Can you put wildness into a formula?

Can you measure the ferocity of problems? And if so, how do you go about it? INVI is trying to develop a wild problem model to support the tools we develop to deal with wild problems. The work is led by Jacob Gerner Hariri, Professor at the Department of Political Science at the University of Copenhagen. We asked Hariri about the role of the panel of policy entrepreneurs in the project and what such a model can be used for in the real world.

- The model tries to put wildness into a formula? Can you really do that?

It's hard, no doubt about it. But as social scientists, we are constantly putting difficult quantities into formulas. Other examples include democracy and bureaucratic efficiency. These are big and controversial concepts, but we still try to formalize them. The trick is to do it in a structured way.

- What is the purpose of the wild problem model? What should it do?

The overall idea is to measure the way problems are wild because it matters for how we address the problems. When a problem is wild, it is difficult to deal with. The job of the model is to tell us what makes that particular problem difficult to deal with. It must therefore be able to distinguish between different ways of being wild.

One example is the context of the problem, which can be extensive and involve actors across different boundaries, including both levels and sectors. The problem may not only be municipal and cannot be limited to just one area of expertise.

But the model should also be able to say that one problem is wilder than another, just like we say that Norway is more democratic than North Korea. And if we stick to that example, the model should be able to point out what makes Norway more democratic than North Korea. In other words, the problem is wilder on certain dimensions, but perhaps not on others.

At the same time, the model should be able to track problems over time and assess whether problems become more or less wild on the different dimensions.

 - Can you tell us a bit about your approach to developing the model?

I take policy entrepreneurs very seriously and have taken the approach that I want a practical perspective on wild problems that we can combine with something objective. It's really important that the model has input from those who have their hands in the dirt. Those who create, implement or evaluate policy and who know something about the problem at hand. So I try to stand on the shoulders of the panel of policy entrepreneurs and their vast knowledge in an objective way.

- What role does the panel of policy entrepreneurs play in relation to the the model?

The panel of policy entrepreneurs plays a crucial role in the model. They ensure the model's connection to reality. For some problems, the panel becomes the primary source for the model to assess the cause and solutions of the problem. Here, I will ask the panelists to formulate their assessment of a problem and then I will objectively quantify their answers. This requires the help of some long-winded quantitative language analysis - the same method that makes Chat GPT work. Specifically, the language model spits out a number based on how much the panel agrees or disagrees with the given problem. In this way, we quantify the panel's assessments as objectively as possible.

Jacob Gerner Hariri is leading the work to develop INVI's wild problem model.

- What do you measure? What are the units to measure complexity?

The model measures four dimensions. The first is whether people agree on the cause of the problem. For example, what causes young people to be unhappy? Is it screen use, character pressure, alcohol culture, loneliness or the parents? This constitutes one dimension of wildness.

The second dimension is about the solution. What should we do about the problem? For example, we roughly know the causes of the climate crisis, but there is great disagreement about the solutions. So that's another way in which a problem can be wild.

In the third dimension, we look at the actor landscape. If there are many actors working on the problem and it cuts across sectors and levels, it may indicate that the problem is more wild.

The fourth dimension involves a polarization measure, which assesses whether there is political disagreement about the problem. Here, we select a representative sample of citizens in the political unit that the problem concerns, such as Holbæk Municipality or Region Zealand. We then ask these citizens how significant the problem is and how much they would pay to solve it.

- When can we start using the model?

We can more or less do that now, but there is a lot of validation work ahead of us. I like to say that it's like dry swim - children first learn to swim on the edge of the pool, and then we have to see what happens when the child enters the water. Whether they swim or sink. The model is at that stage now. I've been working on it for two months and have developed a model on paper, but we haven't tested it yet. There is probably some technical stuff in the calculations that needs to be refined. And if the model says that it's harder to build infrastructure than to solve the climate crisis, then something is wrong.

- What can the model do in the world?

In the research world, I haven't seen anything that has tried to measure wild problems before. Some have asked subjectively about the phenomenon, but I would say that this is a huge contribution to the international literature on wild problems. It will be a model that can be applied across countries, problems and sectors.

Hopefully, the model will result in better problem solving. Not all problems can be fixed with a hammer. Some require a screwdriver, others require conversation or more time. Some have many causes and require the involvement of a wide range of stakeholders. The goal is that the model can shed light on this so that we can design tools that are better suited to solving the problems we face.