Perhaps Agency is Economical

I’ve been thinking a lot about agency lately, for the most part my lack of it.

Which, objectively, feels like a juxtaposition given the year I’ve had. I’ve made more progress, learnt more and put myself in more interesting environments than I have in any other year. Yet I still constantly think: what else could I be doing with this time?

My assumptions alternate between dissatisfaction and choice overload. More frequently, though, I’m wondering whether agency is more complicated than just having choices.

Agency tends to be treated as something psychological: discipline, decisiveness, the willingness to act. But a large part of it is material. Money doesn’t just buy things. It buys optionality. It buys the ability to leave a job, spend time learning something new, build something that might fail, or survive a bad decision without it becoming the be all and end all.

In that sense, agency isn’t distributed equally.

But there’s another side to it. More options don’t necessarily make choosing easier. Sometimes they make every decision feel like the rejection of ten other possible lives. I feel that increasingly now: work, learning, building, money, health, relationships, curiosity. Each one contains a version of life I could lean further into, and each hour spent on one is an hour withheld from another.

So perhaps too little optionality constrains agency, while too much can make agency harder to exercise.

I keep coming back to the idea that life is partly choosing what you’re willing to suffer for. Every now and again I return to Mark Manson’s question → what pain do you want in your life? Every route has its own cost.

But even that framing feels incomplete, because not everyone gets the same menu of suffering.

And I’m not yet sure how we’re supposed to know which costs are worth accepting.

I’ve been working on a project about Britain and how we imagine its future, asking real questions about our economic, political and social direction.

The idea surrounds turning vague collective beliefs into something visible, collecting evidence around questions that are surprisingly difficult to answer cleanly: housing affordability, productivity, wages, energy, migration and what the UK might look like five or ten years from now.

We say things like “house prices will keep rising”, “AI will remove certain jobs” or “Britain is getting poorer” quite casually. It gets more interesting when you have to put some structure around that belief: what exactly are you predicting, over what time period, and what evidence would change your mind?

I’ve got agents gathering new reports, datasets and arguments for me to work through. At this stage, I don’t necessarily care about finding an answer. I’m collating pieces and finding patterns. A chart can imply one story, while a different time horizon produces another.

I’m always interested in the foundations: what data gets chosen, what assumptions sit underneath it, and what would actually have to happen for somebody to change their mind.

I’ll share bits of the research and the build here as I go rather than the whole project.

This week, I’ve been using:

• Metaculus → — mostly to study how good forecasting questions are scoped and resolved.
• The ONS house price to earnings dataset → — for grounding broad claims about housing affordability in something measurable.

My main focus right now is defining the resolution criteria before the prediction. If nobody can agree afterwards on what counted as “true”, the question was badly designed in the first place.

More soon.

Maybe agency isn’t having endless choices.

Maybe it’s being able to decide which choices deserve the cost they ask of you.