‘PAC’ - A competition theory to unpick a dilemma

The recent Z/Chevron merger, which reduces from four to three the number of big players in the New Zealand retail petrol market, required the Commerce Commission to tackle the thorny matter of tacit collusion.

The recent Z/Chevron merger, which reduces from four to three the number of big players in the New Zealand retail petrol market, required the Commerce Commission to tackle the thorny matter of tacit collusion. 

For most mergers, this issue scarcely arises and instead the regulatory focus falls squarely on unilateral effects (the ability of the merged entity to raise price independently). But the hurdle in this case was a high one as the market exhibits a number of features which make it vulnerable: homogenous products, a small number of similarly sized competitors with similar cost structures and an ability readily to observe each other’s prices. 

Indeed, the market has developed around a shared processing and distribution network in the form of the country’s only refinery and the associated distribution and storage facilities. 

Tacit collusion is practiced by competitors too clever to enter into an explicit and therefore unlawful “contract, arrangement or understanding”.  In essence, we are talking about learned deviousness. By definition, tacit collusion cannot be observed, it can only be inferred. The practical difficulty for the regulator is how to reliably distinguish it from vigorous competition. They often look pretty much the same. 

Consider the following scenario:

Petrol Retailer A operates a service across the road from and in direct competition with Petrol Retailer B. Overnight, the price of crude oil rises significantly, raising the input costs for both businesses. The following day, Petrol Retailer A observes that Petrol Retailer B has increased prices on the board by 10 cents per litre.

Petrol Retailer A now has two choices – match the price rise to restore margins or hold prices to win market share. If Retailer A holds price, Retailer B will likely respond by reversing its price rise to stay competitive with Retailer A. The result then would be that market shares stay the same, but margins are reduced and both parties are worse off. So Retailer A may well choose to match the price rise.

What just happened there? From one angle, this looks like tacit collusion in the sense that one competitor accommodated the price rise of the other, instead of challenging it. From another angle, the decision to match the price rise was in fact motivated by the knowledge that competition is so fierce that Petrol Retailer B would never sit idly by and watch its rival win market share so easily. So, which is it: collusion or competition? 

To help regulators to navigate this troubling dilemma, economists have developed an economic framework for thinking about whether observed parallel accommodating conduct (PAC) is collusive or competitive. PAC theory has gained greater currency since 2010 when it was included in the US Horizontal Merger Guidelines.

The most intriguing approach of the PAC model is that an implicit agreement, or mutual understanding, can be formed by participants in oligopoly markets simply by learning commercial sequences over time. 

Specific parties adopt the role of “leader” and the others follow according to “unwritten rules” which have been established over time through an aggregate set of experiences among market participants, often including some form of initial signal or announcement and then consistent “punishment” of deviation. 

Through this mechanism, prices gradually float up and sit above the competitive level. It is particularly noted that “punishment” need not take the form of a substantial price drop. It could just be a reversion of a price to the pre-price rise level – as per the example provided above. 

The reality is that this particular articulation of how tacit collusion manifests doesn’t take us far. It still requires a binary judgement call as to whether observed parallel conduct should be attributed to an implicit agreement or normal competition. 

The initial “signal” is inherent in retail markets where prices are able to be monitored freely. It is of course inevitable that firms will observe the conduct of their competitors and seek to identify patterns and understand motivations. In doing so, they are not necessarily seeking to understand the “rules of the game” through which they can collude. Rather, they are gathering intelligence to outwit the competition and steal market share. 

In some cases, the smart play is to follow a price rise. That does not mean there is an implicit agreement to follow a price rise.  Similarly, the smart play might be to reverse a price rise, if it is not matched. There is no particular reason to think of that as “punishment”. It is just as plausibly a common sense decision to avoid hanging out in the market with a higher price consciously ceding market share to a competitor. As such, at minimum, compelling additional evidence should be required in order to justify attributing observed parallel conduct to tacit collusion.

Government data in the Z/Chevron case suggested that petrol margins had grown significantly across the industry in recent times. However, it was also evident that market shares were volatile, which cut across the finding of an implicit agreement.  

In the end, the majority within the Commission kicked for touch and (rightly) decided that the divestment by Z of 19 retail outlets in locations where aggregation between the Z and Chevron was most problematic and the relatively benign competitive profile of Chevron meant there would be no change to the overall state of competition in the market post-merger:

We find that much of the evidence could be viewed in different ways, one consistent with coordination and one consistent with competition.  But when viewed in the round we consider that even on the view that is most adverse to Z, that evidence points only to the possibility of coordination occurring in these local markets in the form of an implicit agreement between competitors to accommodate each other’s price rises.  The evidence is not strong or unambiguous and the evidence we have reviewed could also be consistent with competitive outcomes.
Accordingly there was no need to make a finding on the base level of coordination currently prevailing. So, notwithstanding that these issues were front and centre in the Z/Chevron case, the difficulty of pinning down tacit collusion – of distinguishing it from robust and competitive parallel conduct – remains. 

Neil Anderson is a partner at Chapman Tripp and Lucy Hare is a senior associate. Both specialise in competition law and advised Z on the Chevron merger.

Recent articles & video

Thomson Geer partner: Bruce Lehrmann suit teaches a lesson about bringing unmeritorious claims

Australian Ethical picks up Altius Asset Management with KWM's guidance

Griffith Hack welcomes patent attorney, patent associate to Perth team

Independent review calls for more investment in Aboriginal and Torres Strait Islander Legal Services

European Union officially approves landmark AI legislation

UK judge predicts artificial intelligence will transform role of expert witnesses

Most Read Articles

Thomson Geer partner on the biggest impact of the Bruce Lehrmann defamation ruling

Pearce IP's AU litigation head shares the most important lesson she learned from her old job

G+T guides Mitsubishi UFJ Trust and Banking on $2.1bn pickup of Link Group

HWL Ebsworth reveals new leadership and management structure