GABRIEL MAKIN: Proposals for stabilising future coalition governments

The collapse of the coalition-led government in Johannesburg last week, marked by the successful motion of no confidence against DA mayor Mpho Phalatse, will raise fresh doubts about the prospects of success for coalition governments broadly in the country. This is especially relevant with the 2024 national election now just more than 12 months away. 

According to our polls, next year’s election could see the ANC’s share of the national vote slip below 50% for the first time. Our most in-depth poll data showed that in a turnout scenario on a par with that of the 2019 national election support for the ANC might have slipped to 49% had an election been held towards the middle of 2022. It also showed that in Gauteng and KwaZulu-Natal the ANC stood to lose its majorities, though in the latter the election would be exceptionally close-fought. The data further showed that a significant proportion of voters welcomed the idea of a future coalition government and that the prospect of a deal between the ANC and the DA was particularly popular.

Those polls join a growing body of evidence that SA is well advanced in a process of its one-party-dominant political configuration being supplanted by a future of coalition governments. This was to be expected, as it is the natural consequence of the country’s proportional representation electoral system. Parties are represented in legislatures based on the proportion of votes they receive, as opposed to constituency systems, in which the party with the most votes in a particular voting district occupies the relevant seat in a legislature.

With the past year’s experience of coalition government under the belt, and after consulting a range of leading experts on the subject, the Social Research Foundation has produced a short report on measures SA legislators might consider to strengthen the stability of future experiments in coalition government. Four such measures are proposed.

The first is for the introduction of electoral thresholds. These are commonly deployed in proportional representation electoral systems to ensure that political parties meet a minimum level of support to enter a legislature. SA has no explicit threshold requirement, but there is an implicit threshold because parties that obtain too few votes to win a single seat are excluded from legislatures. At national level, parties generally need 40,000 to 45,000 votes for a seat in parliament, though small parties can get a seat with just 30,000-35,000 votes because of the proportional representation formula used to calculate seats.

The absence of a threshold allows microparties with extremely low levels of support to sometimes exert outsize sway as kingmakers in government, thereby exerting power over national politics and policy on a scale comparable to that of far larger parties with considerable voter support.

The level at which a threshold should be set is important. A 5% threshold is at the upper end of international practice. This is the case in New Zealand and Germany. In Israel the figure is closer to 3%. Set at a level of 5% an SA threshold would exclude a broad cross-section of parties, including the African Christian Democratic Party, the IFP and the Freedom Front Plus, which have become fixtures of the political landscape since 1994 and serve clear constituencies. A more realistic level might therefore be closer to 1%.

Where smaller parties resist the threshold on the basis that it is undemocratic, the counter should be put to them that what happens in other jurisdictions applying thresholds is that the microparties may band together and form joint lists to meet the threshold.

The second proposal is for binding coalition agreements. Coalition agreements should be legally binding for the period of office of the coalition government, and treated in the same vein as a commercial contract, which can only be terminated under specific circumstances and then only with the application of financial penalties or the approval of previously agreed mediators or courts.

A third proposal is to establish a coalition ombudsman. Disputes will arise even where binding contracts are in place, and to avoid lengthy court battles the office of an ombudsman for coalitions could be used. When a coalition partner feels sufficiently aggrieved to consider breaking a coalition agreement the ombudsman would decide whether the reasons given for termination are justified or not.

A further proposal is to limit the frequency and number of motions of no confidence that may be brought against a sitting government. This will promote government stability and ensure that changes of administration take place chiefly via scheduled elections.

Academic writing on coalitions is largely unanimous that they are most effective where a culture exists — something close to a moral code — of holding coalition agreements together in the best interest of a society. In the absence of such an established culture the building of formal enforcement mechanisms may go some way towards building such a culture in the longer term.

We believe should SA’s legislators choose to consider these proposals they will go some way towards shoring up the stability of future governments and thereby sustaining public confidence in democracy.

• Makin is an associate at public policy think-tank the Social Research Foundation.