The 20th-century political philosopher John Rawls coined the term “veil of ignorance”: a thought experiment used for reasoning about the principles to structure a society based on mutual dependence. In that thought experiment, you are asked to consider which principles you would select for society, but you must select as if you had no knowledge ahead of time what position you would end up having in that society. Almost as if you are not born yet. So the choice is made from behind a “veil of ignorance”, which prevents you from knowing your ethnicity, social status, gender, sexual orientation, etc. This forces you to select principles impartially and rationally.

Political parties need to do that thought experiment when debating how we need to approach coalitions. It then goes like this: consider which principles you would select to govern the issue of coalitions, but select as if you had no knowledge ahead of time what position you would end up having in or outside a coalition. You don’t know how many seats you will have in a council or legislature. It may be one, it may be 20 or may be 100. Or it may be none at all. 

Behind that “veil of ignorance”, let’s discuss coalitions. We have now accepted them as a possibility. Our Constitution insists on an indirectly elected mayor, premier and president, and our electoral systems are based on proportionality. The combination of those two features means that coalitions are a possibility.

Read more in Daily Maverick: Political parties divided as ANC, DA push alternative to proportional representation system

Coalitions have the potential to bring communities together. A coalition government of multiple political parties will represent different political ideologies, cultures and ideas. This can help to bring communities together, because they see their representatives working together. 

Another advantage is that a coalition government is forced to adopt compromise-based policies. These policies may be smarter, and last longer than policies adopted by one majority party, which may be abolished as soon as that party loses power. 

However, in practice, coalitions are known mostly for their instability in local government. Conflict emerges in the coalition, motions of no confidence flow thick and fast, and leadership vacuums emerge. As a result, decisions are postponed, municipal administrations are thrown into paralysis and the council becomes preoccupied with politics instead of governance and service delivery, with devastating consequences for the dignity of residents and communities.

Capitalising on benefits

So, how do we capitalise on the benefits of coalitions, yet minimise the disadvantages? 

In my view, we need (1) A limited number of legal measures and (2) To adopt a number of general principles or, as I prefer to call them, conventions. 

When we talk about regulation we need to distinguish between three things:

First, we must reconsider some of the legal features that make us prone to political instability. We need to debate electoral thresholds: do we cherish the wide diversity of voices in the legislature or do we limit fragmentation? Is the balance struck right at the moment?

Read more in Daily Maverick: We should investigate potential legislation to manage conflict within coalition governments

Should we limit motions of no confidence, be it in the form of limits to the number one member may raise, to the periods in which they can be tabled, or the requirements for a motion to be valid?

Is it wise to elect office bearers in secret? Does that not push coalition negotiations even further away from the public and into the proverbial smoke-filled rooms where nefarious deals are struck?

Political parties need to do the ‘veil of ignorance’ thought experiment when debating how we need to approach coalitions. (Photo: iStock)

We need to allow time for coalition negotiations to take place before office bearers must be elected to avoid the tail wagging the dog.

Second, we need to discuss whether there is space to regulate the making and management of coalitions. This is more complicated and we need to be reluctant because this amounts to regulating political behaviour.

Should we insist on coalition agreements that are formalised, and made public? Should we go even further and insist on certain minimum requirements for coalition agreements?

But do we want to regulate the distribution of political offices according to proportionality? We already do so for those municipalities that operate according to the executive committee system.

Can we make the process of negotiations more predictable? At the moment it is a Mexican standoff in which it is not clear who makes the first move. What about the principle that the largest party should make the first move?

Should we consider limits to the number of attempts a council can make to form a coalition or elect a mayor? Should there be a mechanism to dissolve the council after the umpteenth attempt? Or should there be default positions? 

For example, in Spain, if a candidate does not have a majority in the first round of voting for a mayor in the council, the leader of the biggest party automatically becomes the mayor. Perhaps we would argue that that is undemocratic. But with us, the law says that we must try over and over again and, if need be, it is decided by lot. 

Should we insist on coalition agreements that are formalised, and made public? Should we go even further and insist on certain minimum requirements for coalition agreements? And can we actually enforce those?

It has been argued that we need forms of external support or oversight for municipalities with coalitions: what should the roles be of our provincial MECs overseeing local governments in their provinces, and is there a role for Chapter 9 institutions such as the IEC and the Auditor-General?

Third, we need to examine ways to insulate the administration from coalition turmoil. Perhaps we focus too much on wanting to reduce the scope for changes in government, and too little on how to ensure that when those changes happen, they don’t paralyse the administration. The most important thing here is to have an administration that is resilient during a turnover, to use a rugby term.

If the time for coalition negotiations is extended, can we regulate some sort of holding pattern in which those decisions that are critical to keep the wheels of government turning, are taken in an apolitical fashion? If so, what would that look like?

How do we avoid administrators and accounting officers being involved, willingly or unwillingly in the politics of coalitions? For example, is it wise to give the accounting officer the duty to preside over the election of a speaker, if that position may make all the difference in a council that’s on a knife edge? Or give him or her the duty to declare vacancies on the council even when that council is on a knife edge?

How do we discourage a coalition change resulting in a “purge”: a very high turnover of senior managers, a sudden spike in disciplinary charges, etc? Can we institute measures to stop this from happening?

General principles

Much more important than legal changes is the adoption of general principles or conventions. They don’t need to be laid down in law. But these are the conventions that we agree on, behind a veil of ignorance, as principles that should guide the formation of coalitions.

For example, should we not agree that coalition negotiations must deal with policies and programmes first and positions later? First try to find common ground on what you want to do once you’re in government, and only then on who will occupy which position.

Should we not adopt the principle that any proposal or counterproposal that attempts to distribute departments, posts or tenders will not be entertained?

Should we not confirm that each coalition partner matters, regardless of their size, and each partner must be visible in the coalition and be allocated meaningful roles? 

That brings us to the controversial issue of proportionality and positions. Should the mayor, premier or president come from the largest party in the coalition? Or is it okay that the amount of power you have in a coalition is inversely correlated to your electoral support? If we do agree on this, how dogmatic do we want to be about it? Are exceptions possible, such as when a smaller actor can clearly rise above the parties?

Read more in Daily Maverick: Coalition country

Last, we need to learn from our own experience. We want to be ready for it, when coalitions come to national or provincial government. 

I am not making any predictions, but the experience we have in local government must assist us. However, local government is different from national and provincial governments. The constitutional configuration is different. 

Municipal councils must take executive and administrative decisions. They appoint managers, release land, approve projects and policies, etc. Parliament and provincial legislatures don’t make those types of decisions. They focus on passing laws and exercising oversight. Also, the president, once elected by a majority of the National Assembly, is not only the head of government but also the head of state: that complexity doesn’t apply to a mayor. 

And the electoral system is different: there are no constituency representatives or by-elections in the national and provincial spheres. This means the practical and legal solutions may have to be different. However, the overall principles and conventions that we will emerge with at the dialogue are equally applicable. 

Let’s debate these, and other proposals, objectively and rationally, behind that “veil of ignorance” and for the benefit of the people of South Africa. DM

This is an edited version of a contribution made at the National Dialogue on Coalition Government, convened by Deputy-President Paul Mashatile in collaboration with the Dullah Omar Institute (University of the Western Cape, 4 to 5 August 2023).

Jaap de Visser is the Interim Research Chair in Multilevel Government, Dullah Omar Institute, University of the Western Cape.