Minding the digital economy’s narrowing gaps
Informational asymmetries between buyers and sellers have long been known to impair market performance. But thanks to digital technology and the large, accessible pools of data that it generates, these informational gaps are closing, and the symmetries are declining.
Until recently, market formation has been circumscribed by physical and geographical boundaries. A prerequisite for a market to form is that buyers and sellers are able to find each other, and this process has traditionally been accomplished in physical spaces like bazaars, stock exchanges, stores, or dealerships (albeit with intermediaries using phones and fax machines to facilitate transactions). Things started to change with eBay, the original model for many online marketplaces. Suddenly, geographical boundaries no longer operated as insurmountable barriers between widely dispersed buyers and sellers.
Going cashless
As online marketplaces developed, however, it soon became clear that additional information issues would need to be addressed for these markets to function effectively. For example, because it is difficult for buyers to detect variations in quality among sellers and among goods and services offered online, more information was needed to capture the reliability or trustworthiness of market participants. The problem is essentially the same for both buyers and sellers, with the former worrying about receiving what he pays for and the latter worrying about being paid.
It is precisely this kind of bilateral information asymmetry that prevents market formation or limits market exchange in the first place. Hence, a number of digital-payment platforms initially were created to address online markets’ fundamental “trust” problem. Following the model of escrow systems that are familiar in real-estate transactions, e-commerce platforms created intermediaries that they hoped would be trusted to collect and hold payments from buyers until delivery of the goods or services had been confirmed.
Algorithms and AI
As more and more “stuff” appeared in online marketplaces, users started having difficulties finding what they were looking for, because they could not browse through options in the same way that one does when shopping in a physical store. To address this issue, online platforms developed search algorithms and recommendation engines based not only on individual users’ browsing and purchase history, but also on behavioural data from all other users. These algorithms have been further improved by advances in AI and increases in the volume and quality of data. Search and recommendation engines are a partial solution to the “matching problem”, and thus a key source of online market performance. They add value for both buyers and sellers, and boost transaction volume substantially, especially for lesser-known sellers and brands.
A final informational challenge relates to access, specifically giving consumers accessible online identities and tracking records that signal their attractiveness as counterparties in a variety of market settings.
Credit is a good example. In the offline world, people and businesses have track records and financial histories that hypothetically could be used to underpin credit or insurance markets.
The problem is that these offline records tend to be scattered and inaccessible, whereas in the digital economy — especially following the high penetration of mobile payments and e-commerce — they become easily retrievable and far more useful. Like knowledge, data is non-rival: using it does not diminish its value for further use or for use by multiple parties. AI algorithms can be deployed to assess and price credit for people and businesses with no collateral and little prior contact with the traditional non-digital economy and financial sectors. As in platform-based evaluation systems, informational gaps are reduced and incentives are improved, while market access is expanded for households and small businesses.
In short, data-driven digital markets have evolved from struggling with informational gaps to having higher informational density than their offline counterparts, leaving fewer information gaps and asymmetries. The accessibility of digital data allows for new screening mechanisms and signalling behaviour that are frequently missing in the offline world.
Of course, highly accessible stores of data come with their own real and much discussed risks, and these must be addressed in order to achieve the potential efficiencies and inclusivity benefits on offer.
After all, the institutions (including governments) that collect data and act as digital gatekeepers must be trusted, too. At a minimum, they must be subject to enforceable regulation that provides clear definitions of individuals’ rights with respect to transparency, data use, privacy, and security. Here, arguably, we are making progress, but we still have a long way to go.
Michael Spence, a Nobel laureate in economics, is professor of Economics at New York University’s Stern School of Business and senior fellow at the Hoover Institution