On 17 January, Commissioner Vestager hosted a one-day conference on ‘Shaping competition policy in the era of digitisation’. As well as providing inputs to the conference on market power in digital, data and algorithmic competition, several members of Oxera’s team attended the event. Our reactions to the four main themes are summarised below.
Competition, data, privacy and AI
In the digital age, competition concerns increasingly relate to the growing ability of firms to undertake exclusionary or exploitative behaviour. A common theme is the potential for certain data to give an unprecedented competitive edge to dominant firms. But is this really a source of exclusionary power for firms, or does it simply reflect healthy competition?
Access to data is often cited as one of the main bottlenecks for competition in artificial intelligence (AI). New algorithms and AI typically need ‘training’ using historical data before they can be used, giving the digital incumbents a significant advantage over rival entrants. Better algorithms may improve services for consumers, which, combined with network effects, can lead to ‘winner-takes-all’ outcomes (in the short run, at least). But not all data is created equal. With a wide range of dimensions to consider—such as the replicability, durability and specificity of the data—it can be hard to determine which aspects of data (if any) represent a truly essential facility for competition. Data itself has little value, either to firms or to the individuals creating it. Rather, it is the insight data provides that is of real value. For example, an individual purchase is of little interest, but a long-term pattern can paint a valuable picture of consumer preferences. However, with different firms using diverse datasets to deliver similar data-based services (such as online ad-targeting), it calls into question the extent to which any dataset is uniquely powerful; and suggests more focus should be given to the role of highly skilled labour to find ways to process that data.
Various consumer exploitation risks have also been highlighted. These range from ‘nudging’ consumers to keep them locked into an ecosystem, to personalised prices and the sacrifice of privacy when using online services. These risks may be exacerbated as technology moves consumers ever further from the ultimate decision-making process. However, the economic implications of these practices are not yet fully understood. For example, while privacy protection can be welfare-enhancing, there are also valuable social benefits from sharing data—such as the optimisation of traffic networks or energy grids—that should not be overlooked.
Consumers’ differing attitudes toward privacy and data ownership further complicate the assessment of what an appropriate share of the value created by data should be. Effective public policy and regulation can play a positive role by maintaining transparency and building trust in the market—helping stimulate firms to innovate and compete, while ensuring consumers are free to make an informed choice as they engage with digital services. GDPR is a step in this direction, but consumers’ behavioural biases (such as simply accepting all cookies to access a website) may be limiting its effect.
Digital platforms’ market power
Digital platforms typically feature strong ‘network effects’, meaning the more users they have, the more valuable they are to everyone. While this can lead to limited competition in the market, it can also stimulate fierce competition for the market. New platforms typically begin as niche players serving a specific need, before expanding to compete more broadly. This market contestability acts as an important check on the power of the incumbents, suggesting policy and regulation should focus on ensuring the ability to enter and grow is maintained.
As well as preventing exclusionary behaviours—such as exclusive or restrictive contracts—this can mean fostering an environment that promotes real innovation. More research is required to properly understand the factors that drive innovation in digital markets. For example, is the chance of being acquired a significant driver of investment? Do antitrust agencies need to be more ‘participative’ to prevent the threat of enforcement from inhibiting new ideas? What can behavioural economics tell us about a new entrant’s chances of success?
Departing from traditional economic definitions, a platform’s power might also be considered in terms of the control it exercises. Some platforms play an increasingly significant market governance role—determining what gets sold, who gets to sell, the rules of engagement and the price to be paid for access to consumers. In some cases, the actions of these platforms have the power to shape markets offline as well as online. This has led to calls for more regulation to set limits on firms’ behaviour and guide the functioning of digital markets before competition concerns arise; as well as for a more ‘holistic’ approach to the assessment of digital platforms, reaching beyond just the consumer welfare standard to incorporate questions of privacy, democracy, fairness and society. Competition policy may not be the best tool to address many of the concerns raised by digitisation in a timely manner.
Competing with data
Much is made of the value created by digital firms using data to better ‘match’ consumers with businesses; and there are undoubtedly social benefits in this regard. However, many traditional industries have long used data to help price and tailor their goods and services. This is leading to a blurring of the lines between markets as digital firms find they enjoy a ‘data advantage’ in certain sectors.
This new competition is forcing traditional industries to adapt, with more personalised offers becoming increasingly prevalent. In many cases, this leads to more tailored products at more competitive prices. However, in some cases—such as insurance, which relies on the pooling of risk—further increases to personalisation could put the fundamental business model under threat.
If regulators and policymakers are to properly manage these risks to business and society, they must assess the effects on a case-by-case basis to identify and address specific issues. While more general provisions (such as GDPR) seek to protect consumers, they can also have the unintended effect of frustrating entry into data-driven markets. The global nature of data collection and processing is also significant, meaning international agreements may be needed to properly shape the use of data. In the absence of these, variations in data policy risks becoming another tool that countries can use to attract multinational corporations.
When considering the appropriate degree and form of data protection, it is important to hold in mind the range of social benefits that data sharing can enable. A traditional role of policy and regulation is to unlock the positive spill-over effects that private actions alone would overlook. For example, sensor data from the Internet of Things could offer many opportunities for society to benefit from wide-ranging data sharing. However, this would first require careful thinking about the interaction of private and public data. Importantly, consumers are often found to hold different opinions on data sharing for social reasons compared with commercial reasons.
One option being proposed that could enable competition while still protecting consumers is collective data trading. With appropriate ownership laws, businesses and individuals could entrust their data to a centralised data broker, which could then provide that data on commercial terms. For consumers, this collective bargaining could increase their ability to extract value from the data they provide; while for firms, it would ensure a level playing field for personal data. However, on top of the legal and ethical issues this raises, it might also run the risk of creating another point of discrimination for consumers. Would ‘high-value’ and ‘low-value’ consumers be compensated equally by data brokers (as they are currently by digital service providers), or would divisions emerge as data brokers established the most valuable types of consumer for their data customers?
Preserving innovation through competition
Preserving innovation in digital markets can mean several things. On the one hand, it can mean promoting innovation on the platform, to improve the range of goods and services available. On the other hand, it can mean innovation of the platform, to improve the way in which those goods and services are being offered. Ultimately, however, consumers gain whenever the overall experience of acquiring and using a valued product or service improves.
The question of maximising innovation in digital markets is set against a backdrop of rising concentration and price-cost margins in all sectors of the economy. The evidence suggests this is due to the increased success of the most productive (i.e. highest-margin) firms in these markets. Combined with the ‘winner-takes-all’ nature of many platform businesses and a virtuous circle of technology advantage, this increasingly results in the complete exclusion of less efficient competitors.
However, with the competition policy toolbox skewed towards assessing price competition and margins, it can be hard for authorities to properly consider the role innovation plays in an agile and rigorous way. There remains a great deal of scope for quantitative research in this area as cases unfold. In the absence of greater economic clarity, courts and authorities are left to make normative judgements about innovation, such as what incremental benefit consumers gain from a new product or service.
For their part, governments and authorities could facilitate greater innovation with a combination of: agile antitrust guidance, providing more certainty for firms navigating digital markets; and ‘patient’ finance, promoting innovation in return for benefits (such as time-limited monopoly power).
Competition policy undoubtedly has an important role to play in shaping Europe’s digital markets. However, the focus for competition policy should remain on maximising consumer welfare through free and open markets that allow new products, services and benefits to emerge. At the same time, competition policy has its limits. The speed and complexity of digital markets mean many issues of competitive dynamics—as well as wider social concerns such as issues of privacy, democracy and fairness—may be more effectively tackled through carefully designed incentives, or interventions such as the Commission’s proposed platform regulation.