IN DEFENCE OF PROPORTIONAL REPRESENTATION

This brief looks at the electoral task team report and its relevance with regards to the outcomes of the latest local government elections.

Sep 20, 2016

The ANC says it’s always called for a review of the proportional representation system and is not just doing so now because of its decline in support.

-Eyewitness News, reporting comments by Gwede Mantashe, 15 August 2016

The ANC has been in power for twenty-two years, but the government has only ever commissioned one study of our electoral system. Cabinet resolved on 20 March 2002 that an Electoral Task Team should be established to draft the new electoral legislation required by the Constitution. It should “formulate the parameters of new electoral legislation and draft it in order to prepare for the scheduled National and Provincial elections of 2004 or any earlier election, should the need arise, and include political parties in its consultations with stakeholders. This Task Team was chaired by Dr F van Zyl Slabbert.  

The Task Team reported in January 2003. It presented a majority and a minority report, each with its own draft legislation. From that day to this, no government action has been taken in response to the report. So yes, the ANC called for a review more than a decade ago, but the intensity of its desire to deal with the issue through the government has been, shall we say, variable.

Section 46(1)(d) of the Constitution specifies that the electoral system for the National Assembly must result, in general, in proportional representation. This section can be amended or deleted through processes prescribed for the amendment of the Constitution but, while it remains, ordinary legislation dealing with the electoral system must respect it.

One alternative system for national elections, to the one we currently have, could simply be modelled on the current local government electoral system. Half the councillors in a municipality are elected on a first past the post basis in individual wards. The other half are allocated on the basis of closed party lists, the allocation of seats for all parties being determined by the proportion of the votes going to each party, after seats have been allocated to independent candidates who have sufficient votes to be represented .This arrangement combines the benefits of accountability among the ward representatives with representation of all councillors reflecting the proportion of votes cast.  Applied to the national system, it would entail the division of the country into 200 constituencies. It would clearly meet the requirements of Section 46(1)(d).

But it is not the only system available for national elections and, in fact, the Electoral Task Team’s majority report proposed something different. The majority report recommended that 300 seats in the National Assembly be elected in large constituencies which would return between three and seven representatives each, depending on the number of registered voters in each constituency. The remaining 100 seats would be allocated on the basis of closed party lists in such a way that the total representation of each party in the National Assembly would be proportional to the number of votes received.  This proposal differs from the local government system applied nationally in three respects:

1. The percentage of members of the National Assembly elected by constituencies would be 75% instead of 50%. The Task Team argued that 100 closed list members would be sufficient to allow for the proportional representation principle to be applied fully.

2. Each constituency would have its own electoral quota and all candidates exceeding it would be elected, even if they were from different parties. For instance, a six member constituency might send three ANC, two DA and one EFF representative to the National Assembly. Each would be directly dependent on the constituency for continued membership in the National Assembly. When it comes to dealing with ‘your MP’, it would be natural for an individual voter to deal with the MP (or one of them) most closely aligned with that voter’s interest.

3. Initially, within each multi-member constituency, parties would put up ä closed list, with a number of candidates not exceeding the number of MPs to be elected in that constituency. Over time, the closed lists could become open lists, presumably through a process akin to a US presidential primary.

That, in a nutshell, was the Electoral Task Team’s majority proposal.  It would satisfy the requirement of Section 46(1)(d). The report has been gathering dust for thirteen years, but the dust could be blown off.  Some of the details of the justification, and possibly a few aspects of the draft legislation, are dated, but the main arguments are as valid now as they were then.

Although proportional representation is protected by the Constitution, it remains in need of justification.  The details of the justification will vary over time. Once neo-apartheid consociationalism had been rejected as a political principle in the constitutional negotiations, the problem emerged of the representation of minorities in the National Assembly.

Given the expectation that race would be the principal definer of political loyalties in the first years of democracy, a first past the post system could be expected to close out minorities unduly from representation, with their resulting alienation from the political system. This was a valid concern and was recognised as such.  And race remains very important in determining political loyalties, but its impact is starting to soften. So the initial reason for proportional representation remains, but has become a little less compelling. Is there anything to replace it as time goes by?

There is a proposition in political science known as Duverger’s law.  It runs as follows:

A first past the post system structured within single-member districts tends to produce a two-party system, while proportional representation favours multiple parties.

This is a ‘tendency’ law rather than a rigidly determinist one.

In turn, multiple parties often produce coalition governments. Since 1945, West Germany and then a unified Germany has more often than not been ruled by a coalition government. Indeed even the conservative Christian Democratic Union was itself a coalition of democratic forces from the Weimar Republic, repressed during the twelve years of Nazi rule. Dutch and Italian governments are invariably coalitions, whose composition changes frequently. In none of the three cases, has coalition rule prevented economic and social development. Indeed, coalition rule has transmuted forces which led to street violence and dictatorship in the interwar years into peaceful political competition, an outcome which nearly all South Africans would applaud.

We have multiple parties represented in the National Assembly, as Duverger’s law predicts. But for reasons originating in our political history, we have had a single dominant party since 1994.  It is now less dominant than it was, and we are starting to see the emergence of coalition rule in local government. This trend may or may not continue, and it may or may not extend itself to national elections. It depends on the emerging preferences of the electorate. 

Alliances, coalitions, fusions and splits are the creative responses to political parties to changing circumstances.Proportional representation will align these developments more effectively with voter preferences than the abolition of Section 46(1)(d). Our political parties accept the outcome of free and fair elections. We have developed a working framework for co-operative governance.  While the details of the electoral system can change over time – and they should do so if change promotes accountability – Section 46(1)(d) was the right choice and it will remain so. We can work within its framework without the roof falling in.

Charles Simkins is Head of Research at the Helen Suzman Foundation.

This article first appeared as an HSF Brief.