Cartel detection in the South African bread market : a review of the studies by the Competition Commission and National Agricultural Marketing Council

Abstract

The South African Competition Commission has analysed most levels of the food supply chain by investigation of alleged anti-competitive behaviour by producers, input suppliers, storage companies, processors and retailers. The numbers of these cases include cartels and, to a lesser extent, restrictive vertical constraints and abuse of dominance position. Sexton (2012) states that the recent development in the agricultural market, where large companies are vertically integrating, renders the perfect competition model inappropriate. This brings doubt that the law based on perfect competition will work in oligopolistic markets. Despite the identified number of detected cartels in the grain industry, it might be possible that some cartels in other food chains are still undetected. Through the application of the market screening approach, the aim of this study is to establish whether the National Agricultural Marketing Council (“NAMC”) and the Competition Commission could have detected the bread cartel using secondary data in the absence of the information from the whistle-blower. As the first step, the study carried out a structural assessment of the bread industry. This assessment indicated that the bread market has a number of factors that may facilitate collusive behaviour. The study found that the history of information sharing played a crucial role for bakeries to coordinate their conduct. The second step was to conduct an in-depth behavioural assessment that focused on bread prices to see whether there has been a structural break in the period under investigation. The idea was to estimate the price equation of brown bread as a function of the SAFEX wheat price, petrol price (cost shifters), and 1 kg of maize meal (demand shifters). The study used the OLS to estimate three regressions using the data for the whole period and two sub-breaks (before the break point, and after the break point) to perform a Chow test. The question that the Chow test asked is: was there a structural break in March 2007, after the Competition Commission received the information from the whistle-blower? In other words, had the price of bread increased or decreased at a certain period without any changes in the demand or cost variable. The Null hypothesis states that there was no structural break, while the alternative hypothesis states that there was a structural break in March 2007. The Chow test result shows that at a 5 per cent significant level, the F-critical value is F_5, 90 = 2.68 and the F test statistics is 20.59 with a p value of 0.00. This indicates that we cannot reject the null hypothesis and conclude that a structural break did not occur in March 2007. The screening approached failed to prove the existence of cartel in the bread industry. Therefore, the study concludes that in the absence of the whistle-blower, it would not have been easy for the Competition Commission and the NAMC (2009) to detect a cartel by just using secondary data. This proves that screening alone cannot prove the existence of cartel without prior knowledge of the conduct and of the industry as a whole. Nevertheless, a market screening approach is important as it can be used as a warning mechanism to detect an emerging cartel, since it can flag potentially suspicious behaviour. Nevertheless, this calls for the policy makers to combine the scoping study by the Competition Commission and the monitoring of food prices by the NAMC, as this will provide the best enforcement tool in detecting cartel behaviour in the food industry.