Memorandum submitted to the Parliamentary Select Committee on Electoral Reform
December 15, 2011

Reducing Unfairness in the Delineation of Electoral Constituencies

Lim Hong Hai, Universiti Sains Malaysia

I. Introduction

1. It is useful to begin by stating three main conditions for elections to be fair. First, the franchise or qualifications of voters must be accepted as fair and those qualified must be accurately included in the electoral roll. Second, the competition for votes, i.e. rules and practices on campaigning, must be fair. Third, votes won by political parties must be fairly translated into seats won. Although less widely understood, the third condition is necessary because elections are contest for seats and not just for votes.

2. Various matters that affect the above three conditions have been raised for the attention of the Committee. I shall focus on the third condition, i.e. the fair translation of votes into seats. I believe that the deficit in meeting this condition has caused the most unfairness in Malaysian elections thus far. I shall identify the main reasons or sources of this deficit and suggest measures for addressing them.

II. The FPTP (first-past-the-post) electoral system: Is it a problem?

1. The translation of votes won into seats won by various political parties importantly depends on the electoral system that is adopted. It is an inherent feature of the FPTP electoral system that we also use that the translation of votes into seats is not proportional to votes: the FPTP system always awards a more than proportional share of seats to the largest party (i.e. which won the most votes) and hence less than proportional shares of seats to other parties.

2. However, this non-proportionality has not been seriously or at all widely questioned in Malaysia. This is one reason for retaining the FPTP electoral system and not replacing it with some system of PR (proportional representation). Two other reasons may be added.
First, the non-proportionality increases the legislative majority of the governing party and thus makes for strong and stable government, which most Malaysians seem to value. Second, the operation of the FPTP system has another desirable effect in the Malaysian context. Candidates in the single-member constituencies usually have to win – or pool – votes from more than one ethnic group in order to ensure victory: this encourages moderation on ethnic issues by candidates.

3. If the FPTP electoral system is not the problem, what is? The short answer is this: It is the way electoral constituencies have been delineated. I now proceed to elaborate.
III. Constituency delineation: The abuses of mal-apportionment and gerrymandering

1. The delineation of constituencies has long been considered unfair because of two practices, namely mal-apportionment and gerrymandering. Both these practices, which can have important effects under the FPTP electoral system, have been systematically exploited for partisan advantage by the governing party. Mal-apportionment and gerrymandering are therefore the main sources of unfairness in the translation of votes into seats and constitute major “abuses” of the electoral system.

2. When appearing before the Committee in Sabah, Mr. Ng Chak Ngoon has presented a graph that shows the unequal numbers of electors among electoral constituencies. Inequality among constituency electorates is mal-apportionment (rather than gerrymandering, as reported in the media). Mal-apportionment means that different numbers of electors elect a representative in different constituencies. The value of each vote is therefore not the same in different constituencies. It is unfair to individual voters in the larger constituencies. It is also unfair as among political parties: as also shown by Mr. Ng, mal-apportionment is extremely large as well as extremely systematic in its effects: non-BN parties have to win a lot more votes to get one seat compared to BN.

3. Gerrymandering is the drawing of constituency boundaries for partisan advantage. It can be done even in the absence of mal-apportionment. Gerrymandering uses two methods – essentially to “waste” opposition votes. The first is to draw constituency boundaries to concentrate or pack opposition voters into a limited number of constituencies. Opposition parties will win these constituencies with large margins or much more votes than needed for victory: these “excess” votes are wasted and not available to help secure victory in other constituencies. The second is to disperse opposition voters among constituencies where they form significant numbers but still a “safe” minority that would be outnumbered by voters who support the governing party. These outnumbered opposition votes are also wasted as they do not translate into any seats for opposition parties.

4. Mal-apportionment has sharply increased in the first constituency revision in 1974 and maintained at high levels in subsequent revisions. Gerrymandering is much harder to nail down. Still, various scholars have also noted signs of gerrymandering beginning with the first revision in 1974 and especially in the last revision in 2003, when it was directed largely against PAS. Mal-apportionment remains the more important source of unfairness in constituency delineation and it has been directed largely against urban areas, in effect largely against non-Bumiputra voters and the parties they support.

5. This is not the time and place to explain how all this has come about. More important is to address the following question: What can – and should – be done to reduce mal-apportionment and gerrymandering and thus ensure fairness in constituency delineation? Taking (i.e. without fully explaining) the present situation as the point of departure, I provide five proposals (or sets of proposals) below. As only to be expected, these proposals concern the issue of who should do what and how in constituency delineation.
IV. Proposals for enhancing fairness in constituency delineation

1. The first proposal is to restore to the EC (Election Commission) the power to distribute or apportion the total number of parliamentary constituencies among the various states. This power was taken away from the EC and effectively vested in the governing party by constitutional amendment in 1973. [Since then, the total number of parliamentary constituencies as well as its apportionment among states has been specified in the Federal Constitution, i.e. Art. 46 on the composition of Parliament. Any change would require amending this Article – which the BN can do on its own before losing its two-thirds majority in the 2008 election.]

2. The second proposal is to constitutionally require the apportionment of parliamentary constituencies among states to be proportional to electorate. This proposal would prevent mal-apportionment of parliamentary constituencies among states. [This proposal and the next would essentially restore the original 1957 constitutional provisions that were repealed in 1962, i.e. Clauses 3, 4 and 5 of Art. 116. The Federal Constitution still retains the numbering of these Clauses followed by a single word, “Repealed”.]

3. The third proposal is to restore clear constitutional limits to rural weightage (i.e. “weightage for area” in favour of rural areas). These limits would apply to – and thus serve to limit mal-apportionment – in the delineation of both parliamentary and state constituencies within each state. Note that, for state constituencies, it is mal-apportionment within each state, rather than among states, that is material. Some background is useful before suggesting the exact limits.

(a) To safeguard the fundamental principle of approximately equal electorates among constituencies, differences in constituency electorates were limited to 15 percent above or below the average constituency electorate at the time of Merdeka. These clear numerical limits were relaxed in 1962 and then removed in 1973 by constitutional amendment: the Federal Constitution now allows, rather imprecisely, “a measure of weightage” in favour of rural constituencies. With clear limits removed, rural weightage has been liberally applied by the EC, even though the communication and other disadvantages of rural areas that form the justification for rural weightage have significantly declined since Merdeka. The EC’s application of rural weightage thus strongly appears less than principled. Indeed, constituency electorates are now so unequal as to raise the question as to whether the fundamental requirement of approximate equality is still being complied with.

(b) For Peninsular Malaysia, therefore, it is not unreasonable to restore the limits to mal-apportionment that existed at the time of Merdeka, i.e. 15 percent above or below the average constituency electorate. Conditions in Sabah and Sarawak would justify wider limits than those for Peninsular Malaysia. My suggestion is to restore the limits that were in force when Malaysia was formed. Those limits allow the largest constituency to have twice the number of electors as the smallest constituency. In effect, they limit inequality among constituencies to one-third or 33 percent above or below the average constituency electorate – which is still a considerable reduction of mal-apportionment from present levels in the two states.
4. The fourth proposal is generally to strengthen the independence of the EC. While important generally, an independent EC is particularly needed to prevent gerrymandering in constituency delineation. It is probably the best available safeguard against gerrymandering.

Strengthening the independence of the EC mainly requires improving the procedure for the appointment of its members. Something similar to what we have recently done with respect to senior judicial appointments is equally needed and useful for improving the appointment of EC members. A useful added measure is for the EC to make use of its present constitutional powers to appoint its senior administrative officials instead of relying completely on the federal civil service.

5. The fifth proposal is generally to improve the procedure for securing parliamentary approval of the revised constituencies delineated by the EC. This can be done by means of three specific changes to the present procedure.

(a) First, the EC should report its recommendations on revised constituencies directly to Parliament for approval. At present (or since the 1962 constitutional amendments), the EC reports to the Prime Minister, who can then make changes to the EC’s recommendations before tabling them in Parliament and also thereafter in order to secure parliamentary approval (see the 13th Schedule of the Federal Constitution). This procedure gives too much power to the Prime Minister and does not inspire confidence that constituencies have been or would be fairly delineated.

(b) Second, require more than the present simple majority (in effect, the support of the ruling party alone) in Parliament for the approval of revised constituencies. To provide two alternatives, require a two-thirds majority of Parliament or (even stronger) concurrent majorities by government and opposition members in Parliament.

(c) Third, if changes need to be made to the EC’s recommendations for securing parliamentary approval, then these changes should be made by the EC – after receiving and considering the views of members of Parliament.

V. Conclusion

Largely because of mal-apportionment, the practice of constituency delineation is the major cause of unfairness in our electoral system. It causes more unfairness than deficiencies in voter registration and campaign rules. Correcting unfairness in constituency delineation is of the highest priority in electoral reform. There are ways – and those I have proposed would go a long way – to ensure fairness in constituency delineation. What is needed and should be demonstrated now is only the will.

I repeat what I once wrote of EC leadership: For persons dedicated to public service, courage and perseverance can be derived in no small measure from the conviction of serving an important national need – and from awareness of their once-in-a-lifetime opportunity for making a meaningful contribution to it.