Alex Constantinou, Andrew Sylvester and Sifiso Mhlaba

Recent revelations of collusion in the construction sector in South Africa have led to widespread anger and debate. In June 2013, the Competition Commission of South Africa reached settlement agreements with 15 of at least 21 construction companies implicated in the cartel. Much less has been heard on the matter in recent months, but this apparent quiet might just precede the storm to come. There are strong indications that various parties are preparing legal action to claim damages against specific cartelists, the quantum of which could be substantial.

What is known by the public at this stage is that these firms disclosed 300 rigged projects between 1999 and 2009 worth a total value of R47.1bn. More than half of this was commissioned by public entities, not least projects such as the World Cup soccer stadia. The manner in which these cartelists achieved higher than ordinary prices was by rigging bids through cover pricing arrangements.

Cover pricing refers to a situation where firms, prior to tendering, meet behind closed doors to allocate projects amongst themselves. The anointed winner places a bid while the co-conspirators dishonestly inflate their bids in order to throw the tender. This has the effect of duping the client into believing the most competitive bid was chosen via a competitive process. The co-conspirators are then sometimes paid a losers’ fee to spread the illegal profits between the cartelists.

While the process revealed cartel activity from at least as early as 1999, the Commission could only legally prosecute cases from September 2006 onwards. In the vast majority of cases, it settled with the construction firms with only 3 outstanding cases remaining. The total value of the penalties agreed through settlements was R1.46bn.

These settlements should not be misunderstood as a measure of, or an attempt to recover the harm inflicted by the cartel, and this money is by no means transferred into the pocket of the harmed consumer. Cartel enforcement has as its goal the deterrence of cartel activity. High penalties not only contribute to this goal, but fast action and a clean slate might have better served the public interest rather than 140 hearings which may produce higher penalties half a decade down the road. With the penalties agreed and paid, the question still remains: “How much harm did the cartel actually cause?”

An effective cartel mechanism increases the price paid by customers whilst reducing the volume and/or quality supplied to the market. There are thus transactions which continue to be made, albeit at an inflated price or poorer quality, and also transactions which are no longer made given the reduced supply. Both harm society directly. The first harm can be calculated as a transfer from customers directly to cartelists’ bottom line. This is effectively theft for which customers are empowered to sue cartelists in civil damages suits.

The second harm can be thought of as collateral damage, which economists call dead weight loss. This hurts the customers who choose not to undertake the project at all, or fewer projects than otherwise, given the more expensive cartel price. It also hurts those producers through diminished sales, and in turn reduces employment and demand on upstream suppliers. The repercussions from this collateral damage can therefore extend across the wider economy and continue over several business periods.

Calculating the actual cost of the transfer from customers must be done on a case by case basis. However, we can get a rough estimate from international studies of construction cartels. We isolated bid-rigging construction estimates from Connor’s 2010 influential and comprehensive survey of cartel mark-up studies and observed that, on average, bid-rigging construction cartels were able to mark-up by 20.4% over the normal non-cartel price. The table below provides a list of all construction sector bid-rigging cartel’s from Connor’s study.

This analysis is based largely on road construction, school buildings, sewer construction and other general public construction projects spanning the USA, Japan, Germany, the Netherlands, France, and Korea. As such, it makes for a reasonable benchmark in calculating an equivalent mark-up for the South African construction market.

To get a more accurate depiction of the harm caused; the value of all the projects need to be converted into real 2013 terms. The table below shows this conversion.

Since the cartel spanned 1999-2009, older projects need to be inflation-adjusted to reflect what they are worth in today’s terms. To this end, we distributed the 300 projects over 1999-2009 according to the number of prescribed and non-prescribed projects in each time period. From September 2006 to 2009 there were relatively more projects than 1999 to August 2006. The Producer-Price Index (PPI) was then applied to obtain the nominal and real values of the projects. Ultimately, the initial quoted R47.1bn value of the projects increased to R70.5bn after it was adjusted into present terms.

With the total value of all the projects converted to 2013 Rands, it now becomes possible to estimate the financial value of the transfer from the customers to cartelists, as well as the scale of the deadweight loss (collateral damage). The calculations are as follows:

Applying an average overcharge from Connor’s study of 20.4% and adjusting project values to 2013 Rands, we calculate that the total cost of this transfer for the 300 disclosed projects would amount to around R11.95bn. Calculating the cost of the deadweight loss is trickier and has been estimated by DNA Economics through the construction of a basic economic model. Making conservative assumptions about cost structures and the willingness of customers to buy more construction projects had prices been lower, we estimate the total deadweight loss to be R466mn. Of this loss, R311mn was incurred by customers, with the remainder falling on producers. This takes the total harm to customers to around R12.3bn. Even if we halve the assumed mark-up to 10.2% in our model, the total harm to customers is estimated at R6.6bn.

The abovementioned transfer and loss figures apply to public and private, including prescribed and non-prescribed projects, i.e. all 300 projects. However, if we isolate the public non-prescribed projects, this gives rise to a transfer from customers to construction companies of R5.9bn and total harm to the public sector of R6.7bn (which includes DWL) at an overcharge of 20.4%.

These are big numbers, and noticeably bigger than the penalties agreed to in the settlements. Had the cartel never existed, tens of thousands of dwellings could have been built or hundreds of rural schools constructed or upgraded. This is because around 60% of the rigged tenders were paid for by the state; it is therefore reasonable to assume that these stolen resources could have been deployed to social projects elsewhere.

Clearly, the effects of the cartel are significant and extend beyond the purchasers of construction projects. It follows that recourse should be available and the law rightly allows those who have been adversely affected by such activity to claim civil damages. However, in doing so, it is normally the case that specific overcharges and the extent to which these charges are passed-through to consumers, amongst other assumptions, would need to be proven on a case-by-case basis. In the short term, this will require more detailed economic analysis and potentially costly disputes. In the longer term, more vigilant procurement will be required. The recent establishment of a Chief Procurement Office at National Treasury is a critical step in the right direction.