For most of my adult life, I had economics filed away as something other people did.

Not dismissively per se but more that I assumed it was a settled discipline, the way physics feels settled if you've never studied it. There are people who know how it works, and the rest of us operate on the outputs. I'd taken one economics module in university, learned about supply curves and rational actors, and concluded it wasn't really for me.

Then I ended up at an economic agency building trade analytics tools, which is a strange way to discover that economics is not settled at all.

I picked up Skidelsky's What's Wrong with Economics partly because I'd started noticing something uncomfortable at work: that the tools I was building involved choices I couldn't fully account for. What to measure. What to leave out. What the data was supposed to say. I kept assuming those choices had been figured out upstream, by people who understood the theory. Skidelsky made me think that was probably the wrong assumption.

His argument, roughly, is that mainstream economics narrowed itself over the twentieth century in ways that had real consequences. The method that won — heavy on mathematics, allergic to ethics, treating the rational self-interested actor as the baseline for human behaviour — was one particular bet about how to study economic life. It became so dominant institutionally that alternatives got squeezed out. And what got lost in that process wasn't just intellectual diversity. It was the capacity to see things the framework couldn't see.

The part I kept returning to was about the rational actor specifically. The idea that you can model human behaviour by assuming people maximise utility subject to constraints — it produces tractable models, and I understand why it won. But it also assumes a picture of human beings that I find hard to recognise in practice. People don't maximise. They're swayed by loss aversion and social pressure and exhaustion. They make decisions inside relationships and obligations that can't be reduced to preferences. The model has to simplify all of that, and the question Skidelsky keeps pressing is: what happens to the things you had to simplify away?

The Locke epigraph at the top has stuck with me. He's saying that for most of what actually matters, we don't get certainty — we get probability, judgment, the dim light of something that might be right. And that this is just the condition of being human, not a problem to fix.

What Skidelsky is accusing mainstream economics of is refusing that. Building formal systems that perform certainty instead. I think I understand why — certainty is useful, especially when you're trying to advise governments. But I also think there's a cost to pretending you have more of it than you do.

I'm still sitting with what this means for my own work. Part of me wants to ask what economic assumptions are baked into the tools I'm building, whose framework I'm working inside without realising it. But I also don't fully know how to ask it yet, which probably means I should keep reading before I start drawing conclusions.

What I can say as I'm typing this is that my relationship to economics has changed.

I used to think of it as a discipline you either belonged to or you didn't. Technical, settled, internally consistent. What I understand now is that it's more like a long argument that hasn't been resolved — about what an economy is for, what counts as evidence, what kind of human being you're assuming when you build a model. The version that got packaged into institutions is one answer to those questions. A bet that paid off. Not the only possible answer.

That shift is harder to sit with than I expected. It's easier, in some ways, to work inside a framework you trust than to keep asking whether the framework belongs there. But I think the asking matters. Even if I don't know yet what to do with it …