“Wearing two hats made it easier for me to connect within government”
PhD defence Alexander Melchior AI Lab for Public Services
Agent-Based Modelling (ABM) is a powerful tool for developing sound policy on complex issues. Yet civil servants rarely make use of it in practice. For his PhD research at the AI Lab for Public Services, Alexander Melchior – a civil servant himself – explored why that is and what needs to change to make ABM more useful for policymakers. “Within the world of civil service, it’s essential to build mutual trust.”
Agent-Based Modelling (ABM) is a powerful method that can help policymakers understand complex issues such as climate change, immigration, and sustainable agriculture. It is a modelling technique that allows one to create an overview and simulate how individual systems interact and what effects these interactions have on human behaviour.
“For example, if you look at Dutch climate policy, there are multiple interrelated issues at play: reducing emissions, accelerating adaptation, and creating social support,” explains Alexander Melchior, who defended his PhD thesis on 23 June. “When you integrate those domains into policy, they begin to influence one another. That’s difficult and complex to manage. ABM allows you to clearly observe the interplay of different policies, making it easier to develop effective ones.”
Gap
What’s striking is that this valuable tool is hardly used by policymakers. “Apparently, there’s a gap between how highly academia values this technique and the recognition it receives from policy developers.” For his research, Melchior drew on his own experience as a policy advisor, conducted interviews with experts, and studied case studies to uncover why policymakers tend to overlook ABM. He also developed a practical framework with a set of requirements to help integrate ABM into policy development.
Melchior conducted his research within the Utrecht AI Labs. In these Labs, researchers are employed by both the university and an external organisation. His position as a civil servant made his investigation considerably easier. “I work as a senior policy advisor at the Ministries of Economic Affairs, Climate and Green Growth, and Agriculture, Fisheries, Food Security and Nature. At the same time, I was working at the university. Wearing both hats made it much easier to establish connections within government. Compared to an academic from outside the system, you already have a head start. That gave me a unique insight into the organisation.”
Informal contact encourages policy staff to share more information
Building relationships is therefore one of the most important conditions for incorporating ABM into policy processes. “Modelling for policy is more than just policy modelling. In the civil service, building trust and maintaining relationships is crucial. Regular informal contact encourages policy staff to share more information and be more receptive to each other’s input. Building a model together also strengthens that relationship.”
It’s also vital to give civil servants a clear course of action. “Workloads in government are high. As a researcher, you can’t just produce a fascinating report and expect a civil servant to figure out on their own how to apply the insights in their policy work. Without a concrete action plan, a report ends up gathering dust on a shelf.”
Transition to electric vehicles
As an example, Melchior points to a 2020 case study in which he explored how ABM could assist in policy for the transition from petrol to electric vehicles. “At the time, progress in the transition was mainly measured using financial indicators. But people were looking for other factors that could also shed light on the shift. So we engaged with many stakeholders in the field to explore what influences consumer behaviour and developed a model based on that.”
Melchior’s research revealed, among other things, that car dealers have a major influence on consumers. “Dealers earn a lot from vehicle maintenance. That’s why they often speak negatively about electric cars if their mechanics aren’t trained to service them. And that can happen if training programmes don’t include it, which might be because maintenance of electric vehicles isn’t part of the official exam requirements. My research led to identifying exam requirements as an indicator for the transition. That offers policymakers a course of action by showing what data should be collected to better monitor the transition, and which levers can be pulled to accelerate it. In this way, the model helps civil servants gain more control over the factors influencing the shift to electric transport.”
I enjoy bridging the gap between science and policy
Melchior looks back on his PhD with satisfaction. “I’m originally a computer scientist, but this topic was highly interdisciplinary. I really stepped outside my own bubble by also looking at public administration, psychology, sociology, and so on. It was challenging, but also fascinating. I enjoy bridging the gap between science and policy, and I’d love to continue doing that now that I’ve completed my PhD. Both sides can benefit from working together, but finding each other remains a challenge.”