I’m a big fan of Tyler Cowen’s book Stubborn Attachments and the philosophy that he outlines there. One of the core principles is that economic growth is more important than any random good thing for a society to have, because in the long run a richer society will be able to afford more of the good things, and economic growth compounds, so by optimizing for economic growth you will perform better in the long run anyway. I basically agree with this and only want to ponder about tweaking little details around the edge.

Wealth Plus

Clearly you don’t just want to optimize for plain GDP. There are other things that are valuable. In Stubborn Attachments Cowen’s term for this is “Wealth Plus” - you don’t just count money in the bank, you imagine adding in some term for free time, happy personal relationships, the other good things in life, et cetera.

I think there is another dimension that matters, though. Some economic activity is an end in and of itself. You pay someone to give you a massage. They give you a massage. It feels nice. The end. Versus, some economic activity builds on itself. It creates something either directly or as a byproduct that will make more economic activity possible. Kind of like “infrastructure”. You pay someone to build a bridge, now you have a valuable bridge, but also the existence of this bridge makes more new things possible, even if you charge for the use of the bridge.

Imagine two economies. In economy A the “end goal” economy grows 2% a year and the “infrastructure” economy grows 1% a year. In economy B it’s reversed. You won’t notice immediately, but before long won’t new innovation become fundamentally more possible in economy B?

I feel like normal economists should be analyzing things this way but I can’t really find anything along these lines.

What Counts as “Infrastructure”?

Last summer there were lots of arguments over what counted as infrastructure because there was some bill funding infrastructure and people were fighting about what should go into it. This is related in spirit - what parts of the economy are worth extra investment in? Although, really I don’t want to try to philosophically influence politics. Just too zero-sum. I am interested in this question more because it affects what I would like to be working on, to invest in, or to encourage my kids to get into.

According to my system I am thinking about here, the “infrastructure” parts of the economy are the ones that help other parts of the economy when they get better. Video games are not infrastructure. Cloud services are infrastructure. Those ones are easy.

Recently China has been trying to steer their economy in more particular ways. Read Dan Wang’s 2021 letter if you haven’t. Are they doing it in the right way? In general, Beijing seems to be anti “consumer internet”. But I think a lot of the consumer internet actually should count as infrastructure.

For example, food delivery. To me, food delivery apps seem like great infrastructure. It just saves you time when you can easily get food delivered, instead of picking it up yourself or using some slower, more inefficient delivery method. Anything that saves people time seems like infrastructure - those people will then be able to spend more time doing other things. Like work more. So I think food delivery should count as infrastructure. And personally grocery delivery is one of my favorite technological advances of the past decade.

The same goes for child care. Child care is pretty clearly infrastructure by this definition because it helps people work. It’s easier to get workers if it’s easy for your workers to get child care.

Tougher Questions

Education is an interesting question. I’m not sure what to think of the Chinese crackdown on online education. In theory, education should be infrastructure. When people learn more and become more intelligent, it helps out all sorts of sectors of the economy that depend on intelligent workers. But, maybe in practice some sorts of online education are like a pure game to get a credential, rather than learning something useful.

Is social media infrastructure? It sort of lets you do something, but it isn’t really helping you work. It’s closer to being an addictive waste of time. It’s kind of insufficient to only think of it from an individual’s point of view, though. From an advertiser’s point of view, targeted ads are great infrastructure. They help a lot of businesses get going. It’s tricky, though, because I think the sort of businesses that are most helped by targeted ads are the sort of consumerist businesses that are “not infrastructure” themselves. Like ads for games, clothes, any sort of random item that you consume yourself. This almost hints at a second category - the sort of infrastructure that helps other infrastructure. Shouldn’t that be a better sort of infrastructure to invest in? Yeah, this line of reasoning both makes sense and kind of reminds me of eigenvalues. There should be some “first eigenvector” that defines the optimal type of infrastructure investment. Maybe there’s some normality condition that our economy doesn’t obey, though…

Is cryptocurrency infrastructure? I was excited about working on cryptocurrency application platform stuff, because it seems like infrastructure. Some of the modern trendy crypto things, like NFTs, seem like they are not infrastructure, and I am less into them. A work of art is not infrastructure, blockchain or no blockchain.

Are hedge funds infrastructure? My instinct is no but it probably depends on what exactly they are doing. There are certainly some financial services which make other businesses possible. And there are certainly some financial entities which are just making money via shuffling stuff around cleverly and nothing else really depends on them.


I think people should be proud to work on “infrastructure”. It is underrated because the creators don’t capture all of its indirect benefits.