Follow us


Professor David Butler is Emeritus Professor of law at Stellenbosch University, South Africa, and was the main advisor to the South African Government on the International Arbitration Act with his work as part the South African Law Reform Commission's work reforming South Africa's arbitration legislation. Professor Butler is a life fellow of the Association of Arbitrators of Southern Africa, and has authored authoritative texts on the subject of arbitration, including his textbook Arbitration in South Africa Law and Practice. He has also previously been chairperson of the Department of Mercantile Law at Stellenbosch and has taught Company Law.

In 2017, South Africa introduced its new International Arbitration Act following a quarter of a century of discussion regarding the reform of South Africa’s legislative regime. In this issue of Inside Arbitration, Director, Jonathan Ripley-Evans interviews Professor David Butler, Emeritus Professor of Law at Stellenbosch University, and main advisor to the South African government on the 2017 International Arbitration Act (the “IA Act”). They discuss the background to this reform and the Act on developments in South African law, as well as the current attitude of South Africa towards investment arbitration. They look at recent court decisions and consider the future of international arbitration in South Africa.

Professor Butler, 2019 saw a lot of buzz around South Africa’s new IA Act. That revision has been a long time coming. You’ve been involved in discussions around the revision since they first began. Could you share with us some of the background?

The reform of South Africa’s international arbitration regime (a process which took over twenty years) was first envisaged in the South African Law Reform Commission (“SALRC”)’s Report of July 1998. The Commission started on this project in 1996.

In 1994 South Africa emerged from an era of isolation and was faced with the difficulty that many of its laws in the area of international trade and investment were outdated and inadequate. This was a serious issue in the context of foreign investment. The old Arbitration Act of 1965 was intended for domestic arbitration and contained no specific provisions for international arbitration. Nonetheless it applied to international arbitration seated in South Africa by default. The main recommendation in the 1998 report was that South Africa should adopt the UNCITRAL Model Law for international arbitration only. Domestic arbitration was the subject of a separate investigation and in 2001 the Commission recommended that South Africa should have a dual system with separate laws dealing with international and domestic arbitration. While we have a new act for international arbitration, new domestic legislation has yet to be enacted.

So how have things changed over this period? Does the IA Act resemble the drafts which were in existence in 1998?

Although the 1998 report was initially met with much enthusiasm, many hurdles (mainly political) hindered progress in the field and led to considerable delay in the IA Act being passed.

The IA Act changed considerably between the first draft and the version that became law in December 2017. There were two primary reasons for that. Firstly, UNCITRAL itself updated the Model Law in 2006 and these amendments had to be carefully considered, and where appropriate, adopted. Secondly, in the interests of harmonisation and so that the South African version of the Model Law would be familiar to foreign users, modifications to the UNCITRAL text were kept to those regarded as essential. The revision process officially commenced through the SALRC in 2013. It was decided that modifications to the UNCITRAL text should be kept to those reasonably necessary for the Model Law to work effectively in South Africa. This therefore included some minor technical adjustments in the light of experience in other Model Law jurisdictions.

If you were to identify the biggest changes brought about by the IA Act, what would these be?

Before the new IA Act, South Africa had no legislation specifically designed for an international arbitration with its seat in South Africa. The 1977 legislation that gave effect to South Africa’s accession to the New York Convention also had some serious technical defects. The new IA Act incorporates the UNCITRAL Model Law which, as we know, aims to create an internationally accepted standard upon which countries can model their local IA laws. As a result, South Africa’s arbitration laws are now aligned with international best practice.

In addition, the IA Act replaces the 1977 legislation on the New York Convention. Chapter 3 of the IA Act now properly recognises and gives effect to South Africa’s obligations under the New York Convention.

In your view, how does the new IA Act compare with similar legislation found in other jurisdictions?

The incorporation of the Model Law brings South Africa’s legislation on international arbitration in line with similar (Model Law-based) statutes in other jurisdictions. In the case of South Africa, although there are some South African-specific adjustments, the IA Act incorporates the version of the Model Law adopted by South Africa, in Schedule 1. The Law Reform Commission strongly favoured this approach, as opposed to the alternative of rewriting the UNCITRAL text in the Act itself.

In contrast, the majority of other Model Law jurisdictions have chosen to incorporate the principles of the Model Law in the body of their Acts, as opposed to putting the adapted text of the Model Law in a Schedule, as South Africa has done. The majority approach can and does cause interpretation difficulties, which South Africa will hopefully avoid.

The IA Act therefore brings the South African IA regime in line with internationally accepted standards. As mentioned, there are, however, certain country-specific adaptations which have been included. For example, section 11 mandates that arbitration proceedings to which a public body is a party, are to be held in public, unless for compelling reasons, the arbitral tribunal directs otherwise.

Given the fact that South Africa is clearly moving towards a more “global” approach to IA, do you have any insight into why SA took the decision to not accede to the Washington (or ICSID) Convention?

Initially, the South African Law Reform Commission’s report of 1998 also recommended that South Africa adopt the Washington Convention and subscribe to the International Centre for the Settlement of Investment Disputes. This recommendation has not been followed.

Many of the Bilateral Investment Treaties which were entered into by South Africa were concluded at the time of South Africa’s transition into a constitutional democracy, mainly in an effort to encourage foreign investment in the country. However, since 2001, a sharp rise in international investment disputes was noted, which caused South Africa to review its BITs.

As a result of this review, in 2013, South Africa’s Department of Trade and Industry reached the conclusion that it does not agree that there is a correlation between BITs and increased inflows of foreign investment. Additionally, South Africa’s current socio-economic policies do not necessarily correlate with the best interests of foreign investors. As such, this could present a fertile breeding ground for disputes under BITs. As a result, South Africa has reviewed and cancelled many of its existing BITs.

its Protection of Investment Act. Section 13(4) of that Act provides only for the international arbitration of investment disputes once all domestic remedies have been exhausted, and only with the consent of the government of South Africa. As a result, the South African government clearly regards international arbitration as a means for resolving investment disputes as an exception to the rule, and not as the preferred mechanism.

It seems fair to conclude that the Washington Convention is no longer a priority for South Africa.

Having had the benefit of almost 25 years of debate on the issue, one would expect that the local legal community would at least be familiar with the changes brought about by the IA Act. Recent debates, however, at a local level, seem to indicate that this is not actually the case. Is this something you have seen?

Yes, I suspect that this is mainly due to the fact that South African practitioners need to acquaint and familiarise themselves with the difference between how domestic litigation and international arbitrations are run. South African practitioners are often wedded to South African local practices, which are not necessarily suited to international arbitrations. It is common to see South African practitioners falling back on the safety net of the manner in which arbitrations were conducted under the previous, domestic focused, Arbitration Act.

As we know, the reality is that international arbitrations operate in a sphere outside of the ambit of the courts (by design). This is not something that comes naturally to the average South African-trained lawyer, where the four corners of the law, precedent and court rules reign supreme.

When conducting international arbitrations, South African practitioners must learn to separate governing law and seat and to not assume that domestic principles of law or procedure will apply. That may mean they have to detach themselves from their usual reliance on the legal system of judicial precedent, the habit of falling back on domestic case law and practice and the reliance on the rules of the South African domestic courts.

For many years South Africa implemented an extremely restrictive policy in relation to the jurisdiction of its courts. This meant that the local courts would only take jurisdiction over a dispute if there was the existence of a “jurisdictional link” to South Africa. Can you explain the justification for this approach?

Traditionally, this narrow approach of the South African courts meant steps had to be taken in order to “found” the jurisdiction in South Africa. For example, when one was dealing with a foreign defendant, an application for an attachment of assets had to be made in South Africa. Now, this would work if assets were available in the jurisdiction. But if no assets were attached, or no other link could be established, the court would then lack the jurisdiction to hear a matter. This applied across the board, including where an application was made to enforce a foreign arbitration award. The main justification for this approach was due to the fact that the courts wanted to ensure the effectiveness of their judgments – in other words, that there was some asset against which the judgment could be executed.

But this historic approach surely contradicts the international approach to arbitration and, more importantly, the requirements of the New York Convention?

Absolutely. International arbitration is designed to operate in an environment that is free from court interference. It is not for a domestic court to decide on the effectiveness of an award. As you say, South Africa has been a party to the New York Convention since 1977 and the court’s approach to the enforcement of awards was at odds with its obligations under that Convention.

The IA Act has sought to remedy this. The Convention has been annexed to the IA Act at Schedule 3, and its obligations under the Convention have been further clarified in Chapter 3 of the IA Act. South African courts are therefore obliged, under the international treaty (which has now been enacted by the IA Act), to recognise and enforce foreign arbitral awards. In addition, and importantly, you will note that the IA Act does not include a “lack of jurisdiction” as a ground upon which the court is entitled to refuse to enforce a foreign arbitral award.

As a result, a failure to recognise an award on grounds not stipulated in the New York Convention constitutes a breach of the country’s obligations.

Could you expand upon our courts’ recent confirmation of the development of the concept of jurisdiction, in light of the new IA Act?

This is probably best summarised by the following quote by Adams J, which is contained in the recent decision of Vedanta v KCM:

I think that the importance of the doctrine of ‘effectiveness’ has become somewhat eroded in the context of jurisdiction. This may very well relate to the fact that the world has become a global society in which the fact that an entity, which has agreed to HERBERT SMITH FREEHILLS SECTION TITLE 33 an international arbitration and the concomitant jurisdiction of the overseeing court, chooses not to be bound by an order of such a court, may be frowned upon by the international business community.

But the argument has been raised that this has created a lacuna in the law – specifically, there is a tension between the concept that South African courts are empowered by local law and local law requires that the court must hold jurisdiction to entertain a matter. On the other hand, international obligations impose a duty to recognise and enforce foreign arbitral awards, regardless of the nationality of the parties and the effectiveness of an award. How is this conflict then resolved?

Well the concern arises from an incorrect or perhaps incomplete understanding of the law. In South Africa, the jurisdiction of the local courts is derived from section 21 of the Superior Courts Act. It is true that matters arising within its area of jurisdiction are to be resolved before that court and historically, applications for the attachment of assets to found jurisdiction were required in order to found jurisdiction over matters not otherwise within the jurisdiction of the court.

But it is often forgotten that section 21 also grants jurisdiction to the court, over all other matters of which it may according to law take cognisance. The commentary to the Superior Courts Act makes it clear that jurisdiction is also to be determined with reference to relevant statutes.

The IA Act is indeed one statute that the court is entitled to rely upon in exercising its jurisdiction and as such, applications for the attachment of goods to found jurisdiction should be unnecessary when seeking to enforce a foreign arbitral award, under the IA Act.

But is this necessarily a principle that is being applied uniformly in practice? Take, for example, the case of Steyn v Tanzanian Government where an application to attach an aircraft owned by the Tanzanian Government was sought (and obtained) in order to found the jurisdiction of the court. In light of cases such as Steyn, can it be said that the principle is always adhered to?

The Steyn judgment is unfortunate in a number of respects, one of which is indeed the impression created that an attachment was required to found the jurisdiction of the court. But it would appear that the application for attachment was used as a tool to freeze the asset which was likely to be removed from South Africa in a short period of time. So, rather than the attachment being a necessity, it would appear that it was used as a mechanism to attach an asset which was likely to be removed.

It is also unfortunate that the learned judge in Steyn found that the arbitration award ceased to exist once it was made an order of court. Whilst one can appreciate why the judge came to this conclusion, such a determination is inconsistent with the New York Convention in that an arbitral award is capable of enforcement in any number of jurisdictions without affecting the nature of the award.

The South African court ultimately found that there was no arbitration award capable of enforcement in South Africa. But South Africa was not the seat of the arbitration and so the local courts are not empowered to determine the validity of a foreign award. Is this not a further issue with the Steyn judgment?

That is correct – The South African court was asked to enforce a foreign arbitral award. Section 16 of the IA Act obliges the court to enforce such awards upon request. Only in limited circumstances can a court refuse to enforce an arbitral award. These are set out in Section 18 of the IA Act.

The court was not empowered to pronounce on the validity of the award – it could only enforce or refuse to enforce the award if one or more of the grounds specified in Section 18 of the IA Act were satisfied. By declaring that the award no longer existed, the court exceeded its authority and usurped the power of the courts at the seat of the arbitration.

But these decisions are not all doom-and-gloom as we often see such decisions in jurisdictions coming to terms with new legislation. Do you think it is important that these decisions are scrutinised and commented upon so that ultimately, the courts can learn from the arbitration community and correct the position in due course?

That is indeed so. As long as practitioners are debating these decisions, we can only hope that, ultimately, we will arrive at the correct outcome.

Download Inside Arbitration Issue 9

Key contacts

Jonathan Ripley-Evans photo

Jonathan Ripley-Evans

Partner, Johannesburg

Jonathan Ripley-Evans