Could Monero switch to a proof of stake model?

POS is fundamentally flawed. It increases centralization because people won't want to validate themselves and instead delegate their stake to "trusted institutions". Essentially, you're recreating the existing financial system. POS is also less secure since you only lose X dollars for misdeeds but have the power to do far more than X dollars of damage. And if all you want to do is damage, that slashed stake is just the cost of doing business. And if people are staking with you, you don't even risk losing your own money. And although the "guarantee" that the validator is doing right comes form the slashing of the stake, the enforcement of that slashing is political (i.e. others have to gang up on you to take your stake). When enforcement is political, wrongful slashing is inevitable. Finally, when a 51% attack is possible with validators, it's imposible to undo it...unlike POW which can call on the community to start a few miners on their PCs.

But mining is sooooo fun!!! ⛏️ Especially with p2pool and https://xmrvsbeast.com

Living in a mud hut is more energy efficient.

Could monero move to proof of stake? Sure. Would it? Probably not. There are several considerations to understand and it's not as simple as "use less energy."

Instead of POS (see my previous post), it would make much more sense to adapt the Nimble Wimble approach of making it so that validation is P2P whenever both parties are online. If I buy from you and you agree, then the purchase should been validated with 2 confirmations. Since we both agree that the transaction is made, it's no-one else's business if the transaction is valid. This would be much less resource intensive than even POS and be faster. If only one party is online at a time, it should be possible to have 1 confirmation and the other confirmation can be delegated to a miner. If neither are online at the time, then the usual POW takes place. Of course, that would mean that all wallets would have to be validators, but since you're only validating your own transactions, it should be light weight.

While I agree that a transaction is mostly the business involving the two parties transacting, they still need to adhere to the network rules. And that's were the miners step in. They don't care which two persons are transactig as long as they pay for it and it's in general a valid transaction (according to the network rules).
I agree with this idea

You do realize that proof of work uses a negligible amount of energy compared to banking institutions, right? Its not even worth talking about because its a non-issue.

The real question is: why dont banks just use Monero and get rid of all their wasteful infrastructure?

Banks are hypocritical. They cry about crypto spending too much energy without even bothering to look at their energy usage.
- What? Yes, they looked. And they spread a ton of disinformation that was extremely effective
  
  The fact that this question exists is testament to them.
  
  Its the same for the tobacco industry and the oil industry. They know that they're the big polluters killing people. The looked, and they hired PR companies to spread disinformation.
Any source on this?
- https://bitcoinmagazine.com/business/bitcoin-energy-use-compare-industry
Missing the point

As far as I know PoW leads to more decentralization (especially in combination with ASIC resistance) and therefore a more secure network. Energy efficiency comes through advancements in chip manufacturing.

Improved energy efficiency only results in a higher hashrate with the same amount of energy consumed. Assuming electricity costs dominate the mining expenses, in an efficient system the value of consumed energy should equal the value of mined coins. If it's significantly less, more miners will join until everything is balanced again.

On the flip side this means it's very easy to calculate the running cost of a financial system based on XMR: it's roughly 0.8% of the supply per year, slightly decreasing in the future thanks to the tail emission.

If you use the same metric to determine a cost for the current central banking system, which targets a real inflation of usually 2%, but it's more like a 4%+ cost of living increase for necessities, meaning using monero is AT LEAST 5 times as efficient. However this is ignoring the fees banks demand at every opportunity, so I'd estimate monero is about 10x as efficient.
- Fair point. The market will find a equilibrium between electricity costs and mining reward. Also compared to the current banking system miners will probably move their business mostly to locations where energy is cheap and highly available.