An active research community, delivering high-quality research on competition policy matters, supports and promotes sound competition policy enforcement in South Africa and beyond. At the CCLE, we aim to build the next generation of researchers and to generate policy-relevant work of an international standard.

Research programme 1:

Cartel damages

Collusive behaviour has received significant attention in South Africa in recent years, with a number of high-profile price-fixing cases involving multi-billion-rand fines. A key aspect of anti-cartel enforcement involves fines and, increasingly, private damages claims. The estimation of cartel damages is therefore a key analytical problem for competition agencies and other practitioners. This research programme studies various dimensions of the damages estimation problem, including:

Cartel dating

In traditional models of cartel damages, court-determined dates are taken to represent the period of cartel effect, even though periods of effect and liability need not coincide. Willem Boshoff and Rossouw van Jaarsveld are collaborating with Dutch researchers Maarten Pieter Schinkel and Maurice Bun to study empirical methods for cartel dating.

Related event: Imperfect forms of collusion.

The impact of time-series properties on cartel damage estimates

While time-series data is often employed to estimate cartel damages, key time-series properties of such data – including the presence of unit roots – are often ignored, despite their potentially significant impact on estimation accuracy. Willem Boshoff and Rossouw van Jaarsveld use simulation models to study how time-series properties affect dummy-variable estimates of cartel damage.

Alternative approaches to cartel damage estimation

Conventional time-series approaches to cartel damage estimates face several challenges when the effectiveness of a cartel varies over time. Willem Boshoff and Johannes Paha collaborate with German researcher Philip Klotz are studying appropriate estimation methods to prevent a mis-classified treatment variable.

Cartel damages in capacity collusion

Current models of cartel behaviour typically do not capture the gradual transition between periods of collusion and competition in a market. Furthermore, post-collusion prices may be quite different from pre-collusion or other counter-factual competitive prices: for example, tacit collusion or strategic behaviour aimed at minimizing damages claims may result in high post-collusion prices. Willem Boshoff and Thomas Fagart develop a theoretical model of capacity collusion and study the implications for cartel damages.

A related aspect of cartels focuses on alternative forms of collusion. Willem Boshoff and Johannes Paha study the economic rationale for one such an alternative form, namely list-price collusion, where firms choose to collude on a non-final price (but compete in setting final prices).

Related event: Imperfect forms of collusion.

Research programme 2:

Merger simulation

Mergers for the purpose of retaining or enhancing market power remain a key concern for competition authorities worldwide. In smaller jurisdictions, data limitations and resource constraints create special challenges in the assessment of merger effects. This research programme connects with the simulation modelling framework developed by Luke Froeb, Steven Tschantz and Greg Werden and coded as an online tool for practitioners. In collaboration with these US-based academics, Willem Boshoff and Roan Minnie are working to study how alternative bargaining scenarios – especially in vertical mergers – shape the predicted merger outcomes in these simulation models.

Research programme 3:

Ex-post assessment of mergers

One of the key properties of South African merger control is the set of non-competition objectives – commonly known as “public interest” objectives – according to which the merger’s impact on transformation, employment and other industrial policy aims are evaluated. Willem Boshoff and Prince Changole are developing a comprehensive dataset of South African merger cases, including their public-interest aspects, since the advent of the new competition law regime in 1998. This dataset will allow assessing changes in the properties of merger enforcement over this period, including in duration of merger proceedings at various stages.