How they stole the Malaysian election

by Adam Carr

At the Malaysian elections on 5 May, the ruling National Front (Barisan Nasional or BN) coalition, led by Prime Minister Najib Razak, polled 5,237,699 votes, or 47.4% of the vote. The opposition People’s Pact (Pakatan Rakyat or PR) coalition, led by Anwar Ibrahim, polled 5,623,984 votes, or 50.9% of the vote. Yet the BN won the election in terms of seats: and not by a whisker, but with the comfortable margin of 133 seats to 89. The PR increased their vote by 2.9%, while the BN vote fell by 3.9%, yet the PR made a net gain of only 7 seats.

How did BN pull off this remarkable feat – 60% of the seats with 47% of the vote, leaving the opposition with 51% of the vote but only 40% of the seats? In a system of proportional representation, the opposition would have won 115 seats, leaving BN with 107. Malaysia, however, uses the system of single-member constituencies, combined with first-past-the-post voting: a system which always produces distorted results. At the 1997 British election, Labour polled 43% of the vote but won 63% of the seats. But the distortion in such systems is always in favour of the party which leads in terms of votes. A distortion in favour of the losing party requires a different explanation.

Anwar has chosen to blame electoral fraud for the result. Attention has focussed on the use of indelible ink which turned out not to be indelible, manipulation of the electoral roll, multiple voting, foreign voters (including an alleged 600,000 Bangladeshis) voting in key seats, and other well-worn electoral tricks. The opposition has also pointed to the almost unanimous support for the government from the print and electronic mass media, all of which is owned by government supporters.

But all these allegations, even if completely true, miss the point. And the point is that the opposition still polled a clear majority of the vote. Perhaps if there had been no electoral fraud, perhaps if the opposition had been given fair coverage in the media, they would have won by an even wider margin. But the fact is that they won despite all these factors. The key question is why the PR’s majority of votes did not translate into a majority of seats, or even come close to doing so.

The answer lies in the gross malapportionment of Malaysia’s 222 constituencies. Rigging constituency boundaries is commonly called gerrymandering: deliberately designing the boundaries of an electoral district so as to assist a particular candidate to win it. This practice is widespread in the United States, and in fact is mandated by the Voting Rights Act, which requires the creation of districts with Black or Hispanic majorities.

But this is not what we see in Malaysia. What we see instead is malapportionment: allocating a disproportionately large number of seats to a state or region which favours a particular party, while allocating a smaller number to a state or region which opposes it. Malaysia’s electoral system is grossly malapportioned and has been ever since independence in 1957. Up until now, this has not mattered much, since the United Malays National Organisation (UMNO), which is the largest component of the BN, has always won, with its allies, a majority of the vote.

Malaysia’s constituencies are malapportioned at two levels. The first is malapportionment among Malaysia’s 13 states and three federal territories. The following table shows the states and territories in order by the number of enrolled voters.

This table demonstrates, first, that the two East Malaysian states of Sabah and Sarawak have a greatly disproportionate number of seats allocated to them compared to the West Malaysian states. They have 22 more seats than their proportion of enrolled voters entitles them to. Secondly, within West Malaysia, the state of Selangor is severely under-represented. It has 12 fewer seats than it would be entitled to under a proportionate distribution. Thirdly, most of the other West Malaysian states, plus the federal territory of Kuala Lumpur, are also under-represented, though not as severely as Selangor. The only exceptions are Pahang and Perlis, which are over-represented, and Perak, which has exactly the representation it would be entitled to.

The political consequences of this malapportionment among the states are clear. BN’s strength lies in the more rural and conservative states of West Malaysia, and even more so in Sabah and Sarawak, which are the poorest and least developed part of the country. As Table 2 shows, the strength of the opposition parties is concentrated in the states (and territory) which are most severely under-represented, while BN’s strength is concentrated in the states and territories which are most over-represented.

It will be seen that there is a clear correlation between over-representation and support for BN. The BN’s four strongest states – Sabah, Sarawak, Perlis and Pahang – are the four most over-represented. Of the PR’s four strongest states (or territory) – Palau Pinang, Kuala Lumpur, Selangor and Perak – one is severely under-represented (Selangor) and two are somewhat under-represented (Palau Pinang and Kuala Lumpur).

It should be noted that the malapportionment of seats among the states is not simply a case of favouring the less urbanised states against the more urbanised states, although this bias is undoubtedly present, particularly in the over-representation of Sabah, Sarawak, Perlis and Pahang and the under-representation of Selangor, Kuala Lumpur and Palau Pinang. It is striking that two of the least urbanised states, Kelantan and Terengganu, are also under-represented. These two states are the strongest areas for the Islamic Party of Malaysia (PAS), which is part of the PR opposition coalition. So not only are the urban areas under-represented, so are rural areas with a tendency not to vote for the government parties.

The second level at which Malaysia’s constituencies are malapportioned is within the states, particularly the more urbanised ones. Here the malapportionment operates against urban areas and in favour of rural areas. This operates in every state except Perlis and Terengganu, which have no significant urban areas, and Palau Pinang, which is mostly urban. In every other state, the largest seat is more than twice the size of the smallest. The disparity is particularly striking in Sarawak and Selangor. Table 3 shows the states ranked by the disparity between their largest constituency and their smallest.

It will be seen that the constituencies fall into four groups.

* 75 constituencies with enrolments above 70,000. Of these, the government won 15 (20%), while the opposition won 60 (80%).
* 58 constituencies with enrolments between 50,000 and 70,000. Of these, the government won 33 (39.7%), while the opposition won 35 (60.3%).
* 59 constituencies with enrolments between 30,000 and 50,000. Of these, the government won 55 (93.3%), while the opposition won 4 (6.7%).
* 30 constituencies with enrolments below 30,000. The government won all of them.

To put it another way, 60 of the opposition’s 89 seats have more than 70,000 voters, while only 15 of the government’s 133 do. Conversely, 85 of the government’s seats have fewer than 50,000 voters, while only 4 of the opposition’s do.

There is, however, a third level of the malapportionment of Malaysia’s constituencies. That is the malapportionment by race. The key fact about the Malaysian electoral system is that it is designed to preserve the power of the Malay Muslim population over all other racial and religious groups, and within that population, to ensure the dominance of the main Malay party, UMNO. Since only 54% of the population are Malay Muslims, and since not all of them vote for UMNO, this requires rigging the electoral system to ensure UMNO’s continued dominance. UMNO supremacy is also safeguarded by an alliance with small parties representing the Chinese and Indian communities (MCA and MIC respectively) in the National Front (BN) coalition.

During the colonial period, large numbers of Chinese and Indians settled in Malaysia, and for a time it appeared that they would come to outnumber of indigenous population. Since independence, however, emigration and the Chinese community’s low birthrate has reduced the proportion of Chinese from 45% in 1957 to 25% today. The Indian community is 7% of the population. The remaining 65% are classed as “sons of the soil” (Bumiputera), but this includes ethnic groups in Sabah and Sarawak who do not identify as Malay. These groups are 13% of the national population, and half of them are Christians.

The Chinese are heavily concentrated in urban areas, which is why most of Malaysia’s cities have large Chinese populations. George Town (Penang) is 66% Chinese, Ipoh 44%, Kuala Lumpur 43%, Kota Kinabalu 39%, Johor Bharu and Kuching 37%, Petaling Jaya and Melaka 35%. On the other hand the population of most of the rural areas of West Malaysia is almost entirely Malay Muslim. In East Malaysia most of the rural population are non-Malay Bumiputera.

This means that the malapportionment of constituencies away from urban areas and towards rural areas in both West and East Malaysia has the effect of corralling much of the Chinese population into a small number of large constituencies, either Chinese-majority, or close to majority. Thus, although the Chinese are 25% of the population, only 30 seats out of 222 (13.5%) have Chinese majorities, while another 23 seats (10.4%) are between 40 and 50% Chinese. The Chinese-majority seats have an average enrolment of 69,646, and the 40-50% Chinese seats have an average enrolment of 82,657. All the other 169 seats have an average enrolment of 55,030. (The 40-50% Chinese seats have larger average enrolments than the majority Chinese seats because most of them are in fast-growing outer-suburban areas, while the Chinese-majority seats are mostly in older city centre areas with stable populations.)

For many years the representation of the Chinese community was divided between the Malaysian Chinese Association (MCA), which is part of the ruling BN coalition, and the Democratic Action Party (DAP), which was the main opposition party. Over the past two elections, however, the great majority of the Chinese have gone over to the DAP. At this year’s election, the opposition won all 30 Chinese-majority seats, 29 going to the DAP and one to Anwar Ibrahim’s People’s Justice Party (PKR). Of the 40-50% Chinese seats, DAP won 9, PKR 7, MCA 3, UMNO 3 and the Malaysian Indian Congress (MIC) 1.

The corralling of the Chinese into large urban constituencies leaves a large number of constituencies in West Malaysia with small enrolments and large Malay majorities. There are 26 seats which are more than 90% Malay, with an average enrolment of 61,780. Of these UMNO won 15, and the Islamic Party of Malaysia (PAS), part of the opposition coalition, won 11. Strikingly, these 11 PAS seats have an average enrolment of 74,494, while the 15 UMNO seats have an average enrolment of 52,458. This reflects the deliberate under-representation of Kelantan and Terengganu, states dominated by rural Malay Muslims, but also the long-time strongholds of the PAS.

There are 32 seats in West Malaysia where Malay voters are between 75 and 90% of the total. Of these, UMNO won 28, the PAS 3 and the PKR 1. It is notable that UMNO’s position is stronger in these seats than in the seats which are more than 90% Malay. That is because none of them is in Kelantan and only one is in Terengganu, the two states where the PAS is strongest. The rest are in other rural areas of West Malaysia, where the great majority of Malays vote for UMNO.

There are 56 seats in West Malaysia where Malay voters are between 50 and 75% of the total. Of these, UMNO won 28 (exactly half), and its allied parties won 6 (MCA 4, MIC 1, Gerikan 1). For the opposition, the PKR won 16 and the PAS 6. These are ethnically mixed seats which the PKR can win by winning votes from all ethnic communities. It is striking that the DAP did not win a single Malay-majority constituency. It is also striking that four of the MCA’s seven MPs were elected by Malay-majority constituencies: in other words, by courtesy of UMNO.

Finally there are 51 seats in West Malaysia where Malay voters are a minority. Of these, UMNO won only two (both with Malay enrolment over 48%). UMNO’s allies won 6 (MAC 3, MIC 3). The remaining 43 went to the opposition: 31 to the DAP, 11 to the PKR and 1 to the PAS.

In East Malaysia, there are 8 urban seats with a majority of Chinese voters: all of them went to the opposition (DAP 7, PKR 1). All but one the other seats went either to UMNO or to its local allies. The PKR won one seat in Sabah with a Bumiputera majority. The government thus held all but one of the 56 East Malaysian seats which have a joint Malay-Bumiputera majority. This alone went a long way to giving the BN its majority, despite winning a minority of the national vote.

Finally, let us see what the result of the 2013 Malaysian election would have been if (a) seats had been distributed among the states and territories in proportion to their population, and (b) parties had won seats in each state and territory in proportion to the number of votes they received.

1 seat shortfall in Kedah not accounted for. Hence the discrepancy in PR 112 and BN 109, Total 211 instead of 212.

In Sabah and Sarawak the UMNO totals include local BN allies. In West Malaysia the UMNO totals includes Gerakan.

Thus the opposition’s 50.9% of the vote would have translated into a small majority, 112 seats to the BN’s 109.

Note: There are inevitably some distortions involved in transferring votes from a single-member constituency system. For example, although Chinese voters are about 25% of the total, the two Chinese parties, the MCA and the DAP, are shown as winning 70 seats (32% of the total) between them. This is because, as noted above, the BN ran MCA candidates in some Malay-majority seats, forcing Malay voters to vote for Chinese candidates. In a real PR election, these voters would have voted for UMNO, so the MCA would have won fewer seats. Conversely, in Chinese-majority seats, many Malay voters voted for the DAP candidate when they would probably have voted for a PKR candidate if given the choice. In a real PR election, the PKR would have won more seats and the DAP fewer. Also, the MIC is shown as winning only 7 seats, although if all Indian voters voted for the MIC it would have won about 16 (7% of the total). In fact many Indian voters vote for the DAP, but the MIC would still probably win more than 7 seats.

It might also be noted that if a Malaysian election did produce this result, UMNO and the PAS would win a majority (116 seats) between them. So it would be possible for UMNO to form an all-Malay government with the PAS and leave all the ethnic minority parties in opposition.