Are Preset Defaults Harmful?

In the search engine industry, there is intense scrutiny - and criticism, investigations, lawsuits and adverse judgements - of the practice of “preset defaults” wherein operators like Apple control entry points (such as iPhones; or Google with Chrome browser) and steer users’ choice of search engine towards the designated default product. With a theoretical model which captures the essence of preset defaults, we produce several insights regarding the regulation of digital markets. While preset defaults are alleged to create and perpetuate dominance, we find that their economic implications are quite nuanced. First, preset defaults have two differently-acting mechanisms that tilt consumer adoption towards the designated firm: i) the inconvenience cost of switching out from the default acts in favor of the designated firm and against the rival, but ii) their influence on “sheep” users who blindly adopt the designated default can elicit less competitive behavior by both firms. Second, preset defaults are more harmful for products with weak or no network effects than for digital products with strong network effects: the direction of outcomes depends significantly on the magnitude of the product’s network effects. Third, short term effects (monetization level or extent of advertising) are more negative than long-term effects (impact of defaults on innovation and quality). Preset defaults do increase market strength of the default firm and make it more difficult for the rivals to compete, however abolishing them or weakening their power may ultimately not benefit consumers, especially when the default firm indeed offers the better product. For products with strong network effects, preset default can increase consumer surplus: the default arrangement’s positive impact on the designated firm’s market share (and increased confidence that the market will tip in its favor) incentivizes the firm to innovate more, knowing that its investments are more likely to pay off. Our results also shed insights on many other applications where preset defaults are common, including PDF reader, web browser, and the next battlegrounds for computing: large language models (LLMs) and chatbots.

This seminar will take place in-person in room T09-67. Alternatively, click here to join the seminar online via Teams.