Scroll to top

Behavioural Agent Based Modelling Of Bank Account Switching 

Decision Lab Case Study with Pay.UK

Increasing competition in the banking sector 

The UK government had a political objective to increase competition in the banking sector. To facilitate and de-risk consumers changing their bank accounts, the government launched the bank account switch guarantee service. However, the level of current account switching remained very low and Pay.UK wanted to understand this and to assess how future incentives might motivate switching and increase the switch rate. 

Pay.UK specifically wanted to test the impact of its idea of a “portable bank account number” – an account number and sort code that stayed with you regardless of who you banked with. 

Capturing the complexity of human decision making  

Our biggest challenge was to create an accurate model involving human behaviour in real-life events. There are multiple layers of complexity – for example, how do we model the life events that are typical triggers for people to consider switching their current account. And then there is a generational factor in the decision-making, as older people are less willing to change accounts than younger ones. All this is further complicated by how each person will react to a variety of news events, bank actions (positive and negative) and the marketing a person will receive both from the bank they are currently with, and competitor banks. 

Turning qualitative data into a set of hard rules 

We developed an agent-based simulation model of people’s decision-making when considering switching accounts, including customer awareness, the pull and push factors, etc.  

To understand the hows and whys of what makes people tick – their decision-making regarding their financial interests – we had to exploit the latest academic research in psychology. The qualitative research into this was undertaken by an academic partner, but we had to convert their findings of human behaviour into hard numbers and rules that we could build into the model, and get results that matched the data. 

By running multiple scenarios, we could assess different intervention strategies – incentives offered, marketing campaigns, implementation mechanisms and more – and determine their effectiveness in getting people to switch accounts and increase banking competition. 

The bad news and the good news 

The simulation that we built showed that the main idea is Pay.UK wanted to test – a “portable bank account number” – doesn’t work. This was important to know and Pay.UK did not introduce this to the UK banking sector, and so avoided a potentially costly failure. 

But our model offered much more: it provided a capability to assess and understand the potential impact of different banking innovations. This included the new “challenger” banks – the likes of Metro and Monzo – which were emerging and predicted the effect this might have on the existing high street banks. Pay.UK use the tool to assess the impact of future banking innovations and guide their strategy.