Notes

How to build a European Data Economy?

• The EU needs a policy framework that enables data to be used (...)
• A framework including, an increase of:
  • Flow of Data (Regulation 2018/1807 + GDPR)
  • Data Access and Transfer (guidance on data sharing, debate on data ownership).
  • Data Interoperability (improving tech interoperability, recommended contract terms to facilitate switching).

Now, Data and Market Power

• Does it exist an automatic relationship between big data and dominance? No, there is not
• Can you argue that a firm holding big data holds a dominant position as well?

  No, it's not automatic, however many times, holding big data is a barrier to entry for new competitors. Indeed, often the companies with the richest datasets are the ones in a dominant position in the market.

To assess dominance:

• Define the relevant market
• The market shares of the firm under scrutiny
• The market shares of its rivals
• The barriers to entry

Dominance is not sanctioned, only the abuse of dominance is sanctioned, for this reason, when in a dominant position it needs to have more attention in its decisions.

Example

• Suppose that a food-player (e.g. JUST EAT) wants to make and market a new flavour of pizza. It needs to know about consumers' tastes. The relevant market to consider is "pizza production" not the market of data on consumers' tastes.
• Where does JUST EAT look for data? The market of data on consumers' tastes is the market of inputs that JUST EAT will use to make and market pizzas.

The fact that JUST EAT holds a huge amount of data regarding consumers' tastes and preferences doesn't put it in a dominant position in the pizza production market. Also in the case JUST EAT wants to enter that market, which can be considered as secondary for him.

The relevant market of the firm is the market of the product/service that the firm is going to realize by using the data (extracting information). Therefore, not only, the mere fact that a firm controls big data does not tell you anything about the markets from where the data come. Also, it does not tell you anything about the markets where the firm works. But this second market is where the firm could hold a dominant position.

This is the reason why there is no automatism between big data and dominance! Indeed, many firms hold big data and are not dominant. Then, it may always happen that:

• A firm controls a very specific set of data
• Those data identify a specific market
• The firm is dominant in that market

Or, it may happen that a firm is dominant in the production of a good that use some data. However, this has nothing to do with the automatism between big data and dominance.

However, we cannot neglect that a company is dominant in the production/offer of a product/service also because of the data that it keeps collecting/generating/producing and that it holds.

Indeed, there is no dominant company that does not hold a huge amount of data.

Does holding data confer dominance? No Can we argue that holding data should count is assessing dominance in a product/service relevant market? Yes Big data can work as barriers to entry protecting markets for products or services.

Data do reinforce the platforms positions in offering services and products on both sides.

Therefore, Big data may work as barriers to entry! Other factors to consider:

• Availability of Data
  • A same information can be found using different data (e.g. search engines requests, social media analysis, purchasing data on online platforms, etc.)
  • This is liable to mitigate market power linked to the holding of vast amount of data.
  • However: need to assess whether data are really substitutable -> do they allow exactly for the same use? Case by Case analysis.
• Volume and Variety of Data
  • Volume and variety are not always a necessity; there can be a level beyond which the usefulness of data remains constant.

Joint German - French report on Competition and Big Data (May 2016)

• Abusive conduct can arise from firm's capacity to derive market power from data that its competitors cannot match (case-by-case analysis needed).
• Mere possession of data does not equate to dominance, depends on how difficult to replicate/usefulness, which depends on the specific market features.

There is a massive debate on the need of slightly changing the way we assess and define the relevant market.

Conclusions on Market Power from Data

• A case-by-case analysis is needed.
• Data usually reinforce dominant positions: they may work as barriers to entry the market of a product/service and these barriers should be considered to assess dominance.