Better Regulating Robo-Advice
‘Robo-advisers’ are online-based advice services that use algorithms to create investment recommendations, often combining this advice with portfolio management. These services deliver financial advice at a fraction of the cost associated with traditional financial advisers. Because robo-advice is delivered online, investors can open robo-adviser accounts all over the globe. This raises issues because financial markets regulation is, despite all harmonization efforts, carried out by national regulators with different supervision approaches and under diverse legal frameworks. In my article ‘Reducing Legal Uncertainty and Regulatory Arbitrage for Robo-Advice’ I discuss some possible approaches to address these issues.
For example, some commentators advocate the creation of regulatory sandboxes for robo-advisers, allowing them to test their business models under an alleviated regulatory regime in constant cooperation with the regulators. However, sandboxes are supposed to facilitate the creation of new business models on a small scale. And robo-advisers are already a vital part of the financial markets. Projections estimate that they will have assets under management worth about US$1 trillion by 2020. This is hardly a market suitable for regulatory experiments.
Another unanswered question is whether the algorithm behind a robo-adviser should be subject to approval by a national regulator. This has been suggested by, for example, German consumer protection groups. The idea is not convincing. Regulators would hardly have the resources to scrutinize thousands of lines of code. Algorithms are evolving, either through machine-learning or because they are optimized on a continuing basis. Regulators would need to analyze and assess every single change, which is unrealistic. Moreover, it is not the role of the regulator to guarantee that the service provided is of good quality. Instead, the regulator should make sure that the business operating the robo-adviser implements procedures for safeguarding the quality of the advice. In addition, the business needs to prove that its staff members have the required skills and expertise to operate, service and supervise the algorithm.
It is my view that commentators and policymakers currently put too much emphasis on public supervision and enforcement of financial markets laws. However, an effective regulatory regime combines elements of public and private enforcement, making use of the particular strengths of the respective approaches. It is also sometimes overlooked that the role of the regulators is to apply the law. The details might differ among countries, but as a rule of thumb national courts are not bound by the decisions of the national regulator(s). It is possible that a court reaches different conclusions about regulatory compliance than does the regulator, leaving the market participant in a difficult position if he had relied on the regulator’s view that was subsequently overturned by the courts.
It is surprising that regulatory arbitrage under contract law has not been paid much attention. The EU framework can be used to illustrate the problem. The public law aspects of the laws on financial advice (for example, authorization for advisers, organizational requirements and ongoing duties) have been harmonized through Directive 2014/65/EU on Markets in Financial Instruments (MiFiD2). In contrast, the Directive does not expressly cover the contract law aspects of financial advice (in particular, remedies for breaches of contract). Thus, there is no harmonization of financial adviser contract laws. EU Member States differ considerably with respect to available recourse and the extent to which private enforcement is engaged. Robo-advisers from a Member State with a lenient liability framework for financial advisers or higher obstacles to private enforcement would have a significant advantage over competitors from other countries. As discussed in my article in more detail, the duties prescribed by MiFiD2 should be enforceable by the client in a civil court. This would be a huge step towards a harmonized liability regime for financial advisers and robo-advisers.
Comparative legal research should establish overarching principles for robo-adviser contract law. These principles should be sufficiently precise to enable their coherent application throughout different jurisdictions without relying heavily on ideas that are exclusive to certain legal families. For example, the opacity of robo-adviser services makes it very difficult for customers to prove that there was a breach of the adviser’s duty of care. Thus, national courts should use their discretion (if any) to alleviate or even reverse the burden of proof for the litigant. In addition, robo-advisors should be liable for the specific risks that come with offering online adviser services. For example, robo-advisers should be liable for the client’s losses caused by server outages or failures of the adviser’s internal systems, even if the adviser is not to blame. In contrast, the robo-adviser should not be liable for general power or internet outages because they are not a specific consequence of offering robo advice, but a general risk for all businesses.
Philipp Maume - TUM School of Management