Allianz SE

While many automated financial advice tools possess the characteristics outlined in this section we disagree with important specifics. In particular
o we understand that only customer-facing tools (i.e. B2C tools) are in scope of the DP, while several hybrid formats (including at least some human interaction) are mentioned in later sections. We would like to point out that it is important to achieve a level playing field among all forms of advice to avoid regulatory arbitrage at the expense of the customer
o the definition of “an algorithm” to link consumer inputs to outputs may be too static (as the simple decision trees mentioned in section 21 on page 12) since many new developments (esp. machine learning approaches) contain aspects with continuous improvement of such algorithms over time. While this may pose additional challenges, regulation should not prevent any innovation in this domain.
o we strongly oppose the excessively broad definition of “advice” as defined in the CP (e.g. in sections 12 and 22), which in effect qualifies any kind of information which helps customers to select a suitable product as advice, even if it only makes a minor contribution to such selection or is merely “perceived to be” advice (see e.g. section 22). This potentially opens debates and open-ended disputes about customer perceptions, especially in cases where only minor orientation is provided. We consider the very consistent definition of advice as (clearly marked) personal recommendation across all relevant acts (MiFID, IDD etc.) as a much more suitable concept. Any other information and marketing material should also be held to certain standards (e.g. be fair, accurate and not misleading, as indicated in many acts), but care should be taken not to attach excessive obligations to each and every kind of information provided, because this may greatly limit the willingness to provide any such service in the first place. This could harm the well-understood best interest of the customer.
o By contrast, we suggest a balanced approach which clearly differentiates between “advice” in the narrow sense and broader marketing and other information or even generic algorithms, that merely assist to narrow down the relevant set of products.
There are numerous new tools currently developed in the markets to offer and market financial products based on automatization. Namely such tools which combine digital offering with transition to human advice, be it in person or via video. Such hybrid models try to be as flexible as possible to allow for switching between the models on demand: e.g. after a first meeting in person, the customer may want to verify specific questions and getting further advised in a digitalized framework. It should be well understood that such developments of future design of automated advice based on clients experience and preferences will continue to develop further.
There are various tools and provider active in the digital markets, which fit the very broad definition of advice used in the DP, for example,
o manufacturers websites contain certain elements which help the consumer in his or her product choice and may or may not recommend a single product but narrow down the product choice price comparison websites for insurance products are typically legally set up as brokers and primarily use customer input to solicit bids from different providers which they present to customers, typically ranked by price and with additional specific information, e.g. on coverage, limitations in in claims settlement.
In any case, we would like to reiterate our statement that the definition of advice as any perceived narrowing down of the relevant set of products (see sections 12 and 22 of the DP) is overly broad. Instead, we advocate sticking to the narrow definition of advice as personal recommendation (as in all relevant acts) and apply the most stringent standards there, while limiting the requirements for less far-reaching assistance provided in narrowing down the product choice.
• Allianz is actively offering automated financial advice tools for end customers in addition to existing human based advice in several markets.
• In many markets web based product information is already available; there is no digital concept yet in place which would operate under the current narrow legal term of financial advice as given in the relevant acts (MiFID, IDD etc.).
• In any case, we anticipate that we will substantially broaden our digital approach to providing financial service in many forms in the future.
• In order to enable industry to develop suitable tools for automated financial advice, the regulatory framework should be further clarified to enable legal certainty.
• Firstly there needs to be an unambiguous and not too broad definition of “financial advice” e.g. compared to pure “information” or marketing, each of them properly governed by transparent rules to enable customers to understand the quality of service they may rely on.
• Consistent regulatory regimes both across sectors as well as cross border are essential to create legal certainty as well as a level playing field across providers and market participants in order to prevent regulatory arbitrage;
• Level playing field should promote scale and cost efficiencies across sectors and cross border and therefore allow for clients access to most suitable products at fair pricing.
• Yes, benefits are generally described correctly.
• It should be well understood that these are potential benefits which may not necessarily materialize in any circumstances. They may in addition go along with another quality of inherent risk: e.g. section 35:”online automated tools may present information to consumers in a short and digestible way” – such information may be less comprehensive, generic and not fully capture the client’s needs.
• As further benefit, advice generated by a recommendation engine which takes into account real-time data on financial markets or other relevant aspects, such as other choices from similar customers (see section 38) will typically be less stable over time than products which rely on more static data as a basis. This in turn may complicate reproducible documentation (see sections 39 and 44).
Overall, automated advice tools may be employed beneficially for consumers and facilitate decisions in many valid ways, e.g. by
o simplify comparison of similar offers in one dimension, e.g. price
o reducing complexity without burdening the consumer with the intricate mechanics of the comparison itself (i.e. “under the hood”)
It is not straightforward to assess which approach will yield the most benefit, which may vary over time and even include combinations, since the approaches are non-exclusionary.
Since automated advice holds substantial potential both to reduce complexity of the decision and to facilitate comparison among different offers it holds high potential for customer benefits in many areas. Whether these potentials are realized substantially depends not only on customer acceptance but also on rigidity of regulation.
We therefore ask that the regulatory framework must allow for innovation.
Price comparison websites typically facilitate price reductions for customers, whereas it is not clear, whether quality (including service quality) remains the same
It also seems that in many countries which adopt rigid conduct regulation regimes such as the Retail Distribution Review (RDR) in the UK automated “advice”, or at least information, selection and acquisition platforms seem to remain the sole cost-effective offer type for some customer segments, in particular at the low end of the income spectrum (also see sections 31 – 33 of the DP).
Broadly yes. It should be noted, however, that the benefits listed in the DP are possibly based on an overly simplistic model which may not accurately reflect all relevant types of financial advice tools. In particular, many mechanisms that help to orient even the perception of the customer in the offer space potentially fall under the overly broad definition of “advice” (see sections 12 and 22).
• Reduction of liability risk resulting from “human factor” (lack of documentation)
• Better documentation by digital tools both details provided by consumer as well as advice and product information shared with client.
• More stringent tailoring of offerings and distribution efforts to niche markets and needs,
• Automated advice tools potentially address different distributional challenges and therefore may be broadly employed for many products and product types.
• For example: Automated advice can facilitate
o the tailoring of offerings (incl. products and services) to very special needs, thereby addressing niche demand
o a broader reach into previously underserved market segments with standard products (based on lower price of access and / or superior access)
o simple products may be sold more due to lower reservations of customers to access such products over a fully-automated platform
o complex products may be sold more due to the (perceived) reduction of effective complexity of the decision process (which takes place “under the hood”)
While there are for all three sectors product categories which are suitable for automated advice or execution only acquisition, we see that for more tailor made, long term or personalized products (e.g. health insurance) the human factor of analysis and advice will continue to play a major role for tailor made distribution.
There is broader access to customers via internet based offering since clients can access information and ordering tools outside opening hours at any time of their convenience.
Financial literacy of customers is gradually increasing. Customers who may have collected certain experience and possess a certain financial literacy therefore can be expected to feel more and more comfortable with automated advice, delivering a higher potential market reach for providers of automated solutions.
• We consider the description being not complete, see Q 15;
• It should be well understood that the whole digital marketplace, its instruments and techniques are in a very fast evolutionary process, which by its nature involves potential risks, both to consumers as well as financial institutions and service providers.
• A wise supervisory approach should carefully observe this evolution, giving room for reasonable gathering of experience and common learning before applying to early judgement.
• In order to support and enable innovation, firstly a principle based approach seems to be more promising than any detailed, prescriptive tick-box control-approach.
Bias of incomplete needs analysis may lead to products suggested, which do not fully meet the client’s needs. Based on transparent segregation between “financial advice tools”, which should provide recommendation only based on sufficient detailed client data, any other informative tool for execution only should respect the customers’ personal willingness to investigate options based on more or less detailed options.
Regulation should not be over-protective and prescriptive but rely on the inherent opportunities for increased financial literacy of consumers using digital information tools before taking financial decisions.
While the general risks are similar across all sectors, there are some particularities which render some risks more or less important in the sectors. Specifically
o (capital) market risks are most significant for investment products, which let the customer participate in that risk
o while individual consumer data may be important for suitability tests in all products, in insurance the consumer individual data are typically relevant for pricing and underwriting processes.
o Tax aspects may play an overall more important role in all those areas which have different tax treatments based on product design.
Clients using internet tools may take their decision to acquire a certain product under the impression of having received financial advice (in the legal sense) while the provider would only act as “provider of information” – the risk that the legal quality of a digital offering (information vs advice) is not fully transparent to the customer may lead to disappointment and dissatisfaction of consumers.
Given the relatively new collaboration of financial product manufacturers with Fin Techs, digital service provider and other financial market participants, the concept of fair and transparent allocation of duties and liabilities may give rise to dispute related costs.
Technical dis-functioning of digitalization or potential Cyber-crime may cause further detriment to financial institutions.
• Technical risks (e.g. with respect to IT security) with all the possible detrimental impact are very particular to platforms providing technical advice.
• Currently the risk of legal ambiguity of terms (information vs advice) is one of the major obstacles, causing liability risk and reputational risk
• Currently it seems that the inherent value of a “regulated financial advice” cannot be identified by the customer’s in comparison to other non-advisory services. For example, comparison websites may actually be perceived by consumers as providing financial advice in the legal sense.
While the general risks are equally applicable to all sectors, there may be particularities with respect to the typical risks involved with respect to particular products, e.g.
o the amount and quality of data required to provide the product, e.g. with respect to objective data (such as income) or subjective data (such as risk preference) for investment advice vs. little information need on a minor insurance cover
o handling of critical personal information on the customer, e.g. for insurance quotes or credit scoring
Too early to assess
• We consider the evolution to find its own dynamics and be very fast.
• We can imagine that there might be start-ups / Fin-Techs developing good looking tools for “financial advice”, attracting clients but without sound organization behind to grant sustainable financial protection.
• Consistent application of existing regulation to all market participants in digitalized financial markets is key to ensure level playing fields and sustainable consumer protection in this new area.
• Automation of advice may be based of simplification and categorization of people and needs; the high level of standardization might may lead to un-suitability of the advice for consumers who deviate from standards. – We will see whether (to the extent e.g. self-learning systems are allowed to be integrated in product design and sell processes) future digitalization will lead to even more detailed and customized servicing. This is a field where be strongly recommend to allow for optimistic gathering of experience before drawing conclusions for any regulatory action.
We assume that there will be strong debates about roles and responsibilities of different market participants and tool providers.
Regulators should take the initiative to capture automated advice similar as any other kind of financial advice under their regulatory regime and ensure level playing field in application of existing framework.
• Allianz appreciates the initiative of ESAs to consult on automation in financial advice. In this field of high speed developments, consistent application of adequate regulatory framework to all market participants is highly welcome to ensure ongoing client satisfaction and consumer protection at an level playing field to traditional human advice or any upcoming hybrid models.
• We should respect that there will be customer groups who still prefer receiving personalized human advice, be it in the traditional form or in hybrid models, where clients first research information on- line and subsequently ask for personal setting, which for example may as well be via skype or other future means.
• Subject to continuous harmonization of regulatory framework for automated financial advice, i.e. no national gold-plating, we assume that digitalization will support access to variety of products for consumers, namely in rural areas or smaller countries of the EU.
• Regulator should take into account that the current status in which the definition of financial advice not harmonized across financial sectors nor adapted to an automated environment creates ambiguity and legal uncertainty in many respects. This is also not transparent to consumers.
• To fully protect consumers, it may be advisable to clearer distinguish between “information”, “marketing” and “advice”. To this end it could be required that those digital providers who merely “inform” make the limitations of their service transparent and have appropriate disclosure in place. Where “merely informative” comparison websites forward to providers and receive financial compensation for such forwarding, this should be made transparent to avoid that consumers may think they would take their financial decisions based on neutral information only.
Dr. Catharina Richter
A