKR	Parti Keadilan Rakyat
DAP	Democratic Action Party
AMANAH	Parti Amanah Negara
PAS	Parti Islam Se-Malaysia
PPBM	Parti Pribumi Bersatu Malaysia
EC	Election Commission of Malaysia
GE12	12th General Election (held in March 2008)
GE13	13th General Election (held in May 2013)

3. Methodology

Prior to the 13th General Election (GE13) we came up with a methodology of predicting election results based on voting patterns in previous elections (https://politweet.wordpress.com/2013/05/05/predictions-for-malaysias-general-election-ge13-pru13/).

Our method relied on mapping polling lane results to individual voters. This process assigned probability values (chance of turnout; chance of voting for each coalition) to the voter that was not affected if they migrated to another constituency. This is important because between GE12 and GE13 527,849 voters migrated between different constituencies.

The impact of voter migration cannot be measured for a single seat just by comparing results of GE12 and GE13 for that seat. An analysis of the whole country needs to be performed. New voter registrations, voters passing away and voters no longer eligible to vote are other factors that require deep analysis.

After GE13 we were able to apply the same estimation method to voters based on GE13 results. By comparing the shift in probabilities we are able to calculate the swing in support for each coalition. Because we base our calculations on individual voters, we are able to calculate shifts in support based on combinations of the following dimensions:

• By Age
• By Race
• By Gender
• By Urban Development Category (rural / semi-urban / urban)
• By Parliament/State Assembly Seat
• By Polling District
• By Locality
• By Seats Won by Specific Parties

The most reliable metric is age because voters are separated into polling lanes based on age. The majority of polling lanes contain between 350 – 700 voters.

By calculating support and swing levels at the individual level we can address the impact of voter migration between seats; voters who have passed away; and new voters registered to vote since GE13.

Because individual voters have their own support and swing statistics, we can calculate the proportion of the electorate that is leaning towards PR (>50% probability of voting PR); leaning towards BN (<50% probability of voting PR); and on the fence (50% probability of voting PR/BN). Voters who are leaning BN might still vote for PR, so this metric is only an indicator of how good the odds are for PR/BN to win support from the electorate.

To see previous applications of this method please read our previous reports at the following links:

• How Barisan Nasional and Pakatan Rakyat Performed with Voters in Sarawak (GE13) (https://politweet.wordpress.com/2016/04/16/how-barisan-nasional-and-pakatan-rakyat-performed-with-voters-in-sarawak-ge13/)
• Evaluating Voter Support in Kuala Kangsar and Sungai Besar Using General Election Results and Twitter (https://politweet.wordpress.com/2016/06/11/evaluating-voter-support-in-kuala-kangsar-and-sungai-besar-using-general-election-results-and-twitter/)
• Analysing Pakatan Rakyat’s Performance with Malay Voters in Peninsular Malaysia (GE13) (https://politweet.wordpress.com/2015/07/27/analysing-pakatan-rakyats-performance-with-malay-voters-in-peninsular-malaysia-ge13/)
• Urban Development Categorisation of Parliament Seats in Malaysia (https://politweet.wordpress.com/2013/05/30/urban-development-categorisation-of-parliament-seats-in-malaysia/)

 

How our method applies to redelineation

For constituency redelineation, the effect is the same as voter migration – constituencies, polling districts and localities may contain different groups of voters before and after redelineation.

Our probabilities are calculated based on General Election results at the Federal level. They are not accurate at the State level in seats that exhibited split voting (where voters chose different coalitions at Federal and State level). However it is still applicable for this analysis as most seats showed similar results for Federal and State elections. The average difference between total votes for each coalition at State and Federal levels was <2%. The one exception is N45. SELAT KLANG where a significant proportion of voters voted for BN at Federal and PAS at State.

Using these probabilities we simulated the electoral performance for Barisan Nasional and the Opposition before and after redelineation to get a sense of how the redrawing of constituency boundaries affected the election result.

4. Selangor State Election Results in GE13

Overview of GE13 Statistics

Total seats: 56

Total voters: 2,048,828

Total ballots: 1,789,863

Total votes: 1,784,732

Voter turnout: 87.36%

The division of voters by ethnic group is shown in the chart below.

Results of GE13

BN Votes: 693,650 (38.87% of total votes)

PR Votes: 1,050,971 (58.89% of total votes)

BN Seats: 12 (21.42% of seats)

PR Seats: 44 (78.57% of seats)

 

Division of Seats Won by Each Party

Party	Seats
BN	12
DAP	15
PKR	14
PAS	15

 

Distribution of Seats by Ethnic Majority

Ethnic Majority	Seats	BN	PKR	PAS	DAP
Malay	34	12	8	14	0
Chinese	14	0	2	0	12
Mixed	8	0	4	1	3

 

Distribution of Voters by Party

The average number of voters per-seat for the state is 36,586 voters per-seat.

The table below shows the number of voters in seats won by each party and the average number of voters per-seat.

Party	Seats	Voters	Average Voters Per-Seat	% of State Average
BN	12	342,291	28,524	77.96
DAP	15	506,674	33,778	92.32
PKR	14	565,892	40,421	110.48
PAS	15	633,971	42,265	115.52

 

By comparing the party average against the state average we can show the distribution of voters among the seats. If the % of State Average is higher that means the seats held by that party are larger than average. If it is lower that means the seats held by that party are smaller than average. If seat sizes were perfectly balanced the % of the state average for each party would be 100%.

From this table we can see that PKR and PAS seats have a higher than average number of voters per-seat.

Share of Vote in Winning Seats

PR Votes (44 seats won): 917,051

PR Share of Vote: 51.38%

BN Votes (12 seats won): 155,560

BN Share of Vote: 8.72%

A minimum of 29 seats is required to form the government. For PR the minimum votes required was 487,985 votes (27.26% of total votes) to win the smallest 29 seats.

Distribution of Voter Support

The chart below shows the distribution of voters based on their probability of voting for BN or PR at the Federal level.

A voter is categorised as Leaning PR if they have a >50% probability of voting for PR or is categorised as Leaning BN if they have a >50% probability of voting BN. Otherwise they are categorised as a Fence Sitter.

49% of voters were leaning towards PR, 27% were leaning towards BN and 24% were on the fence.

Marginal Seats

We define marginal seats as seats where the average support for BN and the Opposition / PR is below 52.5% for both sides. This is calculated as the sum of probabilities of individuals voting BN / Opposition for each seat divided by the number of voters. Any seat that is not marginal is ‘Leaning BN’ or ‘Leaning Opposition’ depending on the average support for each coalition.

Based on this definition, the following 18 seats were marginal in GE13:

Seat Code	Name	Winning Party (GE13)	Avg (Odds of Voting for BN)	Avg (Odds of Voting for Opposition)
N2	SABAK	PAS	49.85	48.78
N7	BATANG KALI	BN	51.25	44.69
N9	PERMATANG	BN	51.36	46.26
N10	BUKIT MELAWATI	BN	46.73	51.81
N11	IJOK	PKR	46.61	52.21
N13	KUANG	BN	49.02	44.13
N16	BATU CAVES	PKR	46.63	51.73
N17	GOMBAK SETIA	PAS	48.21	50.44
N18	HULU KELANG	PAS	48.04	50.89
N24	SEMENYIH	BN	49.44	49.60
N32	SERI SETIA	PKR	47.11	51.30
N33	TAMAN MEDAN	PKR	44.96	51.68
N38	PAYA JARAS	PAS	41.11	52.18
N39	KOTA DAMANSARA	BN	41.59	51.97
N46	PELABUHAN KLANG	PKR	51.15	47.34
N51	SIJANGKANG	PAS	49.23	49.03
N53	MORIB	PAS	50.39	47.75
N54	TANJONG SEPAT	PAS	47.14	50.91

 

Total marginal seats by party:

• 6 BN Seats
• 5 PKR Seats
• 7 PAS seats

N45. SELAT KLANG (PAS) was also marginal but we have excluded it from this list due to the split voting issue.

Swing Values and Seats At-Risk

In GE13, 16 seats swung to BN by an average of 1.88 points within a range of 0.05 points – 5.07 points. 40 seats swung to PR by an average of 6.31 points within a range of 0.48 points – 25.02 points.

For campaigning purposes we can classify the risk of a seat by its swing value and its winning majority. Seats with small majorities are at greater risk of being lost by the incumbent, provided the swing values were small.

However for the purposes of this report we will use a simpler method of only using the winning majority. This will make it easier to compare the GE13 results with our simulation results. The rules are:

1. Any seat with a winning majority of 1 – 1,000 is classified as HIGH RISK
2. Any seat with a winning majority of 1,001 – 2,000 is classified as MARGINAL RISK

The table below lists the number of seats won by each party based on this classification.

Party	Seats	Marginal Risk	High Risk	Total At-Risk
BN	12	4	1	5
DAP	15	2	0	2
PKR	14	0	1	1
PAS	15	1	3	4

 

BN had the most number of seats at risk with 1 high-risk seat and 5 at-risk seats in total. However PAS had 3 high-risk seats in GE13 with 4 at-risk seats in total.

Key points

• PR won 79% of the seats with 51.38% of the vote
• BN won 21% of the seats with 8.72% of the vote
• PKR and PAS seats have a higher than average number of voters per-seat.
• 12/15 DAP seats are Chinese-majority seats
• 8/14 PKR seats are Malay-majority seats
• 14/15 PAS seats are Malay-majority seats
• All BN seats are Malay-majority seats
• PKR has 4 Mixed seats, DAP has 2 Mixed seats and PAS has 1 Mixed seat
• There were 18 marginal seats – 6 BN, 5 PKR and 7 PAS seats
• BN and PAS have the highest number of seats at-risk based on winning majorities

 

 

5. Selangor State Electorate in 2016 Q1

Overview of 2016 Q1 Selangor Statistics

Total seats: 56

Total voters: 2,083,425

Net increase in voters: 34,597

New voters (since GE13): 77,814

Removed voters (since GE13): 48,758

Transferred-in voters from other states (since GE13): 15,164

Transferred-out voters to other states (since GE13): 9,623

Since GE13 there have not been many new voters registered or significant transfers in/out of state. The division of voters by ethnic group is shown in the chart below. The ethnicity of new voters (3.73% of the total) was determined based on their name and should be considered an estimate.

Distribution of Seats by Ethnic Majority

Ethnic Majority	Seats	BN	PKR	PAS	DAP
Malay	34	12	8	14	0
Chinese	13	0	1	0	12
Mixed	9	0	5	1	3
Total	56	12	14	15	15

 

Since GE13 the number of voters and ethnic distribution has remained almost the same, with the only exception being 1 Chinese-majority seat (N14.RAWANG, PKR) becoming Mixed.

 

Distribution of Voter Support

The chart below shows the distribution of voters based on their probability of voting for BN or PR. 49% of voters were leaning towards PR, 27% were leaning towards BN and 24% were on the fence. This is identical to the distribution during GE13.

This creates the expectation that if voter sentiment remained the same as in GE13, the outcome of an election in 2016 would be almost the same. However the change in the number of voters per-seat may give a different picture.

 

Distribution of Voters by Party

The average number of voters per-seat for the state is 37,204 voters per-seat. This is an increase of 618 voters per-seat since GE13.

The table below shows the number of voters in seats won by each party and the average number of voters per-seat.

Party	Seats	Voters	Average Voters Per-Seat	% of State Average
BN	12	342,118	28,510	76.63
DAP	15	517,663	34,511	92.76
PKR	14	577,666	41,262	110.91
PAS	15	645,978	43,065	115.75

(seats won = seats won in GE13)

The table below shows the change in statistics since GE13.

Party	Voters	Voter Population Change (%)	Average Voters Per-Seat	% of State Average
BN	-173	-0.05	-14	-1.33
DAP	10,989	2.17	733	0.44
PKR	11,774	2.08	841	0.43
PAS	12,007	1.89	800	0.23

 

Compared to GE13, the average number of voters per-seat has increased for each Opposition party. For BN the total number of voters and average number of voters per-seat has decreased by a small amount. PKR and PAS seats continue to have a higher than average number of voters per-seat.

Growth by Seat Since GE13

The following graphs show the change in the number of voters in seats won by each party (in GE13).

The total growth in state seats is summarised in the table below.

Party	Seats	Total Gains	Total Losses	Net Change
BN	12	4,207	-4,380	-173
DAP	15	13,340	-2,351	10,989
PKR	14	12,923	-1,149	11,774
PAS	15	12,548	-541	12,007

 

The number of seats with significant gains/losses (more than 1,000 voters) is listed below.

Party	Total Seats	Significant Gains	Significant Losses	Total Significant Seats
BN	12	2	1	3
DAP	15	3	2	5
PKR	14	4	0	4
PAS	15	5	0	5
Total	56	14	3	17

 

Impact of Growth on Seats

After the growth in the number of voters, 1 seat became marginal:

• KUALA KUBU BAHARU (DAP)

The following 2 seats went from Marginal to Leaning PR (having >52.5% average odds of voting Opposition):

• PAYA JARAS (PAS, 52.53% average)
• KOTA DAMANSARA (BN, 53.36% average)

Based on the earlier definition, the following 17 seats were marginal:

Seat Code	Name	Winning Party (GE13)	Avg (Odds of Voting for BN)	Avg (Odds of Voting for Opposition)
N2	SABAK	PAS	49.80	48.87
N6	KUALA KUBU BAHARU	DAP	46.34	49.69
N7	BATANG KALI	BN	50.47	45.64
N9	PERMATANG	BN	51.34	46.35
N10	BUKIT MELAWATI	BN	46.71	51.88
N11	IJOK	PKR	46.49	52.36
N13	KUANG	BN	48.95	44.21
N16	BATU CAVES	PKR	46.57	51.81
N17	GOMBAK SETIA	PAS	47.99	50.67
N18	HULU KELANG	PAS	47.90	51.04
N24	SEMENYIH	BN	49.30	49.74
N32	SERI SETIA	PKR	47.11	51.33
N33	TAMAN MEDAN	PKR	45.01	51.64
N46	PELABUHAN KLANG	PKR	50.99	47.51
N51	SIJANGKANG	PAS	49.25	49.02
N53	MORIB	PAS	50.80	47.37
N54	TANJONG SEPAT	PAS	47.27	50.85

 

Total marginal seats by party:

• 5 BN Seats (decreased from 6 seats)
• 5 PKR Seats (unchanged)
• 6 PAS seats (decreased from 7 seats)
• 1 DAP seat (increase from 0)

N45. SELAT KLANG (PAS) also remained marginal but we have excluded it from this list due to the split voting issue.

 

Key points

The change in Selangor’s electorate due to new voters, removed voters and transferred voters did not significantly change the ethnic distribution or total number of voters for the state.

However the changes in the electorate on a per-seat basis resulted in the following changes that would impact the next election:

• 17 seats had their total number of voters increase/decrease by >1000 voters. This is significant enough to potentially affect the ethnic majority and winning odds for each seat.
• 1 Chinese-majority seat becoming Mixed (N14.RAWANG)
• 1 DAP seat becoming marginal (N6.KUALA KUBU BAHARU)
• 1 Marginal BN seat became Leaning PR (N39.KOTA DAMANSARA)
• 1 Marginal PAS seat became Leaning PR (N38.PAYA JARAS)

 

6. Selangor State Electorate After Redelineation in 2016 Q1

During redelineation the boundaries of existing constituencies are redrawn based on existing polling districts and localities. This involves 2,083,425 voters and 56 seats.

The redelineation process resulted in 443,484 voters being transferred between existing seats. Of these transfers, 165,116 voters (37%) were Malay, 194,859 voters (44%) were Chinese and 77,802 (18%) were Indian. This includes voters transferred between seats won by the same party and between seats from different parties. 19,108 voters (4.3%) were new voters who had their ethnicities determined based on their name.

After redelineation the number of voters and the ethnic composition in each seat changed. This will impact the odds of winning for each party.

The following changes were also made to existing seats:

Seat Code	Old Seat Name	New Seat Code	New Seat Name
16	BATU CAVES	16	SUNGAI TUA
21	CHEMPAKA	21	PANDAN INDAH
26	BANGI	26	SUNGAI RAMAL
36	DAMANSARA UTAMA	36	BANDAR UTAMA
44	SUNGAI PINANG	45	BANDAR BARU KLANG
45	SELAT KLANG	44	SELAT KLANG
52	TELUK DATUK	52	BANTING

 

Distribution of Voters by Party

The average number of voters per-seat for the state remains the same as before at 37,204 voters per-seat.

The table below shows the number of voters in seats won by each party and the average number of voters per-seat.

Party	Seats	Voters	Average Voters Per-Seat	% of State Average
BN	12	324,304	27,025	72.64
DAP	15	652,949	43,530	117.00
PKR	14	570,239	40,731	109.48
PAS	15	535,933	35,729	96.04

 

The chart below shows how average voters per-seat have changed after redelineation.

The chart below shows how the total voters in seats won by each party have changed after redelineation.

The table below shows the change in statistics after redelineation.

Party	Voters	Voter Population Change (%)	Average Voters Per-Seat	% of State Average
BN	-17,814	-5.21%	-1,485	-3.99
DAP	135,286	26.13%	9,019	24.24
PKR	-7,427	-1.29%	-531	-1.43
PAS	-110,045	-17.04%	-7,336	-19.71

 

Compared to the 2016Q1 pre-redelineation statistics, it is clear that the number of voters in BN, PKR and PAS seats have reduced by a net total of 135,286 voters, 81% of which came from PAS seats. The number of voters in DAP seats increased by the same amount, resulting in a 26.13% increase in voters.

The average number of voters per-seat has decreased for BN, PKR and PAS seats while it has increased for DAP seats.

DAP and PKR seats now have a higher than average number of voters per-seat.

 

Distribution of Voters by Seat

The following graphs show the net change in the number of voters in seats won by each party (in GE13).

In summary:

• BN lost voters in 5 seats and gained voters in 2 seats
• DAP lost voters in 1 seat and gained voters in 11 seats
• PKR lost voters in 6 seats and gained voters in 8 seats
• PAS lost voters in 11 seats and gained voters in 2 seats
• The following seats had the largest increases relative to its previous size:
  • KINRARA (+78.55%)
  • SUNGAI PINANG (+91.19%)
  • PANDAMARAN (+115.17%)
  • All of these seats are DAP seats

 Distribution of Seats by Ethnic Majority

Ethnic Majority	Seats	BN	PKR	PAS	DAP
Malay	36	12	9	15	0
Chinese	14	0	2	0	12
Mixed	6	0	3	0	3
Total	56	12	14	15	15

 

The net change after redelineation was an increase of 2 Malay-majority seats; an increase of 1 Chinese-majority seat and a decrease of 3 Mixed seats.

However the actual changes due to redelineation affected 8 seats:

• IJOK (PKR) changed from Mixed to Malay-majority
• BUKIT ANTARABANGSA (PKR) changed from Malay-majority to Chinese-majority
• SERI SERDANG (PAS) changed from Mixed to Malay-majority
• SEMENTA (PKR) changed from Mixed to Malay-majority
• KOTA ALAM SHAH (DAP) changed from Chinese-majority to Mixed
• SERI ANDALAS (PKR) changed from Mixed to Malay-majority
• SRI MUDA (PKR) changed from Malay-majority to Mixed
• TELUK DATUK (DAP, name changed to BANTING) changed from Mixed to Chinese-majority

In summary:

• A total of 5 Mixed seats changed to have an ethnic majority – 4 Malay-majority seats and 1 Chinese-majority seat.
• 2 ethnic majority seats changed to become Mixed.

 

Change in Ethnic Proportion by Party

The ethnic distribution of each seat changed after redelineation. The table below lists the post-redelineation percentage for each major ethnic group for seats won by each party. The change in percentage points is shown in parentheses.

Ethnic Group	BN Seats	DAP Seats	PKR Seats	PAS Seats
Malay	68.57 (+1.09)	22.58 (-3.16)	54.44 (+3.17)	69.34 (+8.72)
Chinese	17.03 (-0.98)	60.89 (+0.19)	27.17 (-0.07)	18.33 (-8.43)
Indian	13.07 (-0.05)	15.26 (+3.16)	17.25 (-3.05)	10.72 (-0.58)

 

The previous section showed that BN, PKR and PAS seats have become smaller while DAP seats have become larger. As a result of redelineation:

• DAP seats represent a larger, more non-Malay population of voters.
• Before redelineation DAP seats contained 74.26% non-Malay voters.
• After redelineation DAP seats contain 77.42% non-Malay voters and grew by 26.13%.
• PAS seats have reduced in size by 17.04% and have a higher percentage of Malay voters with an increase of 8.72 points.
• PAS seats have a higher percentage of Malay voters than BN seats.
• PKR seats still represent a mix of ethnic groups however the percentage of Malay voters has increased by 3.17 points and the percentage of Indian voters has decreased by 3.05 points.
• BN seats have seen little change and still represent a Malay-majority electorate.

 

The charts below show how the proportion of each major ethnic group were affected for each seat. Please note the different y-axis scales for each graph.

Significant changes for BN seats:

1. BUKIT MELAWATI
   6. Drop in percentage of Indians by 6.85 points
   7. Increase in percentage of Malays by 3.35 points and Chinese by 3.56 points
2. JERAM
   8. Increase in percentage of Indians by 8.76 points
   9. Drop in percentage of Malays by 6.95 points and Chinese by 1.84 points
3. KUANG
   9. Increase in percentage of Malays by 9.39 points
   10. Drop in percentage of Chinese by 5.68 points and Indians by 3.76 points
4. SEMENYIH
   6. Increase in percentage of Malays by 6.85 points
   7. Drop in percentage of Chinese by 5.07 points and Indians by 1.67 points

 

Significant changes for DAP seats:

1. BALAKONG
   7. Drop in percentage of Malays by 7.41 points.
   8. Increase in percentage of Chinese by 8.58 points.
2. KINRARA
   1. Drop in percentage of Chinese by 5 points.
   2. Increase in percentage of Indians by 4.21 points.
3. KAMPUNG TUNKU
   5. Drop in percentage of Malays by 5.01 points and Indians by 2.95 points.
   6. Increase in percentage of Chinese by 8.25 points.
4. DAMANSARA UTAMA
   9. Drop in percentage of Chinese by 9.67 points.
   10. Increase in percentage of Malays by 5.54 points and Indians by 4.06 points.
5. SUNGAI PINANG
   17. Drop in percentage of Malays by 17.82 points.
   18. Increase in percentage of Chinese by 10.50 points and Indians by 7.24 points.
6. KOTA ALAM SHAH
   20. Drop in percentage of Chinese by 20.52 points.
   21. Increase in percentage of Malays by 1.04 points and Indians by 19.98 points.
7. TELUK DATUK
   13. Drop in percentage of Malays by 13.32 points.
   14. Increase in percentage of Chinese by 14.61 points and Indians by 1.67 points.
8. There was a pattern in the seat changes where the percentage of Malay voters decreased while the percentage of Chinese voters increased by at least 4% for each ethnic group. This pattern was present in 5 seats.
9. The percentage of Indian voters increased in 7 seats
10. It is worth noting that N47. PANDAMARAN only had a small change in the ethnic distribution despite doubling in size due to redelineation.

 

Significant changes for PKR seats:

1. IJOK
   2. Drop in percentage of Chinese by 2.1 points and Indians by 8.22 points.
   3. Increase in percentage of Malays by 10.37 points.
2. BUKIT ANTARABANGSA
   15. Drop in percentage of Malays by 15.48 points.
   16. Increase in percentage of Chinese by 14.95 points and Indians by 0.97 points.
3. KAJANG
   8. Drop in percentage of Malays by 8.11 points.
   9. Increase in percentage of Chinese by 3.25 points and Indians by 4.65 points.
4. TAMAN MEDAN
   6. Drop in percentage of Chinese by 6.53 points.
   7. Increase in percentage of Malays by 5.75 points and Indians by 0.87 points.
5. SEMENTA
   13. Drop in percentage of Chinese by 13.75 points and Indians by 13.82 points.
   14. Increase in percentage of Malays by 27.83 points.
6. PELABUHAN KLANG
   10. Drop in percentage of Chinese by 10.08 points and Indians by 1.69 points.
   11. Increase in percentage of Malays by 12.20 points.
7. SERI ANDALAS
   12. Drop in percentage of Chinese by 12.53 points and Indians by 19.04 points.
   13. Increase in percentage of Malays by 31.62 points.
8. SRI MUDA
   9. Drop in percentage of Malays by 9.53 points.
   10. Increase in percentage of Chinese by 6.36 points and Indians by 2.89 points.
9. There was a pattern in the seat changes where the percentage of Chinese voters decreased while the percentage of Malay voters increased. This pattern was present in 9 seats.

 

Significant changes for PAS seats:

1. The percentage of Chinese voters decreased by >5% in 12 seats, within a range of 5.6 points – 16.11 points.
2. In 11 seats the drop in Chinese voters was offset by an increase in the percentage of Malay voters. The increase in Malay voters was within a range of 5.07 – 17.45 points.
3. In the remaining seat (N42. MERU) the drop in the percentage of Chinese voters was offset by an increase in the percentage of Indian voters.

Transfer of Voters by Ethnicity and Party

In previous sections we examined net changes in ethnic distribution and the number of voters per-seat after redelineation. In this section we will examine the transfers in detail to identify the origin of total voters, Malay voters, Chinese voters and Indian voters that were affected by redelineation.

The diagram below shows the number of voters transferred between seats won by each party, and the ethnic composition of the transfers (restricted to major ethnic groups). There were no transfers between BN and DAP seats. This diagram does not indicate transfers between seats for the same party (e.g. transfers from one PKR seat to another PKR seat).

The table below lists the statistics used to generate the above diagram. The number of seats involved in the transfers are shown in parentheses under ‘Source Seats’ and ‘Target Seats’.

Source Seats	Target Seats	Transferred Voters	Malay	Chinese	Indian	Malay (%)	Chinese (%)	Indian (%)
BN (3)	PAS (2)	7,318	5,529	813	704	76	11	10
BN (4)	PKR (4)	18,634	8,748	6,037	3,635	47	32	20
DAP (4)	PAS (4)	21,290	14,441	4,259	1,599	68	20	8
DAP (1)	PKR (1)	10,662	8,477	1,429	689	80	13	6
PAS (1)	BN (1)	2,557	2,375	66	93	93	3	4
PAS (7)	DAP (4)	79,005	22,491	40,998	14,716	28	52	19
PAS (9)	PKR (6)	88,256	32,173	44,970	10,325	36	51	12
PKR (2)	BN (2)	5,581	3,415	421	1,723	61	8	31
PKR (9)	DAP (6)	88,233	14,651	48,069	24,546	17	54	28
PKR (3)	PAS (5)	31,165	17,047	6,372	7,279	55	20	23

 

By analysing the transfers between seats owned by each party, we can make the following observations.

For BN seats:

• 5 BN seats lost a total of 25,952 voters to PKR and PAS seats.
  • 5,529 voters (76% of transfers to PAS seats) transferred to PAS seats were Malay
  • 9,886 voters (52% of transfers to PKR seats) transferred to PKR seats were non-Malay.
• 2 BN seats gained a total of 8,138 voters from PKR and PAS seats.
  • 93% of the voters from PAS seats were Malay.
  • Transfers from PKR seats were largely a mix of Malay (61%) and Indian (31%) voters.
• Only 487 Chinese voters were transferred from PKR and PAS seats to BN seats.

For PKR seats:

• 11 PKR seats lost a total of 124,979 voters to BN, DAP and PAS seats.
  • 35,113 voters (28%) were Malay, 54,862 voters (44%) were Chinese and 33,548 voters (27%) were Indian.
• 9 PKR seats gained a total of 117,552 voters from BN, DAP and PAS seats.
  • 49,398 voters (42%) were Malay, 52,436 voters (45%) were Chinese and 14,649 voters (13%) were Indian.
• 92% of the Chinese voters and 73% of the Indian voters transferred from PKR seats went to 6 DAP seats.
• 86% of the Chinese voters transferred into PKR seats came from 9 PAS seats.
  • These transfers mostly went into 3 seats – N19. BUKIT ANTARABANGSA (9,665 voters), N37. BUKIT LANJAN (18,280 voters) and N25. KAJANG (16,067 voters).
  • These 3 seats combined received 44,012 voters (98%) out of the total 44,970 Chinese voters transferred from PAS seats.
• Compared to its gains, PKR lost more voters in more seats with a net loss of Chinese and Indian voters. This caused the percentage of Indian voters in PKR seats to drop by 3.05 points and the percentage of Chinese voters to drop by 0.07 points.
• While the drop in the Chinese proportion is negligible, the distribution of Chinese voters transferred in and out of PKR seats was uneven. As shown in the previous section, 9 PKR seats had a drop in the percentage of Chinese voters.

For DAP seats:

• 5 DAP seats lost a total of 31,952 voters to PKR and PAS seats. 22,918 voters (72%) were Malay, 5,688 voters (18%) were Chinese and 2,288 voters (7%) were Indian.
• 9 DAP seats gained a total of 167,238 voters from PKR and PAS seats. 37,142 voters (22%) were Malay, 89,067 voters (53%) were Chinese and 39,262 voters (24%) were Indian.
• While DAP gained more Malay voters than it lost, the larger increases in the number of Chinese and Indian voters resulted in an increase in the proportion of Chinese and Indian voters.
• The share of Malay voters dropped by 3.16 points, the share of Chinese voters increased by 0.19 points and the share of Indian voters increased by 3.16 points.

For PAS seats:

• 13 PAS seats lost a total of 169,818 voters to BN, DAP and PKR seats. 57,039 voters (34%) were Malay, 86,034 voters (51%) were Chinese and 25,134 voters (15%) were Indian.
• 9 PAS seats gained a total of 59,773 voters from BN, DAP and PKR seats. 37,017 voters (62%) were Malay, 11,444 voters (19%) were Chinese and 9,582 (16%) were Indian.
• PAS seats had a net loss of Malay, Chinese and Indian voters, leading to a 17.04% drop in the number of voters.
  • The greater losses of Chinese and Indian voters compared to Malay voters caused the percentage of Malay voters in PAS seats to increase by 8.72 points.
  • The percentage of Chinese voters in PAS seats dropped by 8.43 points and the percentage of Indian voters dropped by 0.58 points.

Common patterns in the transfers:

• Chinese and Indian voters were transferred from PAS seats to PKR and DAP seats
• Malay voters were transferred from PAS seats to PKR seats
• Malay voters were transferred from DAP seats to PKR and PAS seats
• Chinese and Indian voters were transferred from PKR seats to DAP seats
• Malay and Chinese voters were transferred from BN seats to PKR seats
• Transfers resulted in DAP seats having a higher percentage of Chinese and Indian voters and BN, PKR and PAS seats having a higher percentage of Malay voters

Distribution of Voter Support

Redelineation does not affect the voter sentiment of the total electorate. As it was before redelineation, 49% of voters were leaning towards PR, 27% were leaning towards BN and 24% were on the fence. However the distribution of voter support for seats won by each party has changed. To get a general idea of how much it has shifted, we can compare the before and after distribution of voter support in seats won by each party.

The charts below shows the before and after distribution of voter support in BN seats.

After redelineation in BN seats:

• The percentage of Leaning BN voters increased by 1 point
• There was no change in the percentage of Fence Sitters
• The percentage of Leaning PR voters reduced by 1 point
• The odds of winning for BN are slightly better

The charts below show the before and after distribution of voter support in DAP seats.

After redelineation in DAP seats:

• The percentage of Leaning BN voters was reduced by half to 5 points
• The percentage of Fence Sitters increased by 1 point
• The percentage of Leaning PR voters increased by 4 points
• DAP seats have become much safer for DAP

The charts below show the before and after distribution of voter support in PKR seats.

After redelineation in PKR seats:

• The percentage of Leaning BN voters increased by 1 point
• The percentage of Fence Sitters has remained the same
• The percentage of Leaning PR voters reduced by 1 point
• The odds of winning for PKR is slightly harder

The charts below show the before and after distribution of voter support in PAS seats.

After redelineation in PAS seats:

• The percentage of Leaning BN voters increased by 9 points
• The percentage of Fence Sitters has remained the same
• The percentage of Leaning PR voters reduced by 9 points
• The odds of winning for PAS is much harder

 

Distribution of Malay Voter Support

In a previous report (https://politweet.wordpress.com/2015/07/27/analysing-pakatan-rakyats-performance-with-malay-voters-in-peninsular-malaysia-ge13/) we highlighted that Malay voters were starting to turn away from the Opposition in GE13 in many seats on a national scale. For Selangor the shift in voter support from Malay voters can be seen in the charts below.

Between GE13 and GE12, the percentage of the Malay electorate in the ‘Leaning BN’ category increased by 2%. The percentage of the Malay electorate in the ‘Leaning PR’ category increased by 1%.

The table below summarises the swing in support in seats won by each party in GE13. The swing in percentage points is shown in parentheses.

Party	Seats Won	% Leaning BN GE13	% Fence Sitters GE13	% Leaning PR GE13
BN	12	71 (+7)	8 (-2)	21 (-5)
DAP	15	34 (-9)	28 (-14)	38 (+23)
PKR	14	37 (+5)	14 (+2)	49 (-7)
PAS	15	55 (+1)	9 (-3)	36 (+2)

 

In GE13 DAP showed the greatest improvement in increasing their share of support, however most of it came from voters in the Fence Sitter category. Malay voters swung away from PR in both BN and PKR seats, while PAS made a 2 point gain in their seats.

The table below summarises the change in support from Malay voters in seats won by each party in 2016Q1 before and after redelineation. The change in percentage points after redelineation is shown in parentheses. Please note that the values before redelineation are not the same as it was in GE13 due to changes in the electorate between GE13 and 2016.

Party	Seats Won	% Leaning BN 2016 Q1 After Redelineation	% Fence Sitters 2016 Q1 After Redelineation	% Leaning PR 2016 Q1 After Redelineation
BN	12	71 (+1)	9 (+1)	20 (-2)
DAP	15	21 (-15)	33 (+7)	46 (+8)
PKR	14	38 (+1)	11 (-2)	51 (+1)
PAS	15	61 (+6)	6 (-2)	33 (-4)

 

Redelineation has effectively increased the share of pro-Opposition Malay voters in DAP seats. PAS seats had a 6 point increase in the share of pro-BN Malay voters and a drop in both Fence Sitters and pro-Opposition Malay voters. This is another indicator of how PAS’ odds of success have been reduced due to redelineation.

Change in Share of Voters

By comparing the proportion of voters in the Leaning PR (Opposition) category before and after redelineation, we can measure the change in the share of voters that are leaning towards the Opposition. This will help visualise the impact of redelineation on the probability of the Opposition winning the seat.

The following charts show the change in percentage points in the share of voters Leaning Opposition after redelineation for seats won by each party. Positive numbers are good for the Opposition while negative numbers are good for BN.

The changes are summarised in the table below, based on the scale of change.

Change in % Share of Leaning Opposition Voters	BN Seats	DAP Seats	PKR Seats	PAS Seats
Large Increase (>5 points)	0	4	2	0
Small Increase (0.1 – 5 points)	4	4	4	1
No Change	5	3	0	2
Small Decrease (0.1 – 5 points)	1	2	2	1
Large Decrease (>5 points)	2	2	6	11

 

Following redelineation, PKR and PAS are worse off with 8 PKR seats and 12 PAS seats having a decrease in the share of voters leaning Opposition. As a result the number of Marginal seats increased from 17 to 22.

The following 7 seats became Marginal:

• TAMAN TEMPLER (PAS)
• DUSUN TUA (PAS)
• BANGI (PAS)
• SERI SERDANG (PAS)
• PAYA JARAS (PAS)
• KOTA ANGGERIK (PKR)
• SEMENTA (PKR)

The following 3 seats went from Marginal to Leaning BN (having >52.5% average odds of voting BN):

• PELABUHAN KLANG (PKR, 57.71% average odds)
• SIJANGKANG (PAS, 53.59% average odds)
• MORIB (PAS, 56.45% average odds)

No seats went from Marginal to Leaning PR/Opposition.

The following 22 seats were marginal after redelineation:

Seat Code	Name	Winning Party (GE13)	Avg (Odds of Voting for BN)	Avg (Odds of Voting for Opposition)
N2	SABAK	PAS	49.80	48.87
N6	KUALA KUBU BAHARU	DAP	46.34	49.69
N7	BATANG KALI	BN	50.47	45.64
N9	PERMATANG	BN	52.04	45.74
N10	BUKIT MELAWATI	BN	49.02	49.62
N11	IJOK	PKR	49.55	49.52
N13	KUANG	BN	51.15	40.81
N15	TAMAN TEMPLER	PAS	45.36	47.81
N16	BATU CAVES	PKR	46.12	52.34
N17	GOMBAK SETIA	PAS	48.74	49.91
N18	HULU KELANG	PAS	50.84	48.27
N23	DUSUN TUA	PAS	48.23	50.51
N24	SEMENYIH	BN	50.22	48.71
N26	BANGI	PAS	48.44	49.82
N29	SERI SERDANG	PAS	47.26	51.49
N32	SERI SETIA	PKR	46.24	51.67
N33	TAMAN MEDAN	PKR	47.68	48.74
N38	PAYA JARAS	PAS	47.11	45.39
N40	KOTA ANGGERIK	PKR	46.97	51.96
N43	SEMENTA	PKR	47.64	42.74
N45	SELAT KLANG	PAS	51.40	38.39
N54	TANJONG SEPAT	PAS	47.27	50.85

 

Total marginal seats by party:

• 5 BN Seats (unchanged)
• 6 PKR Seats (increased from 5 seats)
• 10 PAS seats (increased from 6 seats)
• 1 DAP seat (unchanged)

 

Another result of redelineation is that existing state seats now comprise of voters who were previously registered to vote in different seats. 37 out 56 seats contain voters from more than 1 seat. The most affected seat is N45. BANDAR BARU KLANG (renamed from N44. SUNGAI PINANG) which contains voters from 6 different seats. The table below shows the 16 seats which contain voters from more than 2 seats after redelineation.

Seat Code	Name	Winning Party (GE13)	Source Seats
N45	BANDAR BARU KLANG	DAP	6
N19	BUKIT ANTARABANGSA	PKR	4
N25	KAJANG	PKR	4
N9	PERMATANG	BN	3
N21	PANDAN INDAH	PAS	3
N22	TERATAI	DAP	3
N24	SEMENYIH	BN	3
N26	SUNGAI RAMAL	PAS	3
N32	SERI SETIA	PKR	3
N34	BUKIT GASING	DAP	3
N37	BUKIT LANJAN	PKR	3
N38	PAYA JARAS	PAS	3
N43	SEMENTA	PKR	3
N47	PANDAMARAN	DAP	3
N48	KOTA ALAM SHAH	DAP	3
N52	BANTING	DAP	3

 

To understand the impact of redelineation on the odds of BN or the Opposition winning the state elections, we need to perform election simulations based on the 2016 Q1 electoral roll before and after redelineation. We will calculate a best-case and worst-case scenario for BN and the Opposition for the election results and compare the two simulations.

 

7. How elections are simulated

To simulate an election, we take all the individual voters for a given constituency and run them through a process:

1. Each individual will have a probability of turning out to vote; a probability of voting for BN; a probability of voting for the Opposition; and a probability of voting for an Independent/Third party.
2. Our system will generate a random number to test if the voter turned out to vote.
3. If he/she voted, then the system generates another random number to see which party they voted for.
4. The result of the simulation is stored in a database
5. We run 1,000 simulations for each scenario, and use the average results
6. Any seat with a winning majority of 1 – 1,000 is classified as HIGH RISK
7. Any seat with a winning majority of 1,001 – 2,000 is classified as MARGINAL RISK

In the simulations we ran for this report, we adjusted the probabilities to work for a straight fight between BN and the Opposition. For each simulation we also tracked the ethnic divide of the vote e.g. how many Malay voters voted for BN. This was used to estimate ethnic-based swings in support.

 

8. Results of State Election Simulation Before Redelineation

We ran 1,000 simulations per-seat under the following conditions:

• Straight fights between BN and a united Opposition
• No redelineation
• Voter sentiment unchanged since GE13
• Voters vote for the same party at both State and Federal level, based on GE13 Federal Seat results
• Average voter turnout rate based on GE13
• Parties re-contested the same seats as in GE13
• The chance of winning a seat by a given party is based on the number of simulations where that party won the seat
• The upper and lower bounds for results will be shown to give an idea of the best and worst-case scenarios for each coalition

 

A summary of the results is in the table below:

Party	Minimum seats (100% probability)	Maximum seats	Seats gained from other party
BN	12	15	N2.SABAK (PAS, 97% chance) N45.SELAT KLANG (PAS, 100% chance) N46. PELABUHAN KLANG (PKR, 100% chance) N51. SIJANGKANG (PAS, 63% chance) N53. MORIB (PAS, 100% chance)
Opposition	41	44	N10. BUKIT MELAWATI (BN, 100% chance won by PKR) N24. SEMENYIH (BN, 62% chance won by PKR) N39. KOTA DAMANSARA (BN, 100% chance by PKR)

 

Detailed Results Based on Averages of Simulations

Total votes: 1,805,663

BN Votes: 755,893 (41.86% of total votes)

Opposition Votes: 1,049,770 (58.14% of total votes)

BN Seats: 12 – 15 (21.42% – 26.79% of seats)

Opposition Seats: 41 – 44 (73.21% – 78.57% of seats)

 

Division of seats by party

Party	GE13 Seats	2016Q1, Min – Max Range of Seats Won	Potential Loss	Potential Gains
BN	12	12 – 15	2 – 3	3 – 5
DAP	15	15 – 15	0	0
PKR	14	15 – 16	1	2 – 3
PAS	15	11 – 13	2 – 4	0

(seats won = seats won under simulations within a min-max range)

 

Observations:

• BN stands to gain more seats than it loses, with the worst case being it retains the same number of 12 seats
• DAP will retain all its seats
• PKR will see its total number of seats increase by 1 – 2 seats
• PAS will lose 2 – 4 seats and gain no seats

 

Range of Winning Majorities

The table below lists the average winning majorities for seats won by each party compared to GE13.

Party	GE13 Seats	GE13 Average (Winning Majority)	2016Q1, Seats Won	Average (Winning Majority), Min Seats Won	Average (Winning Majority), Max Seats Won
BN	12	2,413	12 – 15	2,023	1,764
DAP	15	13,154	15 – 15	11,914	11,914
PKR	14	6,928	15 – 16	5,904	5,279
PAS	15	5,643	11 – 13	5,216	5,038

(seats won = seats won under simulations within a min-max range)

All parties are expected to win seats with smaller average majorities compared to GE13. The average winning majorities for the maximum number of seats won by each party is smaller than the average for the minimum number of seats. This is because the odds of winning the maximum number of seats are lower and they were won with lower majorities.

Distribution of seats by ethnic majority

Ethnic Majority	Total Seats	BN GE13	PKR GE13	PAS GE13	DAP GE13	BN 2016Q1	PKR 2016Q1	PAS 2016Q1	DAP 2016Q1
Malay	34	12	8	14	0	12 – 15	9 – 10	10 – 12	0
Chinese	13	0	1	0	12	0	1	0	12
Mixed	9	0	5	1	3	0	5	1	3
Total	56	12	14	15	15	12 – 15	15 – 16	11 – 13	15

 

Compared to GE13:

• BN will gain 0 – 3 Malay-majority seats
• PKR will gain 1 – 2 Malay-majority seats
• PAS will lose 2 – 4 Malay-majority seats
• DAP will see no change

Seats At-Risk

The table below summarises the seats considered to be at-risk, based on the average of the winning majorities for each party. The number of seats won if all high-risk seats are lost is shown in the last column. The range for each category is based on the min/max number of seats won.

Party	Seats Won	Marginal Risk	High Risk	Total At-Risk	Seats Won without High Risk seats (Seats Won – High Risk)
BN	12 – 15	6 – 6	1 – 4	7 – 10	11 – 11
DAP	15 – 15	0	1 – 1	1 – 1	14 – 14
PKR	15 – 16	2 – 2	2 – 3	4 – 5	13 – 13
PAS	11 – 13	2 – 2	1 – 3	3 – 5	10 – 10

(seats won = seats won under simulations within a min-max range)

Compared to GE13:

• The number of at-risk seats for BN has doubled from 5 to a maximum of 10, out of which 4 are high-risk.
• The number of at-risk seats for DAP has reduced from 2 marginal risk seats to 1 high-risk seat.
• The number of at-risk seats for PKR has increased from 1 to a maximum of 5.
• The number of at-risk seats for PAS has increased from 4 to a maximum of 5. They are expected to lose 4 – 5 seats, but with 3 seats at high-risk they may end up winning only 10 seats.
• The number of at-risk seats for the Opposition is 8 – 11.

Analysis of Simulation Results

The change in Selangor’s electorate due to new voters, removed voters and transferred voters resulted in some seats changing hands, however the Opposition would still retain control of the state government.

Compared to GE13, the Opposition is able to win the same number of seats with the added possibility of losing 3 seats.  BN will maintain the same number of seats or gain 3 more.

Even without redelineation PAS and PKR are expected to lose seats. The loss of seats can be attributed to small winning margins in GE13; new voters registered in pro-BN polling districts; and split voting in GE13. Both SEMENYIH and KOTA DAMANSARA had 3-corner fights in GE13, which is why a straight fight results in PKR winning the seats from BN.

In the best-case scenario for Opposition parties, DAP and PKR can win enough seats to control the state government without PAS. However if the high-risk seats are lost, then PKR and DAP will still need PAS to govern.

9. Results of State Election Simulation After Redelineation

We ran 1,000 simulations per-seat using the electoral roll after redelineation. The same rules from the previous simulation exercise before redelineation were applied. Please note that any seat names mentioned are using the new names and numbering scheme.

The results are below:

Party	Minimum seats (100% probability)	Maximum seats	Seats gained from other party
BN	17	23	N2. SABAK (PAS, 97% chance) N11. IJOK (PKR, 56% chance) N17. GOMBAK SETIA (PAS, 1% chance) N18. HULU KELANG (PAS, 100% chance) N23. DUSUN TUA (PAS, 5% chance) N33. TAMAN MEDAN (PKR, 15% chance) N38. PAYA JARAS (PAS, 100% chance) N43. SEMENTA (PKR, 100% chance) N44. SELAT KLANG (PAS, 100% chance) N46. PELABUHANG KLANG (PKR, 100% chance) N51. SIJANGKANG (PAS, 100% chance) N53. MORIB (PAS, 100% chance)
Opposition	33	39	N10. BUKIT MELAWATI (BN, 48% chance won by PKR) N39. KOTA DAMANSARA (BN, 100% chance by PKR)

 

Detailed Results Based on Averages of Simulations

Total votes: 1,805,552

BN Votes: 755,613 (41.85% of total votes)

Opposition Votes: 1,049,939 (58.15% of total votes)

BN Seats: 17 – 23 (30.36% – 41.1% of seats)

Opposition Seats: 33 – 39 (58.93% – 69.64% of seats)

Division of seats by party

Party	GE13	2016Q1, Min – Max Range of Seats Won	Potential Loss	Potential Gains
BN	12	17 – 23	1 – 2	7 – 12
DAP	15	15 – 15	0	0
PKR	14	11 – 14	2 – 4	1 – 2
PAS	15	7 – 10	5 – 8	0

(seats won = seats won under simulations within a min-max range)

Observations:

• BN stands to gain at least 7 seats while losing at least 1 seat. This places BN much closer to the required 29 seats needed to control the state.
• DAP will retain all its seats
• PKR will at best maintain its GE13 number of seats and at worst lose 3 seats. Before redelineation PKR was likely to increase its total seats by 1 – 2 seats.
• PAS will lose one-third to almost half its seats and gain no seats. Before redelineation PAS was likely to lose 2 – 4 seats.

Range of Winning Majorities

Party	GE13		BEFORE REDELINATION			AFTER REDELINEATION
	GE13 Seats	GE13 Average (Winning Majority)	2016Q1, Seats Won	2016Q1 Average (Winning Majority), Min Seats	2016Q1 Average (Winning Majority), Max Seats	2016Q1, Seats Won	Average (Winning Majority), Min Seats	Average (Winning Majority), Max Seats
BN	12	2,413	12 – 15	2,023	1,764	17 – 23	2,306	2,055
DAP	15	13,154	15 – 15	11,914	11,914	15 – 15	15,915	15,915
PKR	14	6,928	15 – 16	5,904	5,279	11 – 14	7,048	6,095
PAS	15	5,643	11 – 13	5,216	5,038	7 – 10	2,377	1,935

(seats won = seats won under simulations within a min-max range)

Observations:

• BN stands to gain more seats than it loses, with the worst case being it retains the same number of 12 seats
• DAP will retain all its seats
• PKR will see its total number of seats increase by 1 – 2 seats
• PAS will lose 2 – 4 seats and gain no seats

 

Distribution of seats by ethnic majority

Ethnic Majority	Total Seats	BN	PKR	PAS	DAP
Malay	36	17 – 23	6 – 9	7 – 10	0
Chinese	14	0	2	0	12
Mixed	6	0	3	0	3
Total	56	17 – 23	11 – 14	7 – 10	15

 

Because the ethnic majority of seats has changed, we can’t do a direct comparison with GE13 results. However based on this it is clear that BN and PAS will only represent Malay-majority seats, PKR will represent a mix of different seats and DAP will maintain its representation of mostly Chinese-majority seats.

 

Seats At-Risk

The table below summarises the seats considered to be at-risk, based on the average of the winning majorities for each party.

Party	Seats Won	Marginal Risk	High Risk	Total At-Risk	Seats Won without High Risk seats (Seats Won – High Risk)
BN	17 – 23	5 – 5	2 – 8	7 – 13	15 – 15
DAP	15 – 15	0	1 – 1	1 – 1	14 – 14
PKR	11 – 14	2 – 2	0 – 3	2 – 5	11 – 11
PAS	7 – 10	1 – 1	3 – 6	4 – 7	4 – 4

(seats won = seats won under simulations within a min-max range)

Compared to the results of the simulated election before redelineation:

• The number of at-risk seats for BN has increased further after redelineation – from a maximum of 10 to a maximum of 13, out of which 8 are high-risk.
• Without the high-risk seats, BN will still increase its number of seats won from 12 (GE13) to 15.
• The number of at-risk seats for DAP is the same as it was before redelineation.
• The minimum number of high-risk seats for PKR has dropped from 2 to 0. This is because the odds for BN has improved in 2 previously high-risk (but guaranteed) wins for PKR – N10. BUKIT MELAWATI and N11. IJOK.
• PAS lost many seats and from the 7 minimum seats they are expected to win, only 3 are not at risk.
• If BN or 3-corner fights cause PAS to lose all their high-risk seats, they will only be left with 4 seats. This is a big drop from the 15 seats won in GE13.
• The number of at-risk seats for the Opposition is 7 – 13.

 

Analysis of Simulation Results

The table below summarises the key differences in simulated election results before and after redelineation.

Before Redelineation	After Redelineation
The Opposition wins 41 – 44 seats	The Opposition wins 33 – 39 seats
BN wins 12 – 15 seats	BN wins 17 – 23 seats
Excluding high-risk seats, BN wins 11 seats	Excluding high-risk seats, BN wins 15 seats
BN loses at least 2 – 3 seats	BN loses at least 1 – 2 seats
The number of at-risk seats for BN is 7 – 10 ·       6 – 6 Marginal Risk ·       1 – 4 High Risk	The number of at-risk seats for BN is 7 – 13 ·       5 – 5 Marginal Risk ·       2 – 8 High Risk
The number of at-risk seats for the Opposition is 8 – 11 ·       4 – 4 Marginal Risk ·       4 – 7 High Risk	The number of at-risk seats for the Opposition is 7 – 13 ·       3 – 3 Marginal Risk ·       4 – 10 High Risk
If the Opposition lost all their at-risk seats they can still control the state government.	If the Opposition lost all their at-risk seats they will lose control of the state government.
PKR and DAP combined (30 – 31 seats) can control the state government without PAS seats	PKR and DAP combined (26 – 29 seats) can control the state government without PAS seats in the best-case scenario. However it is less likely due to the smaller number of winnable seats
PKR wins 15 – 16 seats	PKR wins 11 – 14 seats
Excluding high-risk seats, PKR wins 13 seats	Excluding high-risk seats, PKR wins 11 seats
PKR loses at least 1 seat	PKR loses at least 2 – 4 seats
The number of at-risk seats for PKR is 4 – 5 ·       2 – 2 Marginal Risk ·       2 – 3 High Risk	The number of at-risk seats for PKR is 2 – 5 ·       2 – 2 Marginal Risk ·       0 – 3 High Risk
PAS wins 11 – 13 seats	PAS wins 7 – 10 seats
Excluding high-risk seats, PAS wins 10 seats	Excluding high-risk seats, PAS wins 4 seats
PAS loses at least 2 – 4 seats	PAS loses at least 5 – 8 seats
The number of at-risk seats for PAS is 3 – 5 ·       2 – 2 Marginal Risk ·       1 – 3 High Risk	The number of at-risk seats for PAS is 4 – 7 ·       1 – 1 Marginal Risk ·       3 – 6 High Risk
DAP wins 15 seats	DAP wins 15 seats
Excluding high-risk seats, DAP wins 14 seats	Excluding high-risk seats, DAP wins 14 seats
DAP loses no seats	DAP loses no seats
DAP has 1 High Risk seat	DAP has 1 High Risk seat

 

Redelineation has resulted in an electoral map where BN’s odds of increasing their number of seats has improved and the odds of retaining their existing seats improves as well. BN’s gains come at the expense of PKR and PAS, though PAS loses the most. PAS is expected to lose 5 – 8 seats and see their number of high-risk seats increase.

Out of the 7 minimum seats PAS is expected to win, 4 are at-risk. A swing to BN in PAS seats can see PAS reduced from 15 seats in GE13 to 3 seats. This could give BN a potential total of 27 seats, needing only 2 more at-risk seats from PKR in order to win control of the state.

Apart from increasing the average size of their winning majorities, DAP will not see any other gains from redelineation in their existing seats. For PKR they will at best retain the same number of seats (14) as in GE13. It is likely that after redelineation PKR and DAP will have to continue to be on good terms with PAS to retain control of the state government.

10. Exploring the Impact of 3^rd Parties

Based on our simulations we can estimate the impact of a 3^rd party or combination of 3^rd parties that have a race-based appeal. These 3^rd parties can take the form of new political parties that are independent or aligned to Pakatan Harapan (the coalition of PKR, DAP and AMANAH).

For this report we explored a ‘what-if’ scenario where Malay voters turn away from the Opposition to vote for a 3^rd party. Malay voters who voted for BN continue to vote for BN.

This covers the following possible multi-corner fights in the next election:

• PAS contesting in DAP seats
• AMANAH contesting in currently-held PAS seats
• PAS contesting in currently-held AMANAH seats
• Parti Pribumi Bersatu Malaysia (PPBM) contesting against other Opposition parties
• PAS contesting in PKR seats

 

Scenario Details

We calculated the election results of the following scenarios, for the electoral roll before and after redelineation:

• 5% of Opposition Malay votes went to a 3^rd party
• 10% of Opposition Malay votes went to a 3^rd party
• 15% of Opposition Malay votes went to a 3^rd party
• 20% of Opposition Malay votes went to a 3^rd party
• 25% of Opposition Malay votes went to a 3^rd party
• 30% of Opposition Malay votes went to a 3^rd party
• 35% of Opposition Malay votes went to a 3^rd party
• 40% of Opposition Malay votes went to a 3^rd party
• 45% of Opposition Malay votes went to a 3^rd party
• 50% of Opposition Malay votes went to a 3^rd party

The results were calculated based on the previous simulations.

 

Overview of Simulation Results

The charts below show the minimum and maximum ranges of seats won by each coalition under different scenarios where 5% – 50% of the Malay voters turned away from BN and the Opposition. A minimum of 29 seats is needed to control the state government.

Before redelineation:

• The Opposition can sustain a loss of 5% – 20% of their Malay voters to a 3^rd party and still retain control of the state government.
• At 25% the Opposition would retain control with a slim majority of 29 – 34 seats.
• At 35% the Opposition would only win a maximum of 29 seats.
• However BN would be unable to reach the 29 seat line unless the 3^rd party wins 45% of the pro-Opposition Malay voters.
• A loss of high-risk seats won by the Opposition would only result in losing control of the state government if they also lost 35% of their Malay voters
• The 3^rd party would not win any seats even with 50% of the pro-Opposition Malay voters.

After redelineation:

• The Opposition loses control of the state government to BN if 15% of their Malay voters voted for a 3^rd
• At a 10% loss there would be the possibility of a hung assembly if BN performs well.
• At a 25% loss the possibility of BN winning the state becomes certain.
• A loss of high-risk seats won would result in a hung assembly if the Opposition lost 5% of their Malay voters, and losing control of the state government if they lost 10% of their Malay voters
• The 3^rd party would not win any seats even with 50% of the pro-Opposition Malay voters.

The data for the charts are in the tables below:

Before Redelineation

Party	No Chg	5% Malay	10% Malay	15% Malay	20% Malay	25% Malay	30% Malay	35% Malay	40% Malay	45% Malay	50% Malay
BN Min	12	13	15	17	17	22	24	27	27	27	27
BN Max	15	18	21	24	25	27	27	27	27	29	33
Opp. Min	41	38	35	32	31	29	29	29	29	27	23
Opp. Max	44	43	41	39	39	34	32	29	29	29	29

 

After Redelineation

Party	No Chg	5% Malay	10% Malay	15% Malay	20% Malay	25% Malay	30% Malay	35% Malay	40% Malay	45% Malay	50% Malay
BN Min	17	18	24	26	26	30	33	34	34	34	35
BN Max	23	26	28	31	33	34	34	34	35	36	38
Opp. Min	33	30	28	25	23	22	22	22	21	20	18
Opp. Max	39	38	32	30	30	26	23	22	22	22	21

 

Results by Party

The chart below shows the min/max range of seats won by BN for each scenario.

(seats won = seats won under simulations within a min-max range)

 

Category	No Chg	5% Malay	10% Malay	15% Malay	20% Malay	25% Malay	30% Malay	35% Malay	40% Malay	45% Malay	50% Malay
BN Min, Before	12	13	15	17	17	22	24	27	27	27	27
BN Max, Before	15	18	21	24	25	27	27	27	27	29	33
BN Min, After	17	18	24	26	26	30	33	34	34	34	35
BN Max, After	23	26	28	31	33	34	34	34	35	36	38

 

After redelineation BN gains a minimum of 7 seats if 10% of the pro-Opposition Malay voters vote for a 3^rd party. If 10% – 15% of the pro-Opposition Malay voters vote for a 3^rd party, BN is within reach of winning control of the state.

The chart below shows the min/max range of seats won by PKR for each scenario.

(seats won = seats won under simulations within a min-max range)

 

Category	No Chg	5% Malay	10% Malay	15% Malay	20% Malay	25% Malay	30% Malay	35% Malay	40% Malay	45% Malay	50% Malay
PKR Min, Before	15	15	13	10	10	8	8	8	8	7	6
PKR Max, Before	16	16	15	15	15	11	10	8	8	8	8
PKR Min, After	11	11	10	8	6	6	6	6	6	5	5
PKR Max, After	14	14	11	10	10	8	6	6	6	6	6

 

Before and after redelineation, PKR can absorb a loss of 5% of their Malay voters to a 3^rd party. Once they lose 10% PKR will start to lose seats. After redelineation PKR hits a low minimum of 6 seats starting from a loss of 20% of their Malay voters.

The chart below shows the min/max range of seats won by DAP for each scenario.

(seats won = seats won under simulations within a min-max range)

 

Category	No Chg	5% Malay	10% Malay	15% Malay	20% Malay	25% Malay	30% Malay	35% Malay	40% Malay	45% Malay	50% Malay
DAP Min, Before	15	15	14	14	14	14	14	14	14	14	13
DAP Max, Before	15	15	15	15	15	15	15	14	14	14	14
DAP Min, After	15	15	14	14	14	14	14	14	14	14	13
DAP Max, After	15	15	15	15	15	15	14	14	14	14	14

 

Before and after redelineation DAP may lose 1 seat if 10% of their Malay voters vote for a 3^rd party. This loss only becomes certain if they lose 30% (before redelineation) or 35% (after redelineation) of their Malay voters. If they lose 50% of their Malay voters, they may lose another seat.

The chart below shows the min/max range of seats won by PAS for each scenario.

(seats won = seats won under simulations within a min-max range)

 

Category	No Chg	5% Malay	10% Malay	15% Malay	20% Malay	25% Malay	30% Malay	35% Malay	40% Malay	45% Malay	50% Malay
PAS Min, Before	11	8	8	8	7	7	7	7	7	6	4
PAS Max, Before	13	12	11	9	9	8	7	7	7	7	7
PAS Min, After	7	4	4	3	3	2	2	2	1	1	0
PAS Max, After	10	9	6	5	5	3	3	2	2	2	1

 

After redelineation, a 5% loss of PAS’ Malay voters to a 3^rd party results in PAS having their minimum number of seats drop from 7 to 4 and their max number of seats drop by 1. This effectively means they lost 1 seat.

A 10% loss results in PAS’s max number of seats drop further, down to 6 seats.

A 15% loss results in PAS winning 3 – 5 seats, which is half of what they were projected to win.

Before redelineation, even a 30% loss in Malay voters allows PAS to win 7 seats. After redelineation a 30% loss would result in PAS only winning 2 – 3 seats.

 

11. Conclusion

Before redelineation, the Opposition had good odds of retaining control of the Selangor state government. PAS was expected to lose 2 – 4 seats but this was offset by gains by PKR, allowing the Opposition to win 41 – 44 seats. It would even be possible for Pakatan Harapan to win control of the state without needing PAS.

After redelineation many changes took place, giving BN a better chance to win control of the state particularly if there are 3-corner fights.

443,484 voters (21% of the total electorate) were transferred between existing seats. The movement of voters of different ethnicities in different amounts resulted in an increase in ethnic-majority seats. 5 Mixed seats changed to have an ethnic majority – 4 Malay-majority seats and 1 Chinese-majority seat.

This makes it easier for political parties that appeal to specific ethnic groups to campaign in those seats. This also creates a situation where ethnic minorities are a lower priority for campaigners.

The key points on transfers by party are:

• Non-Malay voters were shifted into DAP seats on a large scale, causing DAP seats to grow by 26% and represent a larger, more non-Malay base (77.42% non-Malay, increased from 74.26%).
• Malay voters were shifted into PKR seats, resulting in 3 PKR seats becoming Malay-majority seats and the proportion of Malay voters in PKR seats increasing by 3.17 points while the proportion of Indian voters dropped by 3.05 points.
• PAS seats became smaller by 17%. The movement of non-Malay voters out of PAS seats resulted in the proportion of Malay voters in their seats increasing by 8.72 points.
• BN seats became smaller by 5% and still represent a Malay-majority electorate
• 37 seats contain voters who were previously registered to vote in other seats
• 16 seats contain voters from 3 or more seats, the worst being N45. BANDAR BARU KLANG which has voters from 6 seats. This can lead to confusion for voters on election day as the party they voted for or had worked with previously might not be on the ballot. It also increases the difficulty for campaigners as they have to familiarise themselves with districts from (formerly) different seats.

 

After the voters were transferred, the share of support in each seat changed. PAS lost support in 12 seats while PKR lost support in 8 seats. This caused 7 seats to become Marginal and 3 seats to move from Marginal to Leaning BN.

DAP also lost support in 4 seats but the drop was not significant enough to affect DAP’s odds of retaining the seats. In fact DAP will increase the size of their average winning majorities and gains better odds of retaining their seats through redelineation.

From our simulations and the pattern of voter transfers we know that PAS was the most affected by the redelineation process. Compared to GE13 where they won 15 seats, after redelineation they are expected to win 7 – 10 seats. From that total, 3-6 seats would be won with a majority of less than 1,000 votes, making them high-risk seats.

A swing to BN could result in PAS winning only 3 – 4 seats, placing BN much closer to reaching the 29 seats needed to win the state. PAS has the most to lose in Selangor due to redelineation.

Because of the lower odds for PAS and PKR, BN is likely to increase their number of seats from 12 to 17 – 23. A positive swing to BN would result in additional PAS and PKR seats falling to BN. This means that PKR and DAP will likely have to continue to maintain an alliance with PAS to retain control of the state government.

If AMANAH wins enough seats then Pakatan Harapan will not need PAS seats to govern. However 3-corner fights between BN, the Opposition and a 3^rd party are likely to cause the Opposition to lose control of the state. If 5% – 10% of Malay voters opted for a 3rd party (PAS, AMANAH or PPBM) this can lead to PAS and PKR losing enough seats for BN to gain the advantage.