Deutsche Bank

We understand that the ESAs have taken a cross-sectoral (banking, insurance and securities sector) approach to automated advice in the discussion paper. However, all our comments are limited to the securities sector, as this is the area where we have particular interest and expertise.

We broadly agree with the assessment of the characteristics of automated financial advice tools. We understand that the definition of advice is seen from a customer’s perspective for the purpose of this discussion paper and may not be a reflection of the ESA’s future policy intent. However, we still remain concerned about the lack of clarity linked with the definition of ‘advice’ and believe that there is a strong need for this to be addressed at a broader EU level.

The definition of advice should be linked to the type of output generated by automated tools and that a clear distinction should be made between discretionary portfolio management, investment advice and support tools for self-directed investment decisions. For example: we believe that the following tools assist the customers in making investment decisions (on their own responsibility) and may not necessarily constitute advice:

- Search engines displaying product results based on customers’ self-defined parameters - such engines allow customers to preselect products from a provider’s universe of products, using simple selection criteria. This is also made sufficiently clear to customers by the tools’ design and operating process. We do not believe that any results from such tools not constitute advice because the results displayed are based on what the customer wants to see and not on any specific information submitted by the customer. Such services should not be subject to the more stringent requirements governing the provision of advice.

- Proposed portfolio structure – certain automated tools propose only portfolio structures with diversified asset allocation (based on inputs from the customers) instead of proposing specific products. Since customers are only recommended a portfolio structure, such tools may not constitute investment advice.

In recent years, guidance on the definition of investment advice has been issued by various European authorities such as the European Securities and Markets Authority (ESMA) and its predecessor, the Committee of European Securities Regulators (CESR). However, in the context of automated advice, we would appreciate further guidance from the ESAs that would not only help firms develop new service models but also encourage customers to use automated advice – thus increasing retail investor participation in the EU.
Divergent national implementation of EU MIFID 1 suitability requirement creating obstacles for development of Discretionary portfolio management services

While several fintech start-ups have launched automated advisory services in various countries in the EU, banks in certain other EU member states are hampered by a more stringent implementation of the EU Market in Financial Instruments Directive’s (MIFID’s) suitability requirements.
For example in Germany, rules require banks to prove prior investor knowledge / experience in investing in relevant products before offering a portfolio solution. This implies that after collecting relevant customer information through the questionnaire and designing a customised portfolio, banks have to go back to the customer (often face-to-face) and probe them on any previous experience of investing in products included in the portfolio. If proved otherwise, the portfolio has to be redesigned – this often becomes an iterative process. Thus instead of suitability of the overall portfolio, suitability of the individual instruments needs to be considered, which results in suboptimal portfolio settings and impacts smooth customer experience.

We agree that even though automation in financial advice is not presently observed equally across all EU member states, the phenomenon has the potential to continue to grow and believe that the ESAs can play a role in facilitating this growth across the EU. Harmonisation of rules (eg MIFID) across the EU is one aspect that can encourage firms to offer services across all EU member states. Traditionally many EU member states have taken a rules-based approach to implementing regulation. While this may work well for human advice, for automated advice, a more principles-based approach is desirable. We urge the ESAs to review the existing suitability requirements across EU member states and provide guidance on how these can be harmonised across the EU and encourage national regulators to adopt a more principles-based approach to regulation, which will support EU firms to launch new products or expand existing services to all EU member states.
The discussion paper notes that automated tools enable a wider range of customers to access investment advice. We are of the opinion that both automated and human advices are complementary to each other. While automated advice can assist customers make tailored decisions based on their profiles, human advice can help customers define their investment goals and how much they should be saving to achieve those goals. If customers are offered both services, they can make a choice based on their preference and needs.
The paper further observes that automated services provide advice in a faster, non-time consuming way. For established firms governed by strict conduct regulations, this may not necessarily be accurate. In the EU, suitability requirements under MIFID I (and MIFID II in future) require firms to undertake a comprehensive suitability assessment of clients’ personal and financial circumstances. Clients sometimes pursue conflicting financial objectives, and these have to be discussed and reconciled, before they could be offered a solution. However, this again is linked to whether the output of the tool is considered to be an advice or not (our response to Q1 covers this in more detail).

We disagree with the observation that ‘a well-developed algorithm may be more consistently accurate than the human brain at complex repeatable regular processes, and in making predictions’. While automated advice may be free from any cognitive biases, the experience and responsiveness of a human advisor cannot be underestimated. We also do not believe that the term “personalised feedback” can be attributed to automated advice.

In summary, we believe that a balanced view needs to be taken on the benefits of automated advice, which can complement human advice.
The discussion paper notes that ‘financial institutions may be exposed to litigation and subsequent reputational risk due to faulty automation’. We agree with this observation but that this risk is not limited to automated advice only, but apples to human advisory as well. This also reinforces the point we make in our response to Q1 that it is important to define the point at which an automated proposal in relation to a given product is deemed to be a recommendation and advice in legal terms.
We wish to highlight that so far the focus of financial institutions has been on innovations which can be accommodated within existing legal and regulatory frameworks. For these institutions to make continued investment in technology and launch new automated products, regulatory and legal frameworks will need to be adapted to facilitate the delivery of new products and services through digital channels. Without an appropriate regulatory architecture, providing a level playing field for market participants and ensuring appropriate protection for customers will be difficult.
The paper notes ‘If providers of automated advice tools also offer consumers the possibility to engage with a human advisor as an alternative means to obtain advice, consumers may overuse that alternative means so as to supplement the automated advice on the product/service’. We do not see this as a risk and as noted above, see automated advice as being complementary to traditional human advisory services. We strongly believe that customers would always want an option to choose between automated and human services and in certain instances combine elements of both for an optimal experience.
We broadly agree with the ESA’s assessment of the potential evolution of automated advice, particularly on the observation that usage of automated financial advice tools though currently limited to particular EU jurisdictions, growth in this market is likely. However, for financial institutions across the EU to make investment in this space, further legal clarity is required on certain elements such as definition of advice and regulatory frameworks need to be adapted to facilitate delivery of new products and services. We laud the efforts of the ESAs in assessing the benefits and risks of automation in financial advice and fully expect them to partner with the industry to ensure continued dialogue and take appropriate actions to support growth and mitigate risks.

In conclusion, we wish to highlight that automated financial advice offers an opportunity for the EU market participants to provide cross-border investment services across Europe. Human advisory services usually trigger creation of bank branches in member states. Automated advisory services, offered over the internet, cross-border, potentially reduce the regulatory complexities linked with launching branches. Furthermore, such services can be offered in a more scalable way which reduces cost. US based automated advice providers benefit substantially from a large market with unified regulations, which is further amplified by more than 300 million consumers having direct internet access. EU participants have not been able to mimic this growth in spite of similar opportunities.

The ESAs should develop a template for the development of an EU-wide Robo advisory framework with harmonised suitability requirements. This may take time to deliver, but will be an important pre-requisite for further development of services in this space.
Tomo Ishikawa