As the authors of this paper – which can be found here – say in the abstract: ‘Many of the largest and most successful businesses today rely on providing service at no charge to at least a portion of their users. Free services often delight users, yet also create a series of challenges for competition policy, including impeding entry, inviting overproduction on quality, and increasing the risk of deception and overpayment. This short paper presents these problems, examines the strategies that entrants can attempt when competing with free service, and considers possible regulatory responses.’

The paper’s main message is that, while free services have undeniable appeal to consumers, they can also impede competition and market entry. Competition authorities should be correspondingly attuned to allegations arising out of “free” services and should at least enforce existing doctrines strictly in affected markets.

The paper is structured as follows:

A first section provides an overview of businesses offering free goods or services.

Some business models offer free services to one side of a two-sided market—relying on the now familiar logic that users come to the market to connect with each other and that free services to one type of user can increase the number of such users, creating greater value (and ultimately allowing higher prices) to another type of users. More recently, this approach has become prevalent for a wide variety of online services such as search engines, social networks, and other online media. However, free services need not be limited to two-sided markets. Firstly, some firms justify free services via the data they collect (such as collecting email addresses and/or telephone numbers in return for providing free Wi-Fi). Secondly, some firms provide free services as a trial to pull in users who might later pay for additional benefits (so-called “freemium” businesses). Thirdly, some firms offer free services but simultaneously solicit donations (e.g. Wikipedia and some online newspapers).  Fourthly, other firms offer free services primarily because they operate with extremely low costs (e.g. Craigslist).

A second section describes the competition economics of free.

Price is the most basic and most common form of competition. Trade-offs between price and quality support well-functioning markets: often, a rival can enter a market via a somewhat lower price at a level of quality equal to the incumbent, or via a sharply lower of price at a somewhat lower level of quality. Positive prices thereby operate to make market entry easier. A new entrant can try to offer a similar product or service at a lower price, a natural and widely-used entry strategy, which in turn constrains incumbents’ strategy. Further, prices convey genuine information about the true cost of a good or service, allowing allocative efficiency.

In a market where a firm for whatever reason offers service at a zero price, any price competition would require competitors to embrace negative prices, as is indeed the case in many multi-sided markets. However, negative prices usually present important challenges to firms.  If a firm pays a customer to take its product, the customer may discard it rather than use it, providing no benefit to users on the other side, no benefit to the firm, and often only an exploitative or fraudulent claim to payment.

 Despite these challenges, markets with zero prices bring some important benefits to consumers. Firstly, with price set to zero, consumers avoid paying for valuable goods and services. This benefit is particularly valuable to low-income consumers who can nonetheless use world-class search engines, online encyclopaedias, and news content. Secondly, if a service is provided for free, competition will tend to focus on quality. Free markets thereby accelerate quality improvements and innovation, precisely because quality becomes the main, if not the only, parameter of competition. Thirdly, zero prices may encourage users to multi-home by installing multiple apps and using multiple services. Fourthly, free services can play a useful role in a multi-faceted pricing scheme such as “freemium” business. When a firm offers basic services at zero price, along with one or more improved tiers at higher prices, the free version helps customers try the service and assess quality and match.

The authors then turn to the challenges created by zero prices.

A first concern is that free services tend to undermine paid options and give rise to negative externalities. For example, to the extent that free news services provide lower quality news and reduce the revenues of higher quality news providers, any positive externalities from free news are correspondingly reduced. Further distortions arise if consumers are in some way shortsighted or confused about the benefits of paid offerings. For example, paid search services might provide more accurate recommendations which are less tainted by commercial concerns. Paid information sites might protect readers from the distortions of advertising. Lastly, by reducing tracking and associated information collection, all manner of paid services could increase privacy.

Secondly, the provision of free services may impede entry when combined with network effects. As pointed out above, zero pricing invites competition on quality. In the best case scenario, this can push users towards products and services of distinctively higher quality. At the same time, free services may present material impediments to competition. For one, many two-sided markets provide benefits to one set of users at zero price thanks to revenue that comes from other users. An entrant typically cannot beat “free” offerings by lowering prices because negative prices are often infeasible. Nor can an entrant easily find fee-paying customers on the other side, for lack of a sufficient number of customers on the first side of the platform. In such markets, the combination of free services and network effects makes entry distinctively difficult.

Thirdly, free services invite overproduction on quality. Consider markets where advertisers’ payments support free service to consumers. One business uses funds from one size of the market (i.e. advertisers) to produce a shiny, high-quality bound yellow pages; another charges significant less from the advertisers and produces a less shiny phone directory, achieving the same profit margin as the competitor. However, customers will naturally choose the shiny version of the phone directory, even though the benefit provided by both directories is the same and the social cost of producing the shiny version is higher. Nor is this market structure in any way limited to telephone books; search engines and other online advertising share these incentives. This suggests that markets with this structure systematically tip towards overprovision of benefits to customers above the efficient levels, needlessly driving up final goods prices. Notably, free services on one side of the market can even drive out rivals and push towards a single dominant firm. Returning to the “yellow book” example, notice the challenge faced by the second book publisher—facing little consumer adoption, and hence correspondingly lower ability to attract advertisers, it will likely allow the first book publisher to become dominant in the market. This broadly tracks experience in online advertising markets, where dominant platforms have been stable for some time and where entrants and smaller firms have largely struggled.

Fourthly, free services create risks of deception or exploitation. A consumer paying for a good or service can directly assess the appropriateness of the fee relative to the value provided. However, in markets with zero prices, consumers make no such trade-off evaluation: a service might collect outsized value from using customer data, showing ads, or making other use of a user’s activity, and the user would not know.

A third section looks at plausible entry strategies for competitors into free markets.

Despite the challenges listed above, market entrants do have some potential strategies they can adopt. These strategies, however, all have high costs or low probabilities of success, making market entry infrequent. Furthermore, there are also doubts about the feasibility of adopting these strategies in the face of today’s large and entrenched incumbents.

First, an entrant could invest in trying to replicate the customer base of a dominant platform. For the very largest and best-funded entrants, these strategies seem to be possible, though at exceptional expense and risk. Secondly, an entrant could attempt to outcompete incumbents through sharply increased quality, which would be sufficient to induce switching even if there is no financial incentive to switch. Again, such sharp increases in quality are naturally unusual, making entry less common. Third, an entrant could try to offer an entirely new service that does not directly compete with any incumbent, and thus is not vulnerable to the difficulty of undercutting a free incumbent. This strategy seems to be viable and, yet, only rarely does an entrant devise an entirely new type of offering, of broad interest, with potential far-reaching effects. If entry is limited to these unusually successful new firms, entry will be predictably rare. Fourthly, entrants may find an opportunity when consumers tire of incumbents’ service despite the appeal of “free” – but, here, again, entry will be infrequent. Fifthly, an entrant may see an incumbent’s free service, and respond by charging a positive price and eliminating whatever is unwanted or annoying about the incumbent’s service. These subscription models sometimes get traction, but usually face significant challenges.

A fourth section looks at possible policy responses.

Free services create a heightened need for antitrust enforcement. Because entrants face additional challenges in entering markets with zero pricing, there is a particularly strong case to prevent dominant firms from erecting further barriers to entry or eliminating small rivals. Secondly, policy-makers should take seriously the non-price harms that consumers may suffer. Thirdly, in some instances policy-makers might consider mandating positive prices, e.g. by requiring social platforms to offer paid and free versions of the service, thereby avoiding concerns with regulators setting prices. Fourth, policy-makers might be increasingly sceptical towards mergers and acquisitions, particularly of small potential competitors. Fifth, policy-makers could insist that traditional non-competition remedies remain available.


This is a useful, short primer on the economics of free. Some would undoubtedly argue that it focuses excessively on potential anticompetitive effects of such practices, without putting sufficient focus on the benefits that free services provide. Personally, I think that this is a matter that needs to be assessed in individual cases, at least in the competition law context, before general rules are extrapolated.

Further, and acknowledging that this is only a short primer on the topic, the authors do not engage with some important questions raised by the solutions they propose. For example, how does a concern with eliminating further barriers to entry or eliminating small rivals fit within existing antitrust doctrines? In addition, I do not think that anyone argues that non-price harms are not relevant: the challenge was, and remains, to identify them (e.g. is privacy a cognisable competition harm or not? What if there is social regulation protecting privacy but there is evidence that consumers do not value it as much as lower prices?) and to quantify them so as to balance them with price benefits. Lastly, and at the risk of repeating many of my earlier posts, the authors do not address what should be the balance between competition and other regulatory tools in addressing the challenges they identify. This paper may not be the place to deal with these questions, but I hope the authors engage with them in greater detail elsewhere.

Author Socials A weekly email with competition/antitrust updates. All opinions are mine

What do you think?

Note: Your email address will not be published