Models as the basis for a conversation

The other week I sat down to work out what on earth was happening in our economy. At the time the (now former) Chancellor of the Exchequer and (now former) Prime Minister had announced a (now rescinded) financial plan to which an awful lot of financial people took great exception. That much was clear. But this impacted real people’s lives, and there was also a lot of talk about bond yields and pensions and people’s mortgages and the value of the pound, which clearly connected to this and each other, and it was not clear how—at least, not to me.

So I drew a causal loop diagram (using some extra guidance). I worked my way through a number of explanatory financial articles, and tried to represent each statement in each article with a suitable arrow, or sequence of arrows, linking the various concepts.

My main goal was just to be able to understand the cause-and-effect relationships between the various concepts. But I also had a vague hope that it would be possible to look at the diagram, point to a bit of it, and say, “You see? There! Look at this sequence of events. It’s obvious that if you do this thing then that will be the consequence.”

Although I was successful in the main goal, the secondary hope was lost. I suppose if economics was that easy there would be no economists. Instead, I found an awful lot of apparent contradictions. So for example, if investors sell more bonds then they have fewer bonds, therefore less income from their bonds, and therefore less ability to pay the pensions they owe. On the other hand if they sell more bonds then they have more money, and therefore… greater ability to pay pensions. It can’t be both.

So was my model wrong? Yes, of course. Partly because all models are wrong—they always lack some detail—and also because I probably just made some errors.

But it was also clear that such a model can be the basis of a conversation. Two of us can look at the diagram and discuss what might happen. We can point to the bit that says “greater bond sales” (actually the model says “negative bond purchases”), point to the contradictory consequences, and discuss which one will outweigh the other. We might disagree with each other, but at least we’re disagreeing on the basis of a shared understanding. That helps us focus our discussion. A visual model provides an anchor for some agreement, or at least clarifying disagreement.

Just as we can collaborate better through precise language, we can also collaborate better through common models. They help us understand each other better, and help us focus on our essential differences when we seek the best solutions to our problems.

Causal loop diagram as PDF