Finance, algorithms and vulnerable groups discussed at Finance Watch conference

On the 19th of November, Finance Watch held their annual conference entitled “Ten Years after Pittsburgh: Changing Finance for Good?” in Brussels. The conference addressed three main themes, which are at the heart of Finance Watch’s mandate: Stability, Sustainability and Inclusion.

During the event, Finance Watch presented the preliminary results from a major study meant to shed light on the state of play of financial inclusion across Europe, and understand what are the most pressing needs of financial service users (or non-users, in case of excluded people). Among the most vulnerable groups, the study identified over-indebted people, homeless people, unemployed people and the working poor as the most commonly reported users excluded from access to basic financial services.

COFACE Families Europe, associate member of Finance Watch, was represented by Martin Schmalzried, who spoke about financial inclusion during the first panel session. His intervention covered two main topics: fair pricing and exclusion due to the costs of financial products, and AML/KYC (anti-money laundering and know your customer) as potential obstacles to accessing basic financial services such as a basic bank account.

At present, pricing of financial products is very much tied to the idea of personal liability and identifying more and more individual risk at the expense of socialization of risk. Be it in insurance products or credit risk assessments (as distinct from creditworthiness assessments), we are witnessing an increase in personalization or customization of pricing and product features which are supposed to be “tailored to the person” to better “fit their needs”. But the end results are often counter-productive and can have very negative effects leading all the way to financial exclusion:

  • First and foremost, it is important to dismiss the idea that liability for ones actions rests solely on the individual. As Artificial Intelligence (or rather, complex algorithms looking for correlations between millions of data points) is increasingly being used by financial service providers, it is key to orient its power of analysis to identify structural or environmental factors which lead people to make certain decisions. For instance, there is a clear discrepancy between research looking at causes of over-indebtedness and the data that algorithms are being fed in order to find correlations and trends. Over-indebtedness is mostly linked to unpredictable life accidents such as divorce, health issues, loss of employment or death. This should translate into a higher socialization of risk across the customer base rather than making certain customers pay more based on certain individual characteristics. On a more profound level, one can find macroeconomic factors which explain why certain people are unable to repay their loans. The research of professor Steve Keen shows, for instance, an extremely strong inverted correlation between the level of unemployment and the rate at which new credit is created/issued. What this means, is that the likelihood of individual borrowers repaying their debt is directly linked to whether other people are willing to borrow at the time they have to repay their credit!
  • In insurance, the same kind of reasoning can apply. Instead of using algorithms to identify personal behavioural patterns in order to “punish” people (make them pay higher rates based on certain behaviours categorized as “more risky”), use the algorithms to identify the root causes or environmental factors/external influences on people’s individual behaviours. A clear example could be monitoring the exposure of each individual to advertisements and check whether there is a link with their consumption patterns/personal behaviours identified as “risky”. Instead of increasing the insurance premium those people, the companies responsible for influencing their “at risk” behaviour could be levied/taxed to subsidize the higher insurance premium.
  • Customized or tailor made financial products have several other negative effects, such as preventing the consumer from comparing different products to choose the most suited for their need. It also creates prudential risks as the resilience to shocks is much less clear in a situation where risk is calculated on an individual basis as opposed to covering an entire population on the basis of broad trends.

For more information about the conference, visit the official website here:


Or contact Martin Schmalzried : mschmalzried@coface-eu.org

Translate »