Revolving door and entrance/exit of building.

Getting ahead of the curve: EBA consults on 'robo-advisers'

17 December 2015

Technology will transform the financial advice world, and this transformation has already begun.

The global assets under the management of robo-advisers is forecast to grow to an estimated $450 billion by 2020. Regulators across Europe need to get ahead of the curve because the move to so-called 'robo-advisers' is going to affect millions of retail customers across the continent. Recognising this, the regulators of Europe's single market for banks, financial markets and insurers have recently published a discussion paper on automation in financial advice.  No solutions are offered, but the paper does examine the benefits and risks to both firms and consumers and requests input from those who hope to become involved in this fast-growing industry.

Rise of the robots: benefits and risks

Joint Committee Discussion Paper JC 2015 080 has been published by the European Banking Authority (EBA), the European Securities and Markets Authority (ESMA) and the European Insurance and Occupational Pensions Authority (EIOPA) which are collectively known as the European Supervisory Authorities (ESAs).

The Discussion Paper provides a useful working summary of automated advice: 

The technology behind the automation of financial advice is typically an algorithm. The eventual advice provided is therefore reliant on two key inputs: (i) personal information input by the consumer (generally online, and in a questionnaire format); and (ii) the logic of the algorithm, which ‘decides’ which products or services should then be recommended to the consumer. Automated financial advice tools are often presented in the form of a decision tree, where the consumer responds to a sequence of scripted questions which will generate recommendations based on the consumer's specific responses.

The ESAs note that the use of automation is different in the sectors they regulate: 

• In banking automation of advice is not widespread, rather human advisers are supported by websites providing simulators and calculators which help consumers choose between different products.

• In the securities sector the use of automated financial advice is more mature, with 'robo-advisers' asking investors about their specific circumstances and, based on the answers provided, an algorithm is used to recommend transactions in financial instruments that match the customer’s profile.

• In the insurance and pensions sector new business models have emerged, providing online independent financial advisory services which use algorithms to select pension investments for savers, while insurers' websites provide personalised quotations on the basis of a questionnaire.

The ESAs identified a number of key benefits of automated advice, basically: more consumers have better access to a wider range of service providers and to more consistent up-to-date advice, more easily, more quickly and for less money.

It is likely that providers have already identified these advantages for customers, but it's important that they build these consumer-centric USPs into their business plans if they wish to receive a favourable reception from national regulators.  The ESAs recognition of these benefits will likely help any firm seeking regulatory approval from a member state National Competent Authority for novel automated advice solutions.

The ESAs also emphasised certain advantages for businesses in using automated tools, including that firms operating an automated advisory model might be better able to comply with their regulatory obligations.  For example, automated processes that are documented through an algorithm or decision trees can be more easily audited. The automated tools will, unlike their human compatriots, also provide consistent advice.  Although, if there is a problem with an algorithm and it is not monitored closely, a firm operating an online tool will face immediate systemic problems as any error will affect a large number of consumers and will give rise to a high number of complaints or lead to it undertaking a costly redress exercise. 

Flaws in the functioning of automated tools is a risk identified in the Discussion Paper, along with risks related to consumers having limited access to information and a limited ability to process that information with little opportunity to fill in gaps in their knowledge or to seek clarifications, which would become particularly problematic if 'robo-advisers' become so widespread that consumers have no access to human advisers.  The ESAs also note that there are risks if consumers enter the wrong input data because they have misinterpreted the questionnaire, for example, entering gross yearly income rather than net yearly income, or because the tool is entirely reliant on consumers' response to subjective questions, for example, about their attitude to risk for which they do not have an adequate benchmark.  Consumers might also fail to take action to implement the advice with an appropriate time frame such that the advice becomes unsuitable without them realising.  The Discussion Paper also recognises existing risks that have been identified in the advisor market around independence (which have been addressed in the UK through RDR and will be addressed in Europe through MiFID II).  The Discussion Paper identifies a current problem in the market where complex distribution structures and, for example, the white-labelling of automated tools, mean that consumers are not aware of who provided the advice.  Finally, the Discussion Paper also recognises 'herding-risk' which might materialise if 'robo'advisers' become commonplace and if they rely on algorithms which make a similar set of assumption.  From a prudential standpoint, this could be pro-cyclical and increase volatility in asset classes in which consumers commonly invest.

Changes to regulation?

The Discussion Paper highlights key problems with the European regulation of advice as a whole, not simply with the regulation of automated advice.  While there is no definition of 'automation' in any relevant EU legislation, there is also a different concept of advice in different EU directives, for example, the definition of 'advisory services' in the Mortgage Credit Directive (MCD), which is limited to MCD products rather than banking products generally, is different to the definition of 'investment advice' in the Markets in Financial Instruments Directive (MiFID) and the Insurance Distribution Directive (IDD).

The Discussion Paper also recognises that distributors using online channels must also comply with a plethora of other EU legislation that would not apply to offline distribution, such as the Distance Marketing Directive, the E-commerce Directive and the Online Dispute Resolution Regulation. This means that, despite the cost advantages of online advice, it may lead to higher regulatory set-up and ongoing costs.

At this stage the Discussion Paper does not offer any solutions but it is to be hoped that they may be forthcoming.  The Discussion Paper asks for responses to a list of 24 questions from respondents.  Those concerned that the current un-harmonised regulatory landscape is inhibiting the development of pan-European advice tools, taking advantage of the Single Market, should respond to Question 5 which asks whether there any barriers preventing respondents form offering or developing automated financial advice tools.


At this stage, the initiative of the ESAs is to be welcomed.  The advisory market has been moving towards the 'robo-advisory model' with tentative steps.  However, once market participants have regulatory certainty, I expect the 'robo-advice' market to see a huge amount of investment and to gain substantial traction with consumers.  If consumers are willing to sell their house through online-only estate agents, there is little doubt that they will value the cost-savings and convenience of online, automated financial advice.  Regulators need to ensure that this is safe for the individual consumers whose pensions and savings are at stake, for advisers who could face enormous complaint liabilities and for society which may otherwise have to pick up the tab.