Could #digitalization undermine key regulatory assumptions – yet make regulators’ work even more creative? I hazarded two predictions at the Annual Conference on Policy Challenges for the Financial Sector hosted by the Federal Reserve Board, International Monetary Fund, and The World Bank in Washington, D.C., last week. First, digitalization has expanded the range of financial intermediaries but will soon reduce or eliminate some. For centuries, we relied mostly on banks to intermediate our financial services. Today, we conduct financial services using mobile network operators, search engine and social media platforms, plus countless mobile apps. Innovations like “mobile money” introduced new kinds of intermediaries to serve hundreds of millions of previously unbanked people. We may need fewer intermediaries in the future. As more countries experiment with central bank digital currency, some consumers will want to “self-custody” their digital funds just as they hold paper banknotes. “Smart contracts” could assume some intermediaries’ functions, such as broker-dealers’ market-making. Second, regulators will soon license not just intermediaries, but also code. Today, we assume that the incentives and penalties built into financial regulations influence human behavior in intermediaries to promote safe and sound financial activities. However, smart contracts and artificial intelligence (AI) may undermine that assumption. For example, some code does not require a company’s support or human intervention. Bitcoin is a prominent example, and smart contracts that run on the Ethereum Virtual Machine can self-execute without an intermediary. Similarly, as AI-based tools drive more decisions, we don’t know how sensitive those tools will be to safety, soundness, or reputational risk concerns. How will regulators respond? They may seek to evaluate and license not just institutions, but also code like smart contracts and AI-based tools. If a smart contract lacks compliance mechanisms, regulators may ban its use in the market. Similarly, regulators may conduct “fit and proper testing” for some emerging AI-based tools in banks. They may evaluate the training data sets used to calibrate those models, similar to how they assess the education and preparation of human candidates for senior roles. Despite these challenges, the greater use of code in financial services will likely benefit regulators, too. Better technology will reduce the effort they expend today to gather, collate, and manipulate regulatory data. They’ll focus more on the creative challenge of analyzing and interpreting data in the future. Will I get any of these predictions right? Time will tell. Regardless, I enjoyed speculating about these potential developments with fellow panelists, central bankers, and supervisors from over 50 countries. Many thanks to colleagues at the World Bank for inviting me and to the IMF and Federal Reserve for co-hosting the event.
Thanks for sharing, what are the ramifications of fewer intermediaries? Will the licensing of codes come with occasional conditions for example, relating to concentration, interconnectedness, inter-and-intra-relationships, etc. ?
Very interesting, Chris! Food for thought!
Thanks for sharing Chris!
Executive Director at Morgan Stanley
3wI'll keep this in marinating mode. Great stuff. To take a step back— it’s self-evident that Regulators implement legislation; and legislation typically lags innovation. No matter how concentrated a risk is becoming, we also know that Legislation often precipitates only after a crisis, I.e., some sort of severe conflict between two or more constituents’ interests is the catalyst for action by legislators. If innovation is accelerating (and it is), can legislation keep pace? (Or, stated differently: What set of crises will stimulate legislation?). And if legislation cannot keep pace in the absence of crises, how would/should regulators prepare for the first crisis that will catalyze legislation?