In 2017 the Constitutional Court held in United Democratic Movement v Speaker of the National Assembly and Others that nothing prohibited the Speaker of the National Assembly from allowing a vote of no confidence in the president to be conducted by secret ballot.

The judgment, authored by then Chief Justice Mogoeng Mogoeng, was widely praised at the time as it would allow a vote of no confidence against then-president Jacob Zuma to be conducted by secret ballot, which would protect dissident ANC MPs who voted in favour of such a motion from censure by their party.

Because there is a natural tendency to evaluate a constitutional rule by focusing on its effect on the immediate problem at hand and to assume that it will have the same effect in all other situations, there is always a risk that a rule adopted to fix one problem will cause other problems and make things worse in the long run.

The Constitutional Court acknowledged this problem in the UDM judgment, noting that while secret ballot votes of no confidence might sometimes enhance accountability, they could also have disastrous consequences as they could result in the removal or election of a president, premiers or mayors through “undeserved majorities” secured by “crass dishonesty, in the form of bribe-taking or other illegitimate methods”.

The Constitutional Court did not show the same caution in the Nkandla judgment in which it held that the remedial action imposed by the Public Protector was binding. Its ruling in that case “fixed” the problem of Mr Jacob Zuma ignoring the remedial action imposed by then Public Protector Thuli Madonsela, but did so by arguably creating an even bigger problem.

It did so because it wrongly assumed that future Public Protectors would continue to be honest, competent and fair and that they would not abuse the far-reaching powers bestowed on them by the Nkandla judgment.

It is with this in mind that I expressed some scepticism at the national dialogue on coalition governments about the wisdom of adopting new rules aimed at regulating coalition governments. There is a risk that even good-faith policymakers will look at what happened in Johannesburg and some other problematic hung councils recently and craft regulations to “fix” the problems they have identified in these municipalities, without considering that such regulations may do more harm than good in municipalities where the political dynamics are different.

Moreover, even thoughtful and well-crafted regulations may not stabilise coalition governments as the stability of coalition governments largely depends on the quality of the politicians and political parties involved, and on other socio-political factors — something that legal regulations can do little to change.

Despite these misgivings, I do believe that there is one immediate change to the rules regulating the formation and dissolution of coalition governments worth pursuing. This is to adopt a rule (by amending the relevant provisions of the Constitution and the Structures Act) to require an open ballot for the election or removal of the speaker of a legislative body, as well for the election or removal of the president, premiers and mayors. (Currently, a secret ballot vote is legally required for the election of any of these positions, while a secret ballot is permitted — at the discretion of the Speaker — in votes of no confidence to remove any of these office bearers.)

Requiring open-ballot votes for the election or removal of these office bearers would limit the phenomenon of vote buying and would make such votes more predictable. It would enhance trust between political parties involved in negotiations about the composition of coalition governments in hung legislative bodies, as it would help to ensure that such agreements are not derailed by rogue MPs or councillors who do not toe the party line in secret votes.

While conducting votes of no confidence by secret ballot may, on rare occasions, enhance accountability (as it did in the Zuma case), such situations would arise rarely. In almost all cases, such votes have nothing to do with holding office bearers accountable, and everything to do with gaining or losing government power.

In any event, secret ballot votes may have a negative impact on the quality of a democracy, as such secrecy shields political parties and their elected representatives from being held accountable by those who should count most in a democracy — the voters. When such votes are conducted in secret, it robs voters of vital information that might influence their voting choices in future.

Another proposal — to limit the tabling of votes of no confidence to one every 12 months (or every 24 months) — seemed to have enjoyed widespread (but not universal) support among political parties represented at the dialogue.

At first glance, this seems like a common-sense proposal that will reduce coalition government instability by preventing legislative bodies from ousting the president, premier, mayor, or speaker every few months through votes of no confidence. While amendment of the Constitution would be required to implement this rule at national and provincial level, a mere amendment of the Structures Act would be required to implement the rule at the local government level.

It is important to understand that such a rule will reduce the power of the relevant legislative body vis-à-vis the executive by allowing a government to continue in office for a defined period even when that government no longer enjoys the confidence of the legislative body. One must therefore ask whether there is a likelihood that such a rule will be abused and that it may have unforeseen negative consequences in the future.

For example, in a hung council in an executive mayoral system, the rule may heighten contestation around the election of the mayor. Will this encourage political parties to reach agreement on the formation of a coalition government, or will it have the opposite effect? Will it be possible for the political parties who form part of a coalition government to “cheat” the system by engineering the tabling of a vote of no confidence which they know will not succeed, thus protecting themselves and securing their position in government for another 12 or 24 months?

And what will happen if the rule props up a government despite the complete unravelling of the coalition which ensured the election of a mayor?

These (and other, as yet unanticipated) problems may never arise, and the rule may well do more good than harm. But who could say for sure?

Another proposal that seemed to enjoy significant support from the representatives of political parties at the national dialogue is for the establishment of an independent body to oversee the management of coalition agreements and to resolve disputes between coalition partners in accordance with the terms of coalition agreements. The proposal envisages that coalition partners be required to deposit or register their coalition agreements with this independent body, and that such agreements would be made public.

There are good reasons to encourage political parties who form coalition governments to conclude detailed coalition agreements. Many delegates at the national dialogue bemoaned the fact that coalition negotiations between political parties in hung councils focus primarily (and sometimes exclusively) on the number of mayoral committee positions to be allocated to each coalition partner, and on the specific portfolios each party will be entitled to manage. How these coalition governments will govern and what policies they will implement are seldom discussed.

The proposal is partly aimed at addressing this problem. But it is unclear whether the implementation of this proposal will make much of a difference.

It is important to understand that the Constitution and – as far as the executive mayoral system is concerned – the Structures Act are silent on the formation of a majority or coalition government. Legally the focus is exclusively on the election of the president, premiers, and executive mayors, who then enjoy an absolute right in law to select who serve in their governments. (This power is informally limited by what parties have agreed to and by other political considerations.)

When MPs, MPLs and councillors in hung legislative bodies elect the president, premiers or executive mayors, they do not vote in support of a specific coalition government; they only vote to elect the person who will head that government.

This means a person can be elected executive mayor in a hung council without there being any coalition agreement in place, which sometimes leads to the formation of minority governments which remain in power only because one or more political party which is not part of the coalition refrains from supporting a vote of no confidence in the executive mayor.

Political parties which hold the balance of power in a hung council often do this to keep one or more other parties out of government. For example, after the previous local government election, the EFF supported the election of DA mayors in Tshwane and Johannesburg to keep the ANC out of office without joining the government. The DA governed as a minority government, and survived only for as long as the EFF wanted it to.

While the proposal will force political parties which do reach a coalition agreement to make the terms of this agreement public, and may arguably also encourage parties negotiating with each other to focus less on the allocation of positions and more on agreed policies of the coalition, it cannot force parties to conclude a formal coalition agreement as this is not required for the election of an executive mayor.

It will therefore not prevent parties from concluding a backroom deal to elect a councillor from a tiny party as executive  mayor, with an informal agreement on how that mayor would allocate positions in the government. It will also not prevent political parties from hiding unsavoury aspects of the agreement, such as the allocation of positions or tenders within the administration.

Coalition agreements are in essence political agreements that will only stick when there is a degree of trust between the parties, which means that they cannot (and should not) be made enforceable by a court or any other body. While the proposed independent body may facilitate dialogue between fighting coalition parties, this may not be of much help to stabilise coalitions.

What is required to stabilise coalitions is for politicians and political parties to become more honest and principled: to become less concerned about access to positions and resources, and more concerned about the well-being of the voters they serve.

Only voters can nudge politicians and political parties in this direction by punishing parties for their irresponsible and selfish behaviour at the ballot box.

A far more controversial proposal, fiercely resisted by smaller opposition parties, is the introduction of a modest electoral threshold which would reduce the number of political parties represented in legislative bodies, thus limiting political party fragmentation and making it easier for the smaller number of parties to reach agreement on the formation of a coalition government.

I have previously explained why this is a proposal worth considering, despite its obvious drawbacks, and will not rehash those arguments again.

But smaller parties may not have anything to worry about on this score in the foreseeable future, as electoral thresholds might well be in breach of the constitutional requirement that the electoral system at all spheres of government must result “in general, in proportional representation”.

The Constitutional Court has not yet given a clear indication what this may mean, but because the introduction of an electoral threshold would exclude a large number of smaller parties from representation in the various legislative bodies, the Constitutional Court may well find that it does not meet the requirement that the electoral system in general results in proportional representation.

(The amendments to the electoral system allowing independent candidates to stand for election may well run into similar difficulties as it imposes a threshold on independent candidates, requiring them to win more or less double the number of votes required by political parties in order to get elected to the NA.)

Leaving aside all the technicalities, the stark truth is that politicians and their parties are to blame for the instability of some coalition governments. Legislation – no matter how well-meaning – will not change any of this.

Politicians and their parties will only behave better if they believe this is in their own interest, which will happen when voters punish these politicians and their parties for behaving badly. DM