My coauthors and I provide a thorough, scholarly, and accessible treatment of the socialist calculation debate, emphasizing Mises' and Hayek's contributions as well as their relationship to other diverse thinkers, from Ronald Coase to Michael Polanyi. We also deal with some contemporary contributions.
Yes, and moreover, our preferences are largely opaque to our own self-reflection. Even if you were able to declare your quotas at the beginning of each year and know they'd be fulfilled, it would be far inferior to having things be continuously available at market prices.
Perhaps this is what Mises meant by “human action”. However vast an AI’s capabilities are, it can never simulate human action faithfully enough to plan the economy, since human action by definition involves often-irrational, arbitrary, impulsive decisions that all add up to a gestalt outcome
But assuming, as a thought experiment, that such a god-like AI existed, which knew every nook and cranny of economic activity, where everything was and the requirements to produce anything in any possible way - and hence could calculate the optimal path to maximize whatever criteria it was given (individual standard of living, health, whatever). Couldn’t a socialist argue such a command economy would be an improvement precisely BECAUSE there would be no human irrationality and unpredictability involved? No tulip manias or stock market crashes? I could see a loss freedom admittedly; everyone had better follow the optimal instructions coming from the computerized central planning bureau.
Indeed. But then it wouldn’t be a human economy, would it? It would be humans living inside an economy an AI deems optimal for humans. And for all we know, we actually need the quirky, arbitrary, impulsive decisions for a full life.
It’s like sports. Two teams of robots could “play” basketball perfectly. But it’s neither sports nor particularly entertaining at that point
I hadn't considered this aspect of markets--at least not explicitly. I'm sure Hayek would never endorse it, but market socialism would not fall prey to these same issues. I'm sure he'd have some issues with the whole "social ownership of the means of production" thing.
Great article. Since you're teaching undergraduates, you might consider offering this text as a resource: https://www.cambridge.org/core/elements/abs/socialist-calculation-debate/5E63749F9D34D065193DCF77FC9FD8A9
My coauthors and I provide a thorough, scholarly, and accessible treatment of the socialist calculation debate, emphasizing Mises' and Hayek's contributions as well as their relationship to other diverse thinkers, from Ronald Coase to Michael Polanyi. We also deal with some contemporary contributions.
That is a great book. I have Pete scheduled for an interview on my podcast to discuss that very book!
It seems to me that the missing piece for generating market knowledge is individual people’s preferences.
Yeah that is definitely a big part of it. And part of it is how those preferences interact with inventory and substitutes.
Yes, and moreover, our preferences are largely opaque to our own self-reflection. Even if you were able to declare your quotas at the beginning of each year and know they'd be fulfilled, it would be far inferior to having things be continuously available at market prices.
Not to mention things like intermittent self-deception.
Perhaps this is what Mises meant by “human action”. However vast an AI’s capabilities are, it can never simulate human action faithfully enough to plan the economy, since human action by definition involves often-irrational, arbitrary, impulsive decisions that all add up to a gestalt outcome
This is just an arbitrary assertion. Nothing actually stops an AI from doing this.
Except for the lack of knowledge and the absence of time travel technology. Other than that, you're right.
But assuming, as a thought experiment, that such a god-like AI existed, which knew every nook and cranny of economic activity, where everything was and the requirements to produce anything in any possible way - and hence could calculate the optimal path to maximize whatever criteria it was given (individual standard of living, health, whatever). Couldn’t a socialist argue such a command economy would be an improvement precisely BECAUSE there would be no human irrationality and unpredictability involved? No tulip manias or stock market crashes? I could see a loss freedom admittedly; everyone had better follow the optimal instructions coming from the computerized central planning bureau.
Indeed. But then it wouldn’t be a human economy, would it? It would be humans living inside an economy an AI deems optimal for humans. And for all we know, we actually need the quirky, arbitrary, impulsive decisions for a full life.
It’s like sports. Two teams of robots could “play” basketball perfectly. But it’s neither sports nor particularly entertaining at that point
That's a good summary. It's not an issue of compute, it's an issue of local knowledge.
I hadn't considered this aspect of markets--at least not explicitly. I'm sure Hayek would never endorse it, but market socialism would not fall prey to these same issues. I'm sure he'd have some issues with the whole "social ownership of the means of production" thing.
I agree. In Radical Market I think Posner & Weil make this mistake in the last chapter when they say a supercomputer could solve the economy.
I’ve always been surprised such good economists could say something that seems so wrong