People constantly forget that large multinationals like Amazon are bigger that many countries, are purely authoritarian structures and work as 100% centralized economies.

> Central panning doesn't scale

On the contrary, it scaled pretty well in China and in Soviet Russia, leading to the two cases of the fastest economical growth on the planet. The problem is not scalability.

The problem is that dictatorships (both countries and private companies) exist to benefit those in power. When push comes to shove everybody else is expendable.

Take Walmart as an example. If a country, it would rank 25th economically (above 188 other countries and the likes of Austria, Argentina, Norway, Ireland, and South Africa).[1] Walmartian central planning scales at least that far.

1. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3524078

Nobody is claiming the central planning doesn't produce value - i think the language is ambiguous as to the meaning of "doesn't work".

Central planning fails to give the little people what they want. Walmart's central planning is to ensure the profits continue to flow up to the share holders. The workers are expendable, and get a bare minimum the profits (as wages).

Central planning in a communist regime is exactly the same - it enriches the regime, at the expense of the people.

Central planning is the ultimate version of monopoly. It is what happens when a single company owns the entire country and no longer has to compete for anything. It is the only employer and the seller of goods, for every good. They banned the import of other goods, so they are all you get. Also they make the laws, so you can't even protest without having tanks moving you down.

"I really hate these corporate overlords, what can we do to fix it? I know, what about giving the corporate overlords all the power in the entire country? Yeah, that will fix it!"

> large multinationals like Amazon

I don't know about the internal workings of Amazon specifically, but many corporations are set up as hierarchical "business units" that each operate as separate companies within: selling their products and services within the company. There's still a market, not everything is determined centrally.

> On the contrary, it scaled pretty well in China and in Soviet Russia, leading to the two cases of the fastest economical growth on the planet. The problem is not scalability.

China grew much faster after it instituted market reforms. I'm not as familiar with Soviet Russia, but I'd be surprised if it grew faster than similar market-based countries during the same time period.

Sears mostly collapsed because they decided to make their internal departments independent competing companies. The departments acted more in their own interests rather than in the interests of the whole company...

> but many corporations are set up as hierarchical "business units" that each operate as separate companies within

The overwhelming majority of companies, including Amazon, does not allow different units, departments or teams to compete with each other by providing the same service.

Essentially each BU or even team is a little monopolist.

Soviet Russia actually had to reintroduce markets (https://en.wikipedia.org/wiki/New_Economic_Policy) to dig itself out of the economic hole caused by the hyper-centralized planning of "military communism".

You must not be from a former communist country. The system worked for two-three decades during the initial industrialization phase. When that phase ended growth was hard to find (the whole system was based around factories and moving villagers to cities). That's when the numbers started going down and the system started faking numbers to give the appearance that all is well and nobody could disagree with them.

At a country level (not talking about Amazon) these systems are fragile and don't handle volatility well.

Re: Amazon - you can't compare countries with private corporations.

I wonder if the growth problems that SU came across was more due to external pressures than inherent systemic limitations of a planned economy

> The system worked for two-three decades

Yes, that's my whole point.

> Re: Amazon - you can't compare countries with private corporations.

...and why? The similarity is more than evident.

Amazon has competitors, a central government do not. If the central planning fails to deliver a product then there is nobody there to deliver the product, if that product was food then people starve. This lack of competition is both bad for robustness as I described, and bad for efficiency since it means that people has to get their goods from the government and can't choose to get it somewhere else, so there is no selective pressure to keep the government deliver goods people want. You could say that other countries pressured them via war, but the people couldn't choose to buy American, the competition happened at a way too high level to matter.

Or rather, ultimately the people made their choice, central planning was scrapped and people could but American goods. So in a way the competition helped, people ultimately picked the superior option and the inferior option went under. But it took a very long time for that to happen, capitalism is way faster.

> Amazon has competitors, a central government do not

Countries engage in international trade and also go to war. If anything they compete more ruthlessly.

We definitely do broad-strokes central planning in the US and every other developed country with tax system design, social support systems, and industrial policy.

The Soviet Union sucked, but it lasted an extremely long time.

The moral of the story is central planning kinda does work, just not very well if you do too much of it.

I think it's more that you need layers. A centrally planned layer of regulation and safety nets that provide a base level to keep things running (and people alive), but without the cost and inefficiency of trying to control everything, and then a more efficient, free market–like layer on top that relies on the lower layer to provide the "free" in "free market".

I believe it would be more productive to argue about where the layer boundaries should be, rather than endlessly arguing about whether one or the other layer should even exist. (Because they'll both always exist. People will help each other out even in a free-for-all; and black markets will always come into existence in rigid, fully-planned economies.)

How are the current economic agents solving the NP-hard problem of economy? It seems like they are are just as incapable as a computer (assuming the substrate independence of computation).