Only discrimination involving digital product design should be considered self-preferencing, says CADE's Fernandes

23 March 2023 20:15 by Ana Paula Candil

Brazil | CADE

Only discrimination involving the design of digital products or services should be classified as self-preferencing, said Brazilian competition authority Councilor Victor Fernandes, and business models involving contractual conditions shouldn't be.

He said antitrust authorities generally consider both things self-preferencing conduct in their investigations — an approach the councilor said is "too wide," though discrimination involving business models can be considered "a traditional antitrust violation of discrimination."

Speaking at a virtual event* today, Fernandes from the Administrative Council for Economic Defense, or CADE, said discrimination involving product or service design should be separated into two different categories: “anticompetitive innovation” or “default-setting or tech-tying.”

He cited Google Shopping and Amazon’s “Buy-Box” as examples of anticompetitive innovation, where digital products have some sort of “discrimination mechanism” involving crawling, indexing and registration.

He then cited Google's Android and Facebook Marketplace as examples of products that involve default-setting or tech-tying. He explained in the first example, an app store allegedly hinders app developers from monetizing through other platforms, and in the second, a digital platform allegedly ties its online classified ads service to its personal social network.

Fernandes said it is easier to find efficiencies when the discrimination involves an anticompetitive innovation, like Google Shopping, than when it involves default-setting or tech-tying.

"The damage potential and efficiencies of the two cases are radically different," the CADE councilor said. "It's one thing to design a product, to create a product that can generate anticompetitive potential, but its efficiency potential is more widespread, it brings some kind of more sophisticated technology, be it crawling, indexing, registration. Another thing is the default-setting, where you clearly have two separate products, two distinct markets that are being tied together by the company."

Fernandes also said the company's burden to justify conduct that involves default-setting or tech-tying is higher compared to anticompetitive innovation. "In the first group [anticompetitive innovation] we have more pronounced efficiencies, and in the second group [default-setting or tech-tying] we have a stronger anticompetitive presumption," he concluded.

On the other hand, Fernandes said remedies to mitigate the anticompetitive effects of conduct that involves default-setting or tech-tying are easier to implement than anticompetitive innovation. "When we talk about default-setting, the remedy is simpler, you break the default, or you give people options to choose all the other products they want to use as a complement. But when we talk about an anticompetitive innovation in a product, the remedy is difficult to implement because it will involve re-designing the algorithm and understanding what are the transparency options that such an algorithm uses for crawling, indexing, etc.," he said.

The councilor said CADE's ongoing probes involving alleged self-preferencing have more to do with default-setting or tech-tying than with anticompetitive product design.

CADE is investigating whether Google and Apple have abused their dominant positions in the mobile services market with their Android operating system and iOS applications, respectively.

The antitrust agency is also looking at whether Brazil-based online food delivery platform iFood has discriminated against meal voucher operators on the iFood platform. Customers that use the iFood platform to order food can pay for the service with their meal vouchers from iFood Benefícios (iFB) or competing firms. "This case is about default-setting and not anticompetitive innovation," Fernandes said.

* "Self-preferencing," Ibrac, Brazil, March 23, 2023.

Related Articles

No results found