MoneyFarm SIM S.p.A. and Legália Studio

Broadly, we agree with the assessment of the characteristics of automated financial advice tools outlined by the ESAs.
In detail, with specific reference to the statements contained in the DP, we would like to share some considerations:
- “The automated tool is used directly by the consumer, without (or with very limited) human intervention”, the level of automation of the service offered by MoneyFarm may vary depending on the determination of the investor: the user is in the condition to complete the process to obtain the advice without any human intervention; at the same time, the user can request human intervention during any step of the process, while MoneyFarm never asks the user to provide his/her contact details in order for MoneyFarm to contact him to arrange a personal conversation.
In any case, the intervention of the individual advisors is confined within specific tasks (aimed at clarifying doubts along the questionnaire and within the scope of activity of MoneyFarm) and cannot concern broad financial advice on any topic.
As a general remark, our market experience shows that quantitative-only (or “fully automated”) investment systems usually do not achieve resounding success, and many institutions are moving towards automated system based on a qualitative approach, which provides a human intervention during the process.

- “The output of the tool is, or is perceived to be, financial advice”, the notion of financial advice is key.
Today, MoneyFarm does not issue straight recommendation to the individual investor to enter into a specific transaction and the algorithm is only used to assign the client to one of the 6 available risk profiles, each of which corresponds to a pool of assets (a bucket of ETFs). The financial products allocated to each bucket are selected by human advisors periodically (once every two months or more often is required) and regardless of the identity of the individual investors (who will be directed to the basket that best fits the profile generated). In other words, the recommendation is not a truly personalized indication to invest in a specific product chosen in consideration of his/her individual situation, rather a guideline towards a pool of assets that may fit the requirements of a category of investors. Therefore, the actual advice given may be deemed as a lighter form of advice compared to the narrow definition under the majority of EU legislation.
In respect of the description contained in the DP, MoneyFarm may be therefore considered as a risk-profiling-tool that captures information to categorise consumers, for example by risk profile and investment horizon. In accordance with point 24 of the DP, the advice provided may be the same as every other consumer who is in the same category of investor (i.e. with similar risk profile and investment horizon).
In addition, in line with point 25 (“the level of specificity will also depend on the universe of products/services considered by the tool”), the level of specificity of the advice rendered by MoneyFarm should be deemed limited since the algorithm only considers a limited range of products (ETFs) when providing the final advice.
To this regard, in order to avoid deterrents to the smooth development of such a positive phenomenon, it would be helpful to consider a modulation of the regulation according to the type of advice provided, as explained below (see under 24 below).
This solution would make easier to encompass future changes in the financial activities that will be automated. To give a practical example, MoneyFarm is undergoing a reorganization that will expand the company business to include forms of individual asset management, which may require a higher level of control compared to the current activity. In such scenario, imposing the same level of regulation at the initial stage would be likely to hinder the prospect of growth of the company in the future.
MoneyFarm provides digital wealth management solutions to clients in Italy and UK. The company has been operating in Italy and UK and offers both discretionary and non discretionary services. We have a model driven customer scoring and portfolio allocation that are integrated by remote CRM (to support the customer scoring) as well as an investment committee (to support the model portfolio allocation). Both models have been developed in house.
We have identified a strong business opportunity to provide an alternative investment solution to those investors who have found traditional financial advice services unavailable or uneconomic or who have been dissuaded by poor investment manager performances and traditional management fee structures.
Our business objective is to offer a professional well-run institutional standard investment management solution to anyone who wants it. It uses technology to bridge the “advice gap” for the customers described above, providing simple web-based customer interaction (on-line platform) supported by a telephone help facility as well as by the possibility for our clients to meet our advisors in person at our offices. MoneyFarm offers customers investment advisory and execution services applying a low and transparent fee structure.
The MoneyFarm ambition is to become, in the medium term, a leading European player in the digital retail financial products space: the success of the model in UK and Italy will be a template for us to expand further on a pan-European level in the future. Actually, we have about 3,000 customers and over €160 million of assets under management.
We agree that, in theory, automated tools offer access to advice to a wider range of consumers.
In particular, with reference to points 32 and 33, in principle we agree with ESAs’ statement that automated tools can give younger or less affluent consumers greater motivation to act upon financial matters that they would not have if they were using a human adviser.
However, in our experience, MoneyFarm’s clients are mainly conscious savers, searching for safer, less burdensome and more direct investment methods than the traditional channels. On average, (i) MoneyFarm’s users are 48 years old, (ii) the capital invested is around EUR 30,000, and (iii) the amount invested through MoneyFarm is normally only a portion of their whole investment portfolio.
As to point 34, we agree with the statement that, since automated financial advice tools are usually available online, it is “easier for consumers to access a wider range of advice providers, including from other jurisdictions”.
However, we have some reservation about the conclusion that automated financial advice tools “more readily facilitate cross-border transactions, compared to human advice”. Based on MoneyFarm’s model, users are allowed to invest only in ETFs listed on the Italian Stock Exchange, and no cross-border transaction is permitted. It could therefore be pointed out that robo-advising does not per se foster cross-border transactions, but it rather depends on the strategy adopted by each advisor when setting the algorithm.
MoneyFarm’s determination to narrow down the available assets is grounded on the its consumers’ friendly approach, i.e. to give advice to its clients on products that embed low costs, that are transparent, liquid and allow diversification (thanks to the broad variety of underlying available). An ETF, for example, permits to cover the global stock markets at a cost of only 0,25% and can be bought and sold every moment during normal trading hours on the stock exchange.
With reference to point 31, under “Benefits relating to costs” we entirely agree with the ESAs statement that “automation in financial advice could decrease the cost of providing advice”. In our case, the user pays only a fixed fee (1,25% for investments under 3000 euros; 0,7% for investments between 3000 and 200.000 euros; 0,5% for investments over 200.000 euros), which includes all the costs associated with investment, disinvestment, rebalance or any kind of operation on the portfolio, as well as the entry and exit fees.
We have observed the benefit of obtaining financial advice in a “faster, easier and non-time-consuming way” (point 35). Indeed, MoneyFarm questionnaire can be filled in on its website 24/7 and it takes just a few moments before the user receives the advice and is able to act upon it.
As to the “benefits relating to the quality of service” (point 36), MoneyFarm business model, according to which the fact that the algorithm is used only to allocate the investor to a specific risk profile to which a certain basket of assets is associated (see under 1 above), per se narrows down the margin of discretion of the human advisor, substantially reducing the risk of a human error. This implies a consistency in the accuracy of the advice provided to a number of users.
In addition, the possibility to go through the whole process without human intervention (unless required by the user) certainly excludes the potential pressure of a personal relationship with the advisor (point 37).
This clearly constitutes also a benefit for financial institutions: indeed, it reduces the exposure to litigation and reputational risk, as highlighted by ESAs at points 42 and 43 of the DP, under “Benefits to financial institutions relating to the quality of service” for financial institutions.
In our experience, we can confirm that automation allows the use of up-to-date market information (point 38) and the fast updating of the value of the portfolios. Indeed the maximum delay of the updates carried out be MoneyFarm is just 20 minutes after the data are available.
MoneyFarm keeps records of all the processes. In every moment, users are in the condition to print a summary document with all the questions and the answers they have given during the questionnaire (point 39).
The above audit trail represents a very important tool also for the advisor, as it helps, in the event of a consumer complaint, to retrieve evidence of the whole process and demonstrate that, when the advice was given, it was suitable for the client (point 45, under “Benefits to financial institutions related to the quality of service”, and point 80, under “Risks to financial institutions related to the functioning of the tool”).
We agree, in principle, with the observation (point 41, “Benefits relating to the size of the potential client base”) that financial institutions may benefit from automated tools by increasing their client base thanks to the wide reach of on-line channels. In our specific case, even if nothing prevents us to access customers from across the EU, MoneyFarm has decided to limit its reach to Italian residents only, in consideration of the fragmentation of the various national legislations, with particular concern for tax treatment of transactions and services involved.
Automated financial advice systems can definitely reduce the cost incurred by financial institutions to deliver financial advice (point 40). As a matter of fact, the automation allows to cut the cost of the “traditional” advice models, that is the cost associated to the creation and maintenance of the distribution network. On average, such cost absorbs 70% of the revenues of financial institutions.
We agree that the deployment of the algorithm makes the provision of advice by financial institutions more easily auditable (point 44, “Benefits relating to the quality of service”).
As a matter of fact, in our case, the circumstance that the logic of MoneyFarm’s algorithm was ex ante available for inspection made easier the vetting of MoneyFarm’s business model by the Italian Banking Authority (Bank of Italy) as well as the Italian Financial Authority (CONSOB) when the authorization to act as financial intermediary was to be granted.
In addition, given the status of authorized financial intermediary, MoneyFarm is subject to the supervision of national authorities, which includes on-site inspections, such as the one-month inspection carried out by Bank of Italy in October 2015. As a result thereof, our systems (including IT) were further improved according to the input from the Authority.
See above our feedback on points 42, 43 and 45.
With reference to the observation that the consumers are exposed to the risk related to their limited access to information and/or limited ability to seek for clarification (point 50), MoneyFarm’s platform is built in such way to mitigate such risk. Indeed, during working hours (8am to 8pm), if the user needs support of staff, he/she can: (i) contact a human advisor using various channels (phone, chat, email) and/or (ii) meet with an advisor in person at the MoneyFarm offices.
In any case, the client is entitled to suspend the questionnaire in any moment and to resume it at the same point after receiving the support needed.
As to the potential presentation of key information as “legal small print”, MoneyFarm makes any effort to avoid the unduly dismissal of key information, also taking into account that MoneyFarm is a regulated entity and as such complies with precise disclosure requirements (point 51).
As regards the risk that the consumer does not understand what kind of input is requested (point 53), MoneyFarm mitigates it through a three-level clarification scheme:
1) The user can display explanatory comments, containing the definitions of the technical terms used in the questionnaire, by simply placing the mouse pointer on the question mark relating to each notion;
2) The user can always access a glossary, which allows him/her to comprehend the economic concepts required for answering the questions, through common language and illustrative slides (“infographics”);
3) The user can always contact a human advisor (phone, chat, email) in order to seek clarification about the technical terms and questions open to interpretation (for instance, in case of subjective questions requiring his/her personal interpretation, such as those relating to the user’s risk profile).
We agree that there is a risk that the user is not fully aware of how his/her input is taken into account by the tool. For instance, referring to certain investment objectives may lead to focus only on some of them, disregarding others that could possibly play a more important role for the investment planning of the user (point 54). In order to mitigate this risk, MoneyFarm automated tool does not consider financial objectives at all, but only the relevance that the user imputes to his/her investment.
As to risk that the user is “not aware that the output produced by the tool is only relevant regarding his/her present situation” (point 56), it cannot be excluded that the client implements the advice after a long period from the time he/she received it, with the consequence that his/her actual circumstance has changed and the advice is no longer accurate. In our case, this risk is however mitigated by the fact that an input to execute is required, so that a reasonable person would update his/her own profile if a new element occurred during the time lapse from the initial input. In addition, in accordance with MiFID, we update the clients’ profile every two years.
While, in theory, there might be a risk that “free” advice “may, in fact, hide cross-subsidisation between the advice given and the final product/service chosen by the consumer” (point 57), in our case the “fee-free” advice, if any, is offered for promotional purposes only (for instance, Christmas or Valentine’s Day marketing promotions).
With reference to the risk of uncertainty about the compensation awarded to the advisor (amount and its potential link to the final product chosen) (point 58), MoneyFarm adopts a fee structure that presents such element that the above risk is excluded: (i) the fees paid by its clients (between 1,25% and 0,5%) are MoneyFarm’s sole source of revenue and this makes our platform strictly independent; (ii) the circumstance that the fee remains the same regardless of the basket of securities associated to the relevant risk profile excludes any interest for MoneyFarm to recommend a riskier investment than what is suitable for him/her; (iii) MoneyFarm does not receive any rebate from the issuer of the products.
In relation to the issue implied by the involvement of different financial institutions performing different parts of the advising process, it should be noted that MoneyFarm is responsible for the large majority of such process (point 63). In addition, in light of its status of supervised entity, the relationships with the other parties (e.g. the hosting provider, the pricing tool provider, the anti-money laundering control service provider or the middle/back offices) are regulated by specific agreements pursuant to the applicable law and regulations. Such formal organization of the relationship between the parties involved has a positive impact in terms of clear repartition of liabilities and makes it easier for consumers to direct queries or complaints to the appropriate entity, thus mitigating the risk highlighted in the DP.
As to the risk connected to the use of personal data of the consumer, MoneyFarm never uses any of its clients’ personal data for other purposes than giving them the advice they request (point 66).
In respect of the risk of error/inadequacy that may arise if the tool is developed “without the input of people who are appropriately qualified and have sufficient understanding of the implications of the financial advice that the automated tools produce as their output” (point 70 and 71), we tried to mitigate it by hiring for the development of the software professionals with experience in the field of financial advice, with only a limited support by IT specialists.
As to the risk of hacker attacks, MoneyFarm devotes significant attention to the safety of its IT systems. Suffice it to say that, on a total of 45 employees, 17 are members of the technology team, which is today the largest one. That said, it is undeniable that, despite we use the most advanced security systems available (the same used by banks all over the world), the risk of a hacker attack still exists (points 72 and 73). In order to tackle even this eventuality, a recovery plan is envisaged, as prescribed for supervised entities, to restore the functionality of the systems in the shortest time possible.
MoneyFarm has mitigated the risk related to consumers making unsuitable decisions because the tool facilitates them to move too quickly through the process (point 74) by making mandatory for the users to submit the questionnaire only when they have responded to all questions. In any event, it is beyond MoneyFarm’s control to verify that the user answers the questions with due attention and taking the necessary time.
As to point 75, initially (when MoneyFarm was an advisory-only platform), we experienced that frequently clients had a lack of motivation to act upon the advice received, since they were not put in the condition to do so in an easy way and in short time. Once MoneyFarm integrated its business model including the execution phase, the above mentioned lack of motivation disappeared: currently, on about 4,000 clients, only 600 still use the advisory-only system. According to the advisory+execution process, the client, after filling in the questionnaire, is able to invest immediately by clicking the relevant option.
With regard to the potential impact of a “herding effect” (point 77), since the object of MoneyFarm advice is limited to one type of securities (ETFs), it is very unlikely a systemic impact of the investment decision taken by MoneyFarm’s clients, even in case of widespread use of its tool. Furthermore, this risk is also mitigated by the presence of human advisors, whose sensitivity is necessary to avoid a mass investment on a single product that would cause detriment to the clients.
After 3 years in business, we have not received even a single complaint by any of our clients.
In connection with the risk that “consumers may overuse that alternative means so as to supplement the automated advice on the product/service” (point 82), MoneyFarm advisors are not allowed to give advice outside the scope of the MoneyFarm mission.
As to point 84 and 85, it should be noted that, being it a supervised entity, MoneyFarm has in place specific legal agreements with the different parties to govern the liability on an on-going basis, so, in principle, the allocation of liability is clear, to the extent that the contracts are properly drafted.
We appreciate ESAs’ awareness of new developments in the financial environment and we support the effort to promote convergence in the regulatory practice in order to promote the safety and soundness of the market.
To this end, we believe that a preliminary consideration concern the clarification of the notion of advice that is relevant for the construction of the regulatory framework. This is key to favour the orderly development of the phenomenon of automation in the financial sector.
On the one hand, we appreciate the effort not to limit the scope of the future regulation “to a narrow definition of advice under a particular EU legislation”. Indeed, if a single (rigid) definition of advice is to be adopted in the legislation, the risk would be to level factual situations too different from each other, which would imply a potential overregulation of those tools that expose users to significantly lower risks. This could discourage automated advice providers to keep or introduce tools that offer only simplified advice, with likely lost opportunities for consumers and businesses.
On the other hand, while it is functional to the success of the discussion paper in order to achieve a comprehensive view of the phenomenon, the broad reference to the consumer perception (of what advice is) is likely to give rise to a concern in terms of uncertainty. By the definition, the outcome of the human perception is affected by many factors (personal, financial, environmental conditions, level of education, experience, age and so on) and is intrinsically volatile. As such, the scope of the forthcoming regulation would risk to be indeterminate or at least a floating concept.
In light of the above, in order to achieve a possible solution that takes into account both the key role of a legal definition and the importance to cover the various robo-advising tools, it could be advisable to envisage an articulate definition of advice, attached to a gradation of liability, based on the type of advice provided. In this sense, we find a useful example in the road taken by the Financial Conduct Authority (FCA") with the Finalised Guidance FG15/1 ("Retail investment advice: Clarifying the boundaries and exploring the barriers to market development"), where it provides, inter alia, a distinction between "focused", "simplified" and "full" advice, differentiating the applicable rules.
---
MoneyFarm is the leading regulated entity providing automated financial advice in the securities sector in Italy and UK via its sister company MoneyFarm Investments Ltd.
Legália is a boutique law firm active, inter alia, in the financial sector and specialized in the “FinTech” field.
---
MoneyFarm - Paolo Galvani, Chairman
Legália - Vito Vittore, Partner | Luigi Bonifacio, Associate"
Vito Vittore
M