David Severn

The characteristics described are broadly right. It is likely that consumer engagement with financial services by digital means varies from one Member State to another. In the UK, for example, the Office of National Statistics reports major growth in use of the internet for banking transactions. The nature of these transactions may not currently constitute “financial advice” but they form a base on which banks might develop their offerings to consumers in the future. Comparison sites used by consumers for mortgages, non-life insurance and credit cards are also well established in the UK. It is not clear how much automated advice is used for more sophisticated and high value transactions in investments and pensions, although there are a number of sites offering consumers the opportunity to hold, and manage, their securities on-line on a non-advised basis and these sites usually have a variety of tools available.
Yes. At present the majority of offerings deal with one aspect of a consumer’s finances in isolation from their total financial and personal needs. This is a point to which I return under risks for consumers
I can offer comments here mainly from a UK perspective. For around 17 years until end 2004 I was a senior UK financial services regulator. During my time with the Personal Investment Authority (PIA) in the 1990s I went out of my way to engage with firms trying to develop digital propositions to see if the rules of the regulator unnecessarily restricted innovation. In a PIA document of 1996 I wrote “ At present the bulk of the business regulated by PIA is conducted wholly or partly by face-to-face interviews, by telephone, or by written communications ... But PIA is increasingly being asked to discuss, and is happy to do so, various proposals for the use of new technology”. In a follow up PIA document of 1997 I wrote “ A particular aspect of IT is the use of the internet. A number of PIA firms already have a presence on the internet but the use which they are making of the medium is embryonic and the volume of business is negligible. But that position could of course change rapidly and in directions which neither the industry or PIA can anticipate”. When I moved to the Financial Services Authority (FSA) I tried to ensure that the Conduct of Business and Collective Investment Scheme rules, for which I had responsibility, were neutral and did not stand in the way of firms using digital means for interaction with customers. On a specific point, while at the FSA my team developed a so called “Basic Advice” process and to test this I commissioned a UK firm, Distribution Technology, to develop software which allowed financial advice to be delivered to those with fairly straightforward needs. I think this demonstrates that in the UK, at least, there has long been a willingness on the part of regulators to consider adapting rules where they might otherwise stand in the way of digital innovation so long as adequate consumer protection remains in place. I also noticed when I was a regulator that there were a few firms –whether through ignorance or mischief – who alleged that things could not be done by digital means when in fact there were no obstacles in the rules of the regulator.
I agree that there are benefits to consumers from advice that is consistently accurate, but this is based on the assumption that the design of the software is such as to produce a good outcome for consumers, another point to which I return under risks. There are also potential gains in consumers paying less but again this is dependent on firms actually passing on to consumers savings which the firm makes rather than retaining savings as extra profits. There may be gains in improved access to advice but this assumes that consumers using automated advice have the application and understanding to make sensible decisions based on information presented to them. It also has to be remembered that even in the UK where there is significant use of the internet there nonetheless remain significant areas of digital exclusion. Finally. It also needs to be remembered that the use of digital means does not solve the problems which some consumers have in engaging with financial services because of poor literacy or numeracy.
The use of digital means may encourage firms, because of lower costs, to maintain contact with customers and offer reviews of their circumstances. This could enhance confidence in the financial services sector and could bring repeat business to firms.As a regulator one of the problems I have observed is that much financial advice remains transaction based rather than the consumer having an on-going relationship with an intermediary or product provider. The risk for consumers is that over time a financial product may no longer meet their needs, asset allocation, or risk attitude but they may be unaware of this unless they receive advice.
I agree with all of the risks identified. In a recent article in the UK trade paper “Financial Adviser” I argued that one of the big issues with automated advice was that it tended, at present, to be based around an isolated financial need rather than looking at the overall financial and personal circumstances of the consumer. The risk is that consumers will buy products that do not fit well with their overall financial needs. Moreover, many offerings are transaction based and do not necessarily provide any later support and advice. As a consumer’s circumstances change, as economic circumstances change, and as any product they have bought may change, it can mean that over time a product that was quite suitable when initially purchased may at a later date become quite unsuitable.
In the same article I pointed to another critical risk that needs to be borne in mind the whole time and that is that any automated advice system is only as good as the people who design it. If a system is designed so that it always results in a recommendation to buy a particular product, no matter what information has been entered by a consumer, then the system becomes what I called an “automatic mis-selling machine”. I have in the past urged the Financial Conduct Authority and its predecessor to consider some form of accreditation system for advice and risk profiling software. My reasoning has been that with a flawed digital system each and every customer of the firm is automatically put at risk, whereas with a firm using human advisers only those consumers who are advised by a less competent adviser are going to be put at risk.
The Discussion Paper does not deal with issues of reliability and security but I suggest that these are likely to be prominent in most consumers’ minds and will be a major factor in the willingness of consumers to use automated approaches and to have confidence in them. I have not suffered any major problem so far but a series of irritating problems caused by internet service provider failings, two different banks who for different reasons failed to execute on-line transactions as basic as debit card payments, and passwords for three new sites which did not work and required the services of the providers’ help desks to resolve. Niggling problems like this hardly encourage confidence in the IT capabilities of firms. Along with other consumers I have read of the more serious data breaches such as that faced by JPMorgan. There was a recent report by Dr Christopher Richardson of Bournemouth University that “ Less than 25 per cent of cyber security applicants are qualified to perform the skills needed for the job”. If issues of reliability and security are not tackled it will undermine consumer confidence in using automated advice systems.
The Discussion Paper does not deal with “legacy systems” which proved to be a significant constraint on both the Government and regulator in the UK. As mergers and acquisitions of firms took place over time it resulted in the position where some financial institutions acquired a large assortment of ill matching IT systems and it was necessary to give firms a significant lead time before they were able to implement any changes in tax or regulatory requirements. It seems probable that with the proliferation of IT the legacy system issue is going to cause problems in the future for consumers, firms and regulators alike.
The main issue here is the scale of the potential detriment to consumers from on-line transactions. Renewing annual car insurance is relatively low value and if a consumer gets a policy that does not fit their circumstances or is far too expensive there is an opportunity to correct the error the next year. Pensions and investments, however, are longer term and higher value and the potential to do lasting damage to consumers’ long term financial health is much greater.
Yes. There is in the UK a tendency for some firms to try and dress up automated systems as “guidance” or “information” in an attempt to shift risks on to the consumer and deny any liability for mis-selling of products. All the research I saw as a regulator suggested that anything which guided consumers to a particular outcome was perceived as “advice”, no matter what description a firm applied to the process, and it was only when firms stuck to unvarnished facts about a product or service that consumers were prepared to regard it as “Information”. There is a risk here that badly designed automated systems could result in claims from consumers for poor advice and the firm concerned then goes out of business, leaving good firms to pick up the burden of funding compensation for consumers.
I suggest that at present automated advice is very much supply led. Firms want to cut costs, speed up transactions, access new consumers and try to generate repeat business from existing customers. So currently consumers are getting whatever it is that firms are deciding to offer them. What I think is missing is any reliable and comprehensive evidence on the “demand” side. What financial business do consumers want to transact on the internet and what business would they prefer to transact by other means, and why is this? What concerns about reliability, security etc do consumers have? How do consumers actually use the internet at present when carrying out financial transactions? Would consumers actually have the patience and application to enter the information (income and outgoings, assets and liabilities, risk attitude etc) necessary to get holistic financial advice?
Although the systems of which I’m aware are quite sophisticated I do not think any of them warrant the description of artificial intelligence or expert systems. I am sure, however, that something along those lines cannot be far off and the EU needs to be ready for such systems before they are launched.
The Discussion Paper could usefully have covered not-for-profit sites intended to help consumers with financial matters. Two significant ones in the UK are the Money Advice Service and the The Pensions Advisory Service.
David Severn