Global Big Tech Enforcement and Impact

Wednesday April 10, 2024, from 3:30-5 p.m. M2: Salons 12 & 13

Cristina’s informal speaking notes

(disclosures: I have advised against Google and Facebook, and for Apple, Amazon & Microsoft)

• Cristina, in Europe, the DMA was enacted in response to the European Commission’s experiences with the multiple tech cases and in reaction to the perception that the remedies in those tech cases were not effective. What kinds of “big tech” enforcement do you think we’ll see in Europe in the future? And what kinds of remedies can we expect to flow from the DMA and any future conduct enforcement? (6 minutes)

• Potential follow-up question: What do you make of the ongoing DMA non-compliance investigations, and particularly the investigation into whether Google is breaching the DMA prohibition of self-preferencing in Article 6(5)?

Europe started early. For a trip down memory lane, 5 years ago almost to the day I organized an informal side event (round the corner at Fig & Olive) based around the idea that Europe was clearly ahead in digital enforcement, the US was doing nothing.  Speakers included Rebecca Slaughter, Rohit Chopra, Max Miller, Tommaso Valletti, Isabel De Silva and the discussion was very much around how Europe was “on the move” while the US was not. 

However, change on the ground since then has been nil – Google Shopping has not changed the fate of CSSs (Comparison Shopping Sites) and search more generally, Google Android has not shifted Google Search from Mobile phones, Amazon’s two cases lasting years ended in opaque settlement & no change, and we just saw the Apple Music Streaming case going for 8 years ending in a nothing burger for Spotify. 

Multiple reasons.

1. Extremely elongated processes
2. Twists and turns on the theory of harm – reluctance to make EXPLOITATION cases but only EXCLUSION when the issue with many of these platforms is EXTRACTION – wonky theories of harm which are pushed back and redone multiple times
3. Remedies – “cease and desist”  is not a consent decree and certainly not a structural  remedy – the firm needs to come back with remedies which the Commission will assess for “compliance” – we are right into that rodeo now with Apple and the Music Streaming case.
4. But I would also add another less discussed and more uncomfortable fact which is super important.  The reality is that CASE SELECTION, issues and the focus of the case are in practice dictated by complainants. (Not to say anything goes but the process is highly responsive to complainants raising issues and providing relentless evidence, submissions, arguments, going to meetings etc. ) The implication is that BY DEFINITION cases are narrow and shaped by the particular issue the complainant cares about.  CSS was not where you’d start if you had a concern about Google Self Preferencing in Search; in Apple you would not focus on Music Streaming Services just because there’s a Spotify’s complaint – so these cases are very narrow because the complainant has his own issue. There’s NO ZOOMING OUT TO THE BUSINESS MODEL, AND HOW YOU TACKLE A BUSINESS MODEL that is a collection of intersecting businesses and practices creating a moat and exercising power in multiple ways. This is the real issue. It’s not a little conduct here in this market, and a little conduct there in this other separate market. So no chance this approach can disperse any power or create real competition – it is only ever just about whether a particular class of complainants can get a bit more of the jointly created pie.

In passing let me mention this is NOT the approach of the FTC’s/ /DOJ’s recent complaints. Will discuss it later but let me make on general comment. It will all rest on the quality of the facts etc etc but they make an effort to put forward an OVERALL theory of the case, of what are the components of the business model and how things work together to protect its power.  So the Amazon complaint brings together all aspects of the business – ecommerce platform but also algorithmic pricing, advertising, logistics – and says there are these mutually reinforcing concerns that support an extractive ecosystem which is now exploiting sellers and consumers. The Apple complaint is looking at how the ecosystem protects itself from innovation and challenge – puts forward a theory (ambitious yes but at least tries) that there are functionalities Apple has frustrated because they could provide a middleware bridge to a challenger -and this has frustrated innovation in smart devices. Wow this is not this little piece – antisteering rules for music streaming apps – it’ a big ambitious effort.

Now let’s go back to Europe. All of enforcement did not do much, enter the DMA.  The problem is the DMA is list of “dos and donts” which is squeezed from a synopsis of the very enforcement cases – narrow, limited – which we have seen fail.  So you can map the rules into cases – self preferencing is a bit of Google and a bit of Amazon,  seller data is Amazon, there’s a host of things that you can map back into the cases.  So what’s different?

Ah the idea is we did not succeed with enforcement, now we tell companies pre-emptively NOT to do something, or to DO something and they will have to do it.  They will have to comply with the law. So this will be self executing, marvellous!  There will be some challenges yes but it will work…

Now it’s early days but two things. First, the challenges of implementation are becoming apparent.  Because if you set out a set or rules but are not VERY SPECIFIC about what companies should do and leave it to THEM to comply, well…you get that they don’t do what THE COMPLAINANTS IN PARTICULAR think they should do. Surprise surprise. “They are MALICIOUSLY COMPLYING!” Well that’s a way to put it but left to themselves companies do one of four things (or possibly all).

• What exactly do you want me to do. Be very specific. Well if you wont tell me what to do I will do what I think the law says.
• I will do this little bit and try you out – how much appetite do you have for a fight.
• No that’s dumb, I just wont do it forget it.
• See you in court.

So on the one hand we are suffering from lack of prescriptiveness. “Plutes are gonna plute”, but regulators gotta regulate. Companies cannot be expected to give access without some reasonable access regime.  Cannot be expected to voluntarily come up with a solution that pleases the complainants who will always scream. If you don’t do that, it’s a game of cat and mouse -indeed a week after the deadline for voluntary compliance, the EC fired off three suspect non-compliance cases and there’s more in the pipeline. SURPRISE.

At the same time by definition, by design this regulation cannot really bring about much more than some more fairness and some more opportunities for entrants, but limited and so if one is really looking for something that deconcentrates markets, that challenges the power, this will not do it. It will be marginal, and take a long time I fear.  Which is why I think the current hysteria around this, the silly communications with special  effects, the emoji tweets, “not all heroes wear capes” etc, the idea this is transformational regulation is delusional.  Something may come of it, but it will take time and not be as transformative as the DMA groupies hope. 

• Cristina, in the context of DOJ v Apple and Epic v Apple, when we think about Apple in Europe, where we have the abuse of dominance music streaming decision, the proposed in-app-payment commitments, and the DMA obligations which overlap. Are effective remedies in place in Europe, and how if at all, should they inform the enforcement in the U.S.? (6 minutes)

This is a hot mess in Brussels right now. A hot potato for the EC to handle.

There’s a pending not yet concluded enforcement case on Appe Pay, which is in the final stages of commitment but not yet finalized.

Then there is the MSS Spotify case, which just concluded after many gyrations with a Theory of Harm that is just saying “cease and desist” from stopping apps from steer users to subscribe outside the App Store similar to the Epic judgment in California. That’s it. Not tackling the 30% commission that has always been the conflagration point for developers. If Apple must allow steering out, then the question is okay, what can Apple charge developers for this? Because developers want to pay zero and Apple says wait a minute, you use my tools, you use my device to distribute your app and give it visibility to 2bn customers, and you think you can have someone subscribe outside of the App Store and pay us nothing?  This is a classic ACCESS PRICING issue for a platform. Been there, of course complicated because it’s all IP – but Apple of course expects to charge, and said in fact you pay me 27%. Spotify hates it and is screaming murder – that the remedy is non compliant and the EC is looking into it. 

AND THEN THERE IS THE DMA. Which in effect has the same no-antisteering rules for App Stores as a whole not just limited to music streaming. But what happens there is not independent right? EC needs to investigate Apple’s antisteering etc  in general as part of the DMA while at the same time investigate whether the proposal for Spotify and MSS is compliant with the Decision. 

Finally Apple also proactively redesigned its business model and came up with a different charge structure – unbundled etc. CTF etc. What do you expect them to do, if they do something proactively and spontaneously without constraints?

This is the sense in which again, IF we decide to engage in regulation then we need to be prepared to be prescriptive. No such thing as light touch regulation which then has the companies doing the right things spontaneously. Either you have a fight before a judge, or you design a different system that makes sense and deals with access regimes etc.

But my point remains is this law capable of doing anything in the way of truly challenging power? It can create some more competition ON THE PLATFORM, not TO THE PLATFORM .

• Cristina, you have been a thought leader on the topic of ecosystems theories of harm. What are the factors that drive the creation of ecosystems in digital markets? And, beyond mergers, do you think it’s important to incorporate this theory of harm into market definition or effects analysis in “big tech” monopolization cases?  (6 minutes)

• Potential follow-up question: What do you make the DOJ’s Google Adtech and Apple monopolization cases and how each case incorporates ecosystem theories of harm?

The focus has been finally placed on ecosystems as “the way that large platforms compete”. The reality is there are constellations of assets and capabilities which are fungible and can be swung across markets – data, customer bases, operating systems, IP, and market power comes from controlling this collection and being able to use it in ways that can forestall competition –

Here the analysis cannot be market by market, but needs to be taking into account how these assets & capabilities can combine to create problematic levels of market power.  Reality is firms are endowed with capabilities and assets – which are the sources of their success/relative advantage – and may be hard to replicate/scarce.  When we think about firms’ competitive advantages in markets, that’s really being driven by capabilities, not by a share in a product market.  1.    Why are we getting there now: because the network of capabilities is changing

The capabilities which markets value have been evolving over the last 20 years. Making markets more interconnected than they were before. This is especially so in digital.  And because they are more interconnected there’s more synergies across markets and there’s more benefit from owning a larger constellation of capabilities. One thing that has happened is there were capabilities that weren’t valued by markets that have become valued by markets. e.g. algorithm that helps you recognise shapes valuable to building a self driving cars, or language

Another thing that has happened is new markets have emerged that we have not seen before. And new capabilities have emerged. One we had not seen before is consumer data. Data storage. Wasn’t so valued 20 years ago but now it is widely deployed across markets. So at the end you get a picture in which markets are much more interconnected than they were before.

In Microsoft/Activision and Booking/etraveli the agencies have started questioning if there is a problem with deals bringing together assets/capabilities but where there is no suggestion of a tie or a bundle inducing users to consume together.  As we will hear today, they were different: in Microsoft/Activision the ecosystem concern was whether the constellation of assets that MSFT owned, OS, cloud, gaming portfolio, data, hardware etc was going to give it such a COMPETITIVE ADVANTAGE in the incipient cloud gaming market when combined with the acquisition of Activision content, that no one could bridge that initial advantage. In Booking the concern was that the acquisition of an adjacent activity would have FURTHER STRENGTHENED the dominance of the incumbent buyer by deepening the moat.

In mergers, the issue is “what is the change the merger makes” – some have called for a limiting principle but of course limiting principles are always invoked when one wants to do nothing. The point is that one does not need a specific leveraging mechanism from A to B, which is how these conglomerate leveraging cases/ conglomerate theories have worked in the past – relying on foreclosure mechanisms. Company X who is the buyer has power in market A and leverages it into market B with some mechanism – tying, bundling etc.  that was the traditional approach. Identify a market from which you leverage, and one into which you leverage. And a precise mechanism. Then the discussion was about the mechanism. Can it induce foreclosure? Vertical arithmetic, calibrated models.

That is a hopeless way. It does not work like that. Because defendants can always produce a little model that says look it would not be profitable to foreclose etc.  Look at my Vertical Arithmetic! But the issue is in what way does the combination of the assets and capabilities of the target together with the buyer makes a difference. This was a big issue in Microsoft Activision until regulators gave up and pivoted into a standard input foreclosure story. But that’s really not satisfactory.

My view is that for very large platform where it is clear there is an ecosystem, there should be just a bar on deals. Enough. Make it yourself. So what if we don’t reap some benefits ? What is the benign history of acquisitions which has created consumer benefit as opposed to VC benefit statup benefit shareholder benefits?  Just stop.

We have a problem with the “direction of innovation” being dictated entirely by a handful of large ecosystems.  Everything that is being created loks like an app morivated in one of three ways. We need policy innovation to articulate the issue, promoted by agencies as stakeholders.