Response to joint Comitteee Discussion Paper on automation in financial advice

Go back

1. Do you agree with the assessment of the characteristics of automated financial advice tools presented in this Discussion Paper? If not, please explain why.

Whilst FECIF agrees in general with the assessment presented, it also feels that further considerations are necessary. Firstly, whilst it may be understandable at this stage to consider the term “advice” in a common and broad sense, which is indeed the way that many consumers perceive it, this could become a very dangerous approach, for those very same consumers. Much of the automation being seen is quite obviously not advice; at the very most, and only in some cases, it is guidance. In markets where automation is more greatly seen, companies offering such services are often extremely definitive in confirming that it is not advice. We think that it is very important for the consumer to know whether they are receiving advice – which they would see as having third party responsibility and liability backing it – or simply guidance; where, if it “goes wrong” the consumer will almost certainly have no recompense and shoulder the whole responsibility of the decision themselves.

Furthermore, automation should merely be an execution-only programme devised by a computer expert under the supervision of a markets’ operator. The algorithms used to increment the transaction cannot be 100% objective and could be biased towards the interests of the provider and not the customer.
Automated tools may be helpful in the first stage of the advisory process, but in later stages they usually need to be complemented with a real personalised service and the interaction of a human advisor. This form of semi-automation is more likely to shape the future and meets investors’ needs (i.e. investors may input all of their relevant details by means of automated devices, and then they refer to a human advisor).

2. Are there any other relevant characteristics of automated financial advice tools?

Whilst the US is often quoted as an example of how full automation will “take over”, the US market has actually moved in a different direction. The term ‘robo’ has often been used to mean an advice process with no human intervention, but the models which are expanding their usage most rapidly in the US do not really fit that mould. They are half-and-half – they have an automated front-end, where the consumer effectively self-serves in terms of putting in all of their details, and then speaks to an adviser on the phone or via Skype for the advice part. It seems that semi-automation is more likely to be the shape of the future, certainly in the advice sphere. Our earlier comments about ensuring that consumers know if they are receiving “true” advice or not are also, therefore, relevant here too.

Also, the potential problem with the algorithms is twofold:
- One, do they include the whole of the market, i.e. all transactions possible, even during crises such as recent times have shown? Or do they exclude some scenarios? If so, which ones?
- Two, if nothing is excluded from the programmes, how do they choose to execute the order or not, and in which order? Would the “too big to fail” institutions get priority on volume over the end investor for instance?

3. Are you aware of examples of automated financial advice tools being used in the banking, insurance, and/or securities sectors? Please provide examples, giving details of their operating process.

We are aware of a number of automated services presently available. In the UK, for example, companies such as Hargreaves Lansdown, Nutmeg, Fiver a Day, Money on Toast and Wealth Horizon all operate in the financial product arena (securities sector). There are also a large number of comparisons sites for insurance products, with Go Compare being one of the highest profile and longest-running entities.

In the securities sector, we also know the following firms in mainland Europe: Marie Quantier, 1.2.3 OPCVM, les meilleursfonds.com.
In the insurance sector, we are aware of many and we know the following firms well: Yomoni, Funshop, Advize.

Some firms provide tools for others from different professional worlds like Harvest, 1.2.3 OPCVM and Netvox.

No data is available as regards to the algorithms. A lack of official controls could affect the rights of the end investors.

4. Do you offer/are you considering offering automated financial advice tools as part of your business model? If so, please briefly describe: i) what type of entity you are, e.g., long established, start-up, a product provider, an intermediary; ii) the service you provide (e.g. to what extent do you integrate human interaction in the tool you provide?); iii) the nature of your clients; iv) your business model; v) who developed the automated tool (i.e. an external company or developed internally?); and vi) the size of your activity and/or forecast activity?

In France, IFAs are obliged by law and regulation to be able to offer a transaction service to clients. This can include a best execution rule, even if they only take the order and forward it to the provider. Automation, in this context, could alleviate the burden of best execution and, via correct procedures, put the responsibility of dealing back with the investor, not the adviser. Transaction becomes the field of the client, advice the responsibility of the IFA.
Some of our members provide advice services in the securities, insurance and banking sector, and real estate. A great part of their activities integrates both human advice and tools, their business model is varied.

5. Do you consider there are barriers preventing you from offering/developing automated financial advice tools in the banking, insurance and securities sectors? If so, which barriers?

We have to take into consideration some barriers: access to technology and cost for automated tools.
Facing the reality, developing an automated tool by a firm needs considerable effort in R&D or, as we understand from the market, or to shoulder incredible costs because of the integration of companies or technologies. Moreover, most of the efficient technologies are not available without licences.

Again, if this comes to the fore, the implied cost cannot be borne by the IFA, and will no doubt end up being paid by the client.
Firms might update the various automated tools and licences, which could be more expensive than the automated tools themselves.
Finally, regarding cross-border activities, the fact that automated tools must comply with the jurisdiction’s law concerned is obvious, but it doesn’t necessarily seem easy to do so.

6. Do you consider the potential benefits to consumers to be accurately described? If not, please explain why.

Generally speaking the potential benefits are accurately described but a few provisos are necessary we feel. For instance, it is generally true that automation can reduce the costs of delivering a service, but – echoing our earlier comments – this service is not always advice (in fact it very rarely is at present). It may therefore not be a benefit to consumers to access such a service if they felt that they were obtaining advice at a lower cost, to only later on find out that they never really received advice at all. They may have paid less but actually they paid for a service that they did not want and/or didn’t feel they were receiving at outset. That would be detrimental to the consumer rather than beneficial. Also, the assumption that some consumers might have greater motivation to act upon financial matters via an automated tool rather than via a human adviser is only likely in a very small number of cases we feel. There is little empirical evidence known to us that this is the case and considerable evidence to the contrary.

Indeed the question is a trade between:
- transaction costs that could be addressed to some extent by automation (for regular users), and
- advisory services which are calculated according to the difficulty of the tasks.

7. Are you aware of any additional benefits to consumers? If so, please describe them.

Where automation lowers the total cost of advice, but still provides human interaction at some stage in the process, this will most likely make such advice and assistance more cost-effectively available to a wider number of consumers. Moreover, interactions among firms could mean an easier exchange of tools and competences, and the monitoring of firms could be easier for National Competent Authorities.

8. Do you see any differences in the potential benefits arising for consumers in each of the banking, insurance and securities sectors?

Within the banking and insurance sectors there are some products and consumer needs that are relatively simple: opening a bank account and comparing car insurance premiums would be two respective examples. For such products and services automation is often ideal, with few potential detriments to consumers.
In the securities sector, however, most products and/or consumer needs are far less simple and thus the potential benefits of automation are less far-ranging, certainly for the whole process. In addition, potentially detrimental aspects rise in such processes.

Also, even for car insurance and the like, the temptation to devise algorithms that eliminate, by rising premiums, certain clients could be significant. Therefore, a genuine broker will find the best deal, while automation may tend to eliminate “border risks”. There is no additional benefit to IFAs, apart from potentially the French execution-only context, as mentioned above.

9. Have you observed any of these potential benefits to consumers? If so, please provide examples and describe the kind of benefit that has accrued.

The online automation of the comparison of, and access to, general insurance products has brought greater price competition into those markets. One of the main reasons that clients have taken to such services, however, is that there is less administrative burden, no need to duplicate paperwork and documents for the sole purpose of additional, unnecessary and repeated due diligence. This also results in no/lesser delay/s in contracting an insurance policy of this nature or opening a bank account because of over-regulated compliance issues. This is a lesson that could be learned elsewhere in our industry perhaps.
Automation tends to thrive where too much administrative and repetitive burdens are asphyxiating IFAs.

10. Do you consider the potential benefits to financial institutions to be accurately described? If not, please explain why.

The benefits to financial institutions seem to have been accurately described. However, it should be noted that this could result in a non-level playing field between institutions and IFAs in the EU.

11. Are you aware of any additional benefits to financial institutions? If so, please describe them.

Other than the above: reducing competition.

12. Do you see any differences in the potential benefits arising for financial institutions in each of the banking, insurance and securities sectors?

As previously stated, automation is most suited to simpler products and services where a consumer needs less technical or industry knowledge. This would potentially seem to favour the banking and insurance sectors more than the securities sector

13. Have you observed any of these potential benefits to financial institutions? If so, please provide examples and describe the kind of benefit that has accrued.

Online banks have benefited from expanding territories for their clients. From above competition deposit accounts rates they have moved to any area where online transactions can be easily made, with minimum or no advice given and/or charged for. This is more the area of execution-only orders, where the customer feels that cost is paramount.

14. Do you agree with the description of the potential risks to consumers identified? If not, explain why.

We agree with the description of the potential risks to consumers that were identified, but would add to those. In a fully automated process the consumer is entirely exposed to their own behavioural biases, to a far greater extent than when human advice is received. Whilst it is true that the human adviser will also have biases, these will usually be far more tempered and controlled due to their professional and industry expertise, and experience.

In the field of behavioural experience, it is also proven that male and female trades vary considerably regarding risk. Having a machine in front of anyone tends to make them think that it is neutral. However, in real life, a male client will be tempered by a female counterpart, and vice versa. Therefore, dependent on whether a man or a woman designs and programmes the algorithms could, in itself create other real and unpredictable biases.

Moreover, IFAs hold professional indemnity insurance, so in the case of, for example, bad or inappropriate advice the profession is fully liable and customers will be compensated. Electronic platforms, whether their service is called comparison" or "advice" bear the potential to unlimitedly multiply bad or inappropriate advice as well as fraud. At the moment robo advisers are less or even unregulated, so it seems possible that we will read about the first scandals in the newspapers quite soon. Who will then compensate for the damage caused? Witha level playing field any robo provider would have to guarantee that at any time they are financially strong enough to pay in case of default.
Furthermore, the financial literacy of the majority of European citizens is poor. But, qualified use of any "self-service" robo portal requires at least minimum financial competence and mathematical skills. Without this consumers can neither compare nor verify the output they receive from any IT device. The majority of EU citizens could very likely become a victim of inappropriate or even fraudulent offers on the internet."

15. Do you consider there to be any risks to consumers missing? If so, please explain.

We believe that client profiling using a mere algorithmic application with automatic response, devoid of an objective assessment, would involve the risk of a kind of self-profiling by the user that, by trial, may complete the automated procedure, in order to obtain a specific product, without an effective evaluation of the suitability of the choice. In short, investors may repeatedly respond to the various questions in the profiling tests until they get the desired profile (probably unsuitable) depending on the products they wish to buy.

For the sake of effective harmonisation with the standards of investor protection already established for “traditional” financial intermediaries, we deem it necessary that, in many cases, investors do not perform all of the investment process by themselves, with no assistance from professionals. The results of the investment process should be validated by specifically identified experts, in accordance with the best interests of the clients.

Pursuant to Article 24, directive 2014/65/EU (MiFID II) and Article 29, directive 2016/97/EU (Insurance Distribution Directive, IDD), investors shall be appropriately informed in good time as to whether they will be provided with a periodic assessment of the suitability of the products they have been recommended. It is necessary to consider that the investor’s financial situation and needs change over time. In fact, online platforms may not provide for any form of engagement and periodic evaluation of the advice given, or alerts that inform the investors of possible significant changes in market conditions. Instead, the constant human relationship developed with an advisor, allows clients to change/update their investment decisions following the course of their lives and that of their families, as well as market developments.

16. Do you see any differences in the potential risks arising for consumers in each of the banking, insurance and securities sectors?

As stated above, we see potentially more risks where the product or service provided via automation is more complex and thus greater risks will almost certainly exist in the securities sector, in many instances.

18. Do you agree with the description of the potential risks to financial institutions identified? If not, explain why.

We agree in general with the description of the potential risks identified. However, we believe that in general the Discussion Paper does not take into account the regulatory requirements established by the MiFID Directive, which consistently regulates human-based investment services. We believe that the same rules should be extended, with appropriate safeguards and adaptations, to this type of service, due to the logic of the level playing field.

19. Do you consider there to be any risks to financial institutions missing? If so, please explain.

All the above reflections about the implementation of the algorithms open a wide avenue to lawyers to look for any built-in inducements that lead their clients to operate the wrong button/decision in an automated system. We would propose additional risks for financial institutions:
- Risk of project delay: if tools are not ready to start on time, there will be additional costs to pay that could create problems for clients and issue for the credibility of the “brand”.
- Every time tools are changed or are being improved, detriment and subsequent claims could increase because consumers do not know how to use these tools.

20. Do you see any differences in the potential risks arising for financial institutions in each of the banking, insurance and securities sectors?

As stated above, we see potentially more risks where the product or service provided via automation is more complex and thus greater risks will exist for financial institutions in the securities sector, in many instances.

21. Have you observed any of these risks causing detriment to financial institutions? If so, in what way?

We are aware of the delay for the launch of new projects or in setting up new tools. In this case, human intervention will often be necessary and essential for resolving this problem, correcting automated tools inefficiency.
Moreover, automated platforms in many cases simply fail to execute existing EU regulation. For example, IFAs have to identify and regularly monitor their clients with regards to Anti Money Laundering. This requires, by its very nature, the physical presence of the consumer at the beginning of any business relationship and, at regular intervals, face to face contact with the consumer during the contract period. In Austria and many other EU countries this is heavily controlled by the National Competent Authorities. Automation, on its own, obviously removes this aspect.

Finally, automated services very often use or are based in locations that are different from the home states of the consumers, which, if issues arise, could make it difficult or even impossible for consumers to fight for redress or compensation.

22. Would you agree with the assessment of the potential evolution of automated advice? Please provide your reasoning.

As previously stated, we feel that the use of the term “advice” is both potentially misleading and detrimental to consumers, and also summarises the online services presently available in an incorrect manner in many instances (particularly in the securities sector). The assertion that there will be further supply of automated “advice”, based in part on developments in the US, is inaccurate we feel. Our research suggests that there is likely to be two main developments in the near future: (i) continued and increasing growth in the automation of parts of the advice process, and (ii) increased growth in the provision of online services that provide information and, in some cases, guidance but effectively assist the consumer to self-service themselves.

One of the driving forces behind increasing automation is obviously because, for both financial services operators and consumers, it tends to reduce the excesses of regulation: less and simpler paperwork due to the processes being electronic for instance, which generally makes them cheaper and quicker. There is a general trend at all levels of society for an “Uberisation” of business activities because it can be cheaper and easier for both operators and consumers. Automation in financial services is a further example of this – but we need to ensure that it occurs within a level playing field of regulation for all advice distribution channels.

23. How do you think that the market for automation in financial advice will evolve in the near future in the banking, insurance and investment sectors? Please also provide details of any relevant data or information to support your views, where available.

Please see our various responses above.

24. Are there any other comments you would like to convey on the topic of automation in financial advice?

At the risk of repeating ourselves, we see a number of key aspects re automation within our industry:

(i) Increased automation in general, across all three sectors.
(ii) Increased automation in parts of the financial advice process. Greater use of automation and technology by human advisers to deliver advice in a more cost-effective manner, potentially to a wider consumer base.
(iii) A need to ensure that automated services are appropriately regulated and overseen to avoid consumer detriment.
(iv) In the (possibly few) instances where advice is actually provided and this is achieved entirely by automation, the regulatory protections around that advice should obviously be as great for the consumer as with human advice.
(v) A highly important need to ensure that consumers understand what service they are receiving. Is it advice in a legal/industry sense, backed by appropriate qualifications and expertise, professional indemnity insurance and regulatory oversight – or is it really just information or, at the very most, guidance that is provided without any of those consumer protections.

Name of organisation

European Federation of Financial Advisers and Financial Intermediaries (FECIF)