This is an edited version of a speech I gave to TechUK on Thursday December 2nd at a panel on the new ‘pro-competition regime for digital markets’. Footage of the speech will be available online eventually.
I’m going to focus on a specific part of the proposed ‘pro-competition regime for digital markets’: merger control.
The reforms proposed by the government would block any deal involving a SMS firm that creates a ‘realistic prospect’ of reducing competition – defined as being a ‘greater than fanciful’ chance.
Under the status quo, the CMA only blocks deals if they think there’s a 50% or more chance of it substantially lessening competition. So this would represent a substantially lower standard.
It is already used in deciding whether to refer a deal to phase 2 review, but the CMA has already said that historical deals that were not referred to phase 2 would likely be blocked by this, like Facebook/Instagram, so I don’t think that’s as good a guide to what would and wouldn’t be blocked as we might like.
I think the CMA has underestimated the risks of this approach, and has been a bit myopic about the effects it has on the wider economy.
It’s sensible for the CMA to take a narrow view when it’s reviewing individual cases, but when it’s proposing legislation I’m pretty disappointed that it doesn’t seem to have considered the harms that it can cause if it is given excessively broad powers.
To understand these harms, I think it’s first worth establishing that mergers often play an important pro-competitive role in tech.
I want to focus on two key reasons why we should not automatically assume that all acquisitions in tech are suspect and anti-competitive.
1. Acquisitions are a key route to exit for entrepreneurs
High-growth tech entrepreneurship is inherently risky. Most startups fail and the VCs who fund them know this.
Only a small fraction of businesses will go on to IPO.
According to data provider Beauhurst, only nine equity-backed startups exited through IPO in 2019.
Last year eight British equity-backed startups were acquired by Microsoft, Google, Facebook, Amazon, and Apple alone.
Acquisitions, in effect, derisk entrepreneurship and investment in high-growth startups, because if your company isn’t going to be the next Facebook they mean you can still make money from it. I don’t think it’s a wise policy to try to force startups to become the next Facebook against their will – but increasingly it seems like the CMA thinks we need to protect startup founders from themselves, and force them into much riskier paths than might be wise.
Because exits are so important to founders and their investors, it should not be surprising that cross-country studies find that restrictions on mergers and acquisitions are linked to declining rates of VC investment. I think it’s likely that if the UK becomes more restrictive, we’ll see the same effect here.
2. Acquisitions facilitate interplatform competition.
Competition in digital markets often takes place between digital platforms that have a strong position in one market and move into another market dominated by another digital platform.
Acquisitions can accelerate this kind of inter-platform competition. Instead of starting from scratch, platforms can use mergers to gain a foothold in the new market, and do so more rapidly and perhaps more effectively than if they had to develop the product in-house.
A few examples:
Google’s acquisition of Android increased competition faced by Apple’s iPhone
Apple’s acquisition of Beats by Dre increased competition faced by Spotify as it accelerated the development of Apple Music
Walmart’s acquisition of Jet increased competition faced by Amazon in e-commerce.
As Ben Evans puts it: “for every Instagram, there’s a PA Semi, which Apple bought in 2008 for $278m. PA Semi is the reason Apple is now creating its own chips to power all of its devices, and why those chips are better than anything from Intel or Qualcomm. Intel had a near-monopoly of CPUs for personal computing for close to 40 years, and now Apple has broken that dominance, because of an acquisition.”
If any of the mergers I’ve just mentioned were blocked then its likely competition in digital markets would have declined, and consumers would be worse off as a result.
In essence, and I hope we all accept this, there is a cost to over-enforcement as well as under-enforcement, and I’m not convinced the CMA is taking that seriously.
The question is whether or not new harms in digital markets justify changes to the merger control regime.
I’m sceptical for three reasons.
1. I’m sceptical of the ability of regulators to identify anti-competitive acquisitions in advance.
I recently saw Tommaso Valletti say: “1,000 and zero: mergers done by GAFAM and mergers blocked in the past 20 years.”
I’ve often heard something similar cited as an argument against the status quo, including from Andrea Coscelli and the Furman Report (Tommaso’s numbers are slightly inflated).
But while these are impressive sounding numbers, the vast majority of these deals were absolutely tiny – half were worth less than $10m and 80% were worth less than $50m. Between 80-90% were acqui-hires of companies with fewer than ten staff members.
When you actually list the deals that people think were problematic, the list is much smaller – Facebook / Instagram and Whatsapp, Google / Waze and Doubleclick, and perhaps a few others.
We can have different views on whether these should or shouldn’t have gone through, but it’s really just a rhetorical trick to use the biggest number possible that, on further scrutiny, really just muddles the question.
It feels probable to me that any rule or burden of proof that picks out the rare anti-competitive acquisitions of startups will also end up picking out many more procompetitive or neutral buyouts.
The question for the CMA and the government isn’t just “Should we have blocked Facebook / Instagram?” – it’s “Are we sure that a regime that leads us to block Facebook / Instagram might not also cause us to block the next Apple / PA Semi, or Google / Android?”
2. Arguments about potential competition tend to prove too much.
Take the classic example of potential competition being snuffed out: Facebook buying Instagram.
If Instagram posed a potential or nascent competitive thread to Facebook when Facebook acquired it, with its photo feed and social features, then so must other services with products that are clearly distinct from Facebook, but have social features.
In which case Facebook faces potential competition from other services like Tiktok, Twitch, Youtube, Twitter and Snapchat, all of which have services that are at least as similar to Facebook’s as Instagram’s.
The loss of a single, relatively small, potential competitor out of many cannot be counted as a significant loss for competition, since so many other potential and actual competitors remain. We can define these markets narrowly, or broadly, but we can’t define them in both ways at the same time.
3. Are new powers necessary?
This week we saw the CMA order Meta to sell Giphy. (I blogged about it here)
One of the key justifications was that Giphy is a potential competitor to Meta in the display advertising market.
I think the decision is quite a stretch, but really it raises a more important question: If the CMA is able to stop this deal under the existing burden of proof, on such speculative grounds, what deals does it want to stop but can’t because the burden of proof is too high? Are we really sure that we want to block deals on even more speculative grounds than Facebook / Giphy?