Big data and fintech: a solution for financial exclusion ?

Taking the various headings of the conference sessions one by one:

1. As to insurance

Some types of discrimination between groups in society are presently forbidden to providers of insurance products, for example discrimination based on gender or race. Others are allowed, for example discrimination in the calculation of car insurance premiums based on age – young/old.

For health insurance, discrimination based on risky behaviour – smoking – is allowed, but equally, there is no regulatory barrier to discrimination based on genetic inheritance, or chronic health conditions.

FinTech and Big Data raises the possibility that new algorithms for calculating insurance risk will be introduced. In particular, these algorithms could be based at least partially on measured behaviour, either indirectly through data captured from other sources such as social media, or directly using data captured from a direct, real-time monitoring of behaviour for example from a black box installed in a vehicle.

Of course, for vehicle insurance, there is already a bonus/malus system in widespread use, whereby a track record of claims determines ex post conditions of access to insurance. With real time monitoring of behaviour, conditions of access to insurance could, theoretically, be determined by the reality of behaviour on a continuous rather than ex post basis – something that could be favourable, for example, for a young driver who drives prudently and well. Pay as you go premiums in other words.

Similar developments might be imagined for health insurance – based for example on the monitoring of life-styles, risky or prudent.

Would consumers be ready to accept direct or indirect monitoring of their behaviour in exchange for better conditions from insurance providers? Would they be taking such decisions in full knowledge and understanding of the implications? How could they make price comparisons with offers made to other consumers?

Would insurance providers be able to manage the complexity of such a detailed segmentation of insurance markets? Would the erosion of the pooling of risks this implies undermine their business model since even the most sophisticated of profiling and monitoring tools can and will give false results in individual cases?

Would the development of such algorithms risk creating new kinds of unacceptable financial exclusion since algorithms may reflect the sub-conscious prejudices of their creators or sponsors and not only objective realities. Should regulators be tasked with checking and approving such algorithms – with what staff, with what expertise? And against what standards – what types of discrimination are acceptable, and which are not?

So many questions, so few answers. But the good news perhaps, is that rapid change in the market for insurance products is thought unlikely. The impact of FinTech is likely to be mostly on the trading of risks within the financial sector, rather than on the interface between the insurance provider and the consumer.

Moreover, cross border marketing, and hence international competition for the provision of insurance products, is held back by the very different cost structures and legal systems in different countries.

One clear positive might be the introduction of artificial intelligence (AI) systems for giving advice to consumers on offers available in the market and their suitability for each consumer’s specific requirements. Potentially, if well designed and genuinely independent of commercial interests, these could be more efficient and trust-worthy than human beings. But again, should there be some regulatory system for verifying this is genuinely the case, and who would do such a job?

Last but not least, what does all this imply for the future role of government as the insurance provider of last resort, or even first resort – for health risks, and for some life events such as divorce, unemployment? Private insurance providers can only be expected to offer coverage for certain categories and types of risk. After all, some types of risk coverage almost inevitably increase the hazard of the risk occurring, and that is an issue that can only be settled by moral and social choices in society.

2. As to payment services, transactions and transfers.

Access to payment services, and transaction and transfer services, is a fundamental need for all citizens in a “low cash” environment. While a completely cash free society seems unlikely and undesirable, access to electronic services is already a basic requirement to lead a normal life, or even access to social security benefits. From this perspective, the right to a basic banking account and related services as it will be established under PSD 2 is an important reform, whose practical application in every Member State of the European Union needs to be closely monitored.

FinTech offers a perspective of further progress in two respects. First the rapid spread of low cost electronic payment and related banking services is of clear potential benefit to consumers. Secondly, FinTech is opening up the possibility of greater competition for transfers and remittances to third countries, of particular benefit to migrants for example, who currently pay high fees in many cases. The regulator could do more to facilitate these developments, in particular by acting to prevent oligopolistic practices obstructing new market entrants, or which prevent cross border access to new providers in the jurisdiction of another country.

The impact of FinTech is both immediate and profound on the traditional providers of banking services. The accelerating closure of bank agencies, and the likely decline in the number of ATMs, impacts especially on rural areas and deprived communities. This creates new forms of exclusion for the populations concerned, and in general, for those who are not digitally literate, or have special needs (poor eye-sight for example), or who prefer not to use electronic services for whatever reason. Is this a matter on which the regulator should intervene, so as to preserve a universal service that is accessible to all groups in society? Traditionally, post office banking has had a social role as well as a commercial role, but privatization tends to privilege the latter.

3. As to credit.

Of course, credit should not be treated as a permanent top-up to an insufficient income so as to enable a normal life. It should be a temporary measure to smooth out over time affordable purchases, or to meet temporary unforeseen circumstances. In an ideal world, individuals and households would be in a position to accumulate savings equivalent to at least three months income, so as to be able to minimize the need for recourse to credit except for major purchases such as housing and possibly, a new car.

FinTech combined with big data offers the possibility of better algorithms for assessing the credit-worthiness of solvent households, based once again on the reality of behaviour rather than on categories of clients. For example, recent records of financial behaviour for example bank statements during the previous three months, are said to be rather reliable measures of credit-worthiness. This could be helpful for those who do not have an established credit history for whatever reason, or who find themselves in a generalized “at risk” category despite their individual behaviour.

Also, making available “apps” to consumers that give them an opportunity to assess for themselves their capacity to borrow, and to better understand what determines their creditworthiness and maybe adapt their behaviour accordingly, could be helpful to all concerned.

However, FinTech and Big Data also bring with them the risk that unscrupulous lenders could target more effectively vulnerable consumers, and so expand still more the market for high cost, or worse, toxic forms of lending.

Given that regulators have been, in many countries, unwilling to act vigorously to suppress usury, this risk must be considered serious. Universally, the poor and vulnerable pay more for credit than those whose needs are less. Should government in partnership with alternative providers do more, if the business case for providing fair and affordable loans to this market is not demonstrated?

Another concern is the widening scope of data on life-styles collected by credit rating agencies. Are these agencies collecting only data demonstrably relevant to assessing credit-worthiness, or are they increasingly engaged in comprehensive data mining for other purposes, such as marketing ideas and products in general? There are many calls for more transparency about their activities, for a right to challenge their assessments of credit rating without risking penalties, for a right to live off-line without being penalized when seeking credit. And of course, for better guarantees that data is kept safe. This is not so much an issue of FinTech as such, as a question of who collects and who owns big data, and for what purposes, and with what consent from those concerned.

Across Member States, the types of data collected and the criteria used to establish credit-ratings are quite different. There is no established orthodoxy as to the type of data it is relevant to collect. It would be attractive if, at a European level, a process could be started to build a consensus around what is genuinely needed, and what is not, and on that basis, to regulate the types of data that credit rating agencies are licensed to collect. A fundamental principle should always be, only collect what is strictly needed for the declared purpose.

The issue becomes even more important as the “internet of things” develops, and a new source of data about personal life-styles becomes available. What do you have in your fridge – does this suggest a healthy life-style?

Finally, in the modern world, there is a proliferation of new credit providers, including in particular, retailers. Access to credit is used as a means of capturing consumers. The business model is not to provide a service to a consumer, but rather as a tool for companies to segment markets and control consumer loyalties. Insufficiently regulated, competitive pressures amongst retailers lead to unsustainable credit offers and a new generation of over-indebted households.

In conclusion:

Not surprisingly, there are more questions than answers. Perhaps it is more important at this stage to be asking the right questions and collecting the right data to assess the reality of problems. It is arguably premature to be looking already for the right solutions, except in those cases where the issues are the most evident and urgent.

Another way to analyze the question is to look at the impact of FinTech and Big Data on specific groups:

• people who are not earning enough income to live a normal life in the society to which they belong. The solution has to be access to a sufficient income, not dependency on credit as a substitute;
• people just about managing who have no savings (such as young families). The best solution is to help households achieve a modest saving potential to reduce their need to call on credit, as well as ensuring access to affordable credit when needed. Is FinTech commercially motivated to play a role?
• people in unstable employment. Can Fintech and Big Data better take into account their prospects over time and provide temporary solutions in a more attractive form, for example within mortgage or pension schemes?
• people with special needs. Cross border marketing might help by creating larger groups of clients and hence a stronger business case for providing tailor-made solutions including by the design of universal services;
• workers mobile across borders and migrants. FinTech can help by increasing competition and reducing the costs of remittances, and by facilitating access to financial services including credit where there is no established credit history in the country concerned;
• finally people without a data profile. What will be the future for them?

At the time of writing of this newsletter, contributions have not been published on-line.