The gains from a lot of the abuse vectors they've mention are either ad-fraud related (your suggestions don't help), or are ways you steal from other creators by taking a larger than intended share. That is true whether the pool is $N,000,000 dollars statically, or a larger fraction than intended of 10000 $2 pools per user.
The other alternative is a pay-per-listen business model, but this kills the company.
My suggestion would however work, if ad money were distributed for plays proportional to how paying users listen, as long as Spotify's fee is larger than the ads share of revenue. I.e. each paying user don't "decide" more than a dollar of royalties per dollar paid.
https://www.musicbusinessworldwide.com/great-big-spotify-sca...
They argue a guy in Bulgaria set up 1200 premium accounts. However, they don't have proof.
Yes, you still run into the issue where 30 second whale songs get a larger than intended share of the royalties from whale-song-listening users.
I'm not sure where the disconnect is here.
Ok. The gains for the abuse is capped at the subscription fee minus Spotifys fees, per account.