Should platform fees be capped?

Regulators are increasingly concerned about digital platforms using their market power to charge high commissions and penalize those who trade elsewhere. During the Covid-19 crisis, for example, New York City and other jurisdictions responded by imposing a 15% ceiling on the rates used by delivery apps. Can such restrictions shield smaller businesses and consumers without harming innovation and investment in online services? As part of TSE Digital Center’s efforts to inform these debates by improving our understanding of platforms’ business models, Renato Gomes and Andrea Mantovani investigate how best to cap platform commissions.

How do platforms exploit their dominance in digital markets?

Platforms provide consumers with important information about trade opportunities. For instance, online travel agencies such as Expedia are an essential tool for discovering hotel offers, while Amazon and other marketplaces inform buyers about sellers they might otherwise struggle to find. A similar informational role is played by ride-sharing platforms, food delivery apps, and matching sites for booking restaurants and babysitters. Most of these platforms allow sellers to set their own prices, but charge a substantial commission per transaction. On Amazon Marketplace, professional sellers pay on average 15% per sale, whereas Booking.com fees range between 15% and 25%. Apple charges 30% in its App Store for music, apps, and e-books.

A crucial challenge for this business model is show-rooming, whereby opportunistic consumers use the marketplace to find their preferred seller before exiting the platform to conclude the trade. These consumers can then benefit from a discount as the seller avoids the platform’s commission. To prevent this practice, many platforms use price parity clauses, which restrict sellers’ ability to charge lower prices elsewhere. Platforms also use “anti-steering” provisions, which limit sellers’ ability to inform buyers of alternative sales channels or, in the case of Apple’s App Store, impose an outright ban on direct selling.

How successful have regulators been in responding to concerns about platforms’ market power?

All price parities are forbidden in France, Italy, Belgium, and Austria, whereas in other countries they are prohibited only for certain online travel agencies. In the US and EU, Amazon removed its price parity clauses following political pressure and complaints about anti-competitive effects. In the EU, the recently introduced Digital Markets Act prohibits price parity clauses for gatekeeper platforms and insists that consumers should be allowed to directly trade with third-party providers.

However, it is not clear that removing price parity or anti-steering clauses produces tangible results. Sellers may still prefer to operate on the platform, fearing it may penalize them otherwise. And by top-listing sellers in exchange for voluntary acceptance of price parity, “preferred partner” programs provide opportunities to bypass bans. Transacting directly with sellers can also be impractical due to physical obstacles or inconveniences that are high relative to the price of the good.

How does your paper shed more light on regulators’ options?

We model a situation in which a platform greatly increases each seller’s potential demand by improving consumer information and convenience. If a firm decides not to join the platform, it faces much lower demand as well as competition from all rivals on the platform. Under price parity, this reduces profits away from the platform, and induces firms to accept high commissions. Ultimately, firms might pay more in commissions than the profit gain generated by the platform’s service. Consumers may also be hurt as prices may increase more than the gain from better information.

In light of this market failure, we study a natural policy alternative: a cap on platform commissions. We first consider mature markets, where the platform raises competition and marginal costs but does not expand demand. We show that the optimal cap equals the expected externality that the platform imposes on other market participants. This rule makes the platform internalize the effect that technology improvements have on consumer and firms. The platform then only invests to improve technology if the cost is less than the convenience and informational benefits for consumers and firms. Looking at online travel agencies like Booking.com, we find that a 25% commission fee is only lower than the optimal cap if the platform can triple hotels’ potential demand. This suggests that a cap on commissions might bind in some markets.

We then study expanding markets, in which the platform brings in new consumers and more sales for all firms. The most competitive industries are most likely to benefit, as demand expansion trumps the increased competition. Counterintuitively, the optimal cap may be lower here as, for the same potential demand, the expansion of consumers’ consideration sets is lower. A booking platform should now quadruple hotels’ potential demand for a 25% commission fee to meet the optimal cap.

Finally, we consider platform competition. By tying a firm’s direct-sale price to the fee it charges on any platform, price parity leaves firms with only two relevant options: delisting from all platforms, or joining all of them. As a result, we find that competition between platforms is ineffective in curbing market power, even under narrow price parity.

What other policy implications emerge from your results?

Much of the current regulatory debate focuses on the use of algorithms to favor specific sellers or the platform’s own products. However, our analysis reveals that, even with an unbiased search tool, platforms can leverage contractual externalities across firms to levy excessive commissions. Curbs on recommendation biases and price caps are therefore likely to play a complementary role in the design of optimal regulation. To prevent platforms from using loopholes, regulators will need to ensure that any price cap relates to the total fee charged by the platform.

Quantification of the cap remains the main challenge to implement cap regulation. If not adequately compensated, platforms may cut investment in consumer information. We show that this is more likely to happen in case of a ban on price parity, which would have the same outcome as capping fees at the convenience benefit of a transaction. A cap at this level would be inefficiently low, as it prevents the platform gaining from the informational benefits it generates. By contrast, our policy tool would incentivize the platform to invest, as its ability to charge a higher fee depends on its production of informational benefits.