> All calculations of this sort are fatally flawed because they assume all the coins in one address are owned by one person.

But don’t the calculations equally assume coins spread out across multiple addresses aren’t owned by one person, when in fact they often are?

Sure but that doesn't offset the first problem, it's just an additional problem. In fact I believe it could move the fake Gini coefficient in either direction depending on how each person splits their money.

If many coin owners share the same address, that is not very decentralized, is it?

that's like saying that the internet isn't decentralized because lots of people interact using facebook. that address isn't decentralized, much like facebook isn't. that doesn't mean that bitcoin or the internet aren't decentralized.

"Decentralized" means little without more context. The issuance of bitcoins isn't a legal monopoly. In this sense Bitcoin is "decentralized". On the other hand the Bitcoin blockchain is a centralized ledger, which is distributed, but nonetheless centralized. So which one is it? Centralized or decentralized? It doesn't matter because at this stage "decentralized" is being used as a mere buzzword rather than to convey a precise meaning.

> that's like saying that the internet isn't decentralized because lots of people interact using facebook

Yes, and I do say that. Decentralization is a lot less interesting if most people just end up centralizing anyway.

Who shares a bitcoin wallet or operates one like a bank? If anything the opposite is more likely, several bitcoin wallets belong to single individuals, probably in particular for some of the high net worth ones in an attempt to obfuscate ownership, which if anything, understates how concentrated it is.

> Who shares a bitcoin wallet or operates one like a bank?

Almost every bitcoin exchange.

It's typical to pool user funds into a relatively small number of addresses.