The public opinion seems to be that the market power of Big Tech companies needs to be controlled, if necessary by splitting them up. It is easy to understand that sentiment. Big Tech companies know almost everything about us and may behave in a way that many people do not like, notably in how they use our private data. Moreover, their market power is a threat to competition in digital markets, potentially leading to higher prices (where the price may consist also of the personal data we provide to Big Techs), a sub-optimal rate of innovation, and efficiency losses in the form of a social deadweight loss.

However, in this blog I would like to explain that splitting up Big Techs has some disadvantages, while the long-term effects are not clear. Before I get to those points, let us first take a step back and address a more basic question.

Can we split up Big Tech companies?
Legally, splitting up Big Tech companies seems difficult but not impossible. Under merger control, ‘structural remedies’ (which imply that parts of a large company have to be divested before a merger or acquisition can be approved) have already been used for a long time. In that context, a divestiture is an ex ante measure, to which undertakings agree beforehand (not always happily), because it allows them to go ahead with the merger or acquisition. Ordering a divestiture under other branches of competition law, however, would function as an ex post remedy, as a kind of punishment for abuse of market power. There are only few examples of that under competition law.

We need to have very good reasons for imposing structural measures ex post. Sanctioning companies in this way under competition law would set precedents. Splitting up companies under competition law should therefore be considered as a measure of last resort, only to be used if there is ample proof of severe and repeated abusive behaviour. The question arises whether we need to use this measure of last resort here, knowing that we can also make use of regulatory options available under consumer protection law (regulation of B2C contracts), IP law (regulating limits to the extent of patent or copyright protection) and privacy law (such as the GDPR), besides the available options under competition law.

What would be the effects of NOT splitting up Big Tech companies?
As I mentioned above, the negative effects of ‘allowing’ the Big Tech companies to be dominant in digital markets include at least the following:

1. the possibility that final consumers or business customers will be exploited (less choice, use of private data, etc.);
2. the possibility that rival firms will be foreclosed from the market; and
3. the possibility that the innovation rate is less than optimal.

However, the analysis should not stop here; all of the above effects first need to be proven. These are not facts, but possible scenarios. For example, many consumers probably do not mind that their private data are being used by Big Tech companies, as long as they receive better products or services in return. We hence need to find out what consumers really want, rather than making simplified assumptions about their preferences. Furthermore, regarding rival firms, regulators and competition agencies should not necessarily intervene when competitors are being foreclosed. We should only protect those rival firms who actually bring something valuable to the market for consumers. It would be inefficient to protect small and inefficient competitors at all cost.

What could be (under-examined) consequences of splitting up Big Tech companies?
While some people may believe that splitting up Big Techs would solve the (potential or alleged) problems mentioned above, I would here like to highlight some potential effects of splitting up Big Tech companies that have been ignored somewhat in the discussion.

First, the long-term effects on competition may be smaller than expected. Digital markets, particularly those for social networks, operating systems and search engines, seem to have characteristics in common with the so-called ‘natural monopolies’ defined in economic theory, which traditionally included markets for particular infrastructures (like railroads) and public utilities (like electricity transmission). In natural monopolies, like in digital markets, producers benefit from economies of scale. Moreover, in digital markets there may be network effects on the side of users: the more consumers are on a digital platform, the better its perceived quality. This also applies for advertisers who make use of the platform to attract consumers. Therefore, some digital markets, like natural monopolies, may be suitable for only one or a few players. Chances are high that in the future consumers will flock again to one and the same platform if we split up Big Tech companies now. This raises the question of how many social networks or search engines society wants.

Second, splitting up Big Techs may have a ‘chilling effect’ on innovation. If companies who gain their market power by offering the best product at a reasonable price would be punished by having to divest parts of their business, ‘competition on the merits’ would be distorted. We should hence do this only if there is strong proof that prices are too high or trading conditions unfair. Of course, it remains important to have actual or potential competition in a market if we aim to reach an ‘optimal degree’ of innovation. In digital markets, such competition may also come from outside the current market. A technology that is dominant now may become outdated in the future (think of Nokia’s dominance before the introduction of smartphones, or of TomTom’s huge market share before integrated navigation systems and Google Maps became the standard); and the same may be true concerning social networks or the way we search for information or buy products online. We cannot predict how digital markets will develop in the future, even when (differently from Nokia and TomTom in the examples above) Big Tech companies are dominant in several different markets at the same time.

Third, while it is often assumed that the market power of Big Tech companies is bad for all other firms, we should not forget that there are many firms who benefit from the current situation. Think of, for example, companies who advertise on Google or Facebook, companies who offer their services via Amazon or Apple, and so on. When intervening in markets, regulators and competition agencies should not forget the economic interests of these firms, often SMEs, who rely on certain outlets for their advertisements and product sales. Splitting up Big Techs may not necessarily improve the situation of these firms. Of course, it remains important at all times to carefully monitor the price setting behaviour and trading conditions (including sharing of essential information) imposed by Big Techs on their business customers.

The way ahead
I would like to stress again that splitting up companies should be a measure of last resort. Instead of doing that, we could further explore the options in ex-ante regulation. This would also be in line with economic literature on natural monopolies, which suggests that we should allow a low number of market actors in these markets, but that regulation of price or quality may be required to protect consumers and business customers.

With the EU heading in the direction of more ex ante regulation in the form of a Digital Markets Act and Digital Services Act, among other developments, this is also the most likely way forward, even if the pros and cons of the Digital Markets Act are still much discussed in the literature. Moreover, we should not forget that in competition law, IP law, consumer protection law and data protection law we already have several options to address abuse of market power and unfair practices by Big Techs either ex ante or ex post. Examples include a stricter control on ‘killer acquisitions’ in competition law, defining limits to the use of patents or copyrights in IP law, or designing better rules on consumer data use and user consent.

Splitting up as a last resort is recommended only in those situations where it is both possible and efficient, that is, where it does not distort possibilities to profit from network effects and economies of scale, and taking into account the actual preferences of consumers and business customers. In other words, there is still a big challenge ahead of us!