Dont get me wrong, I believe OP's research has some flaws if it came to that conclusion, I just havent had time to read it.
It's on-chain data, and there are no flaws in the math, in reality, it's even worse today, looking at the last 7 days, blocks were found are distributed as follows:
Foundry USA 30.3%
AntPool 27.7%
F2Pool 12.5%
ViaBTC10.4%
4 pools mined >80% of all blocks
Take this address as an example
38XnPvu9PmonFU9WouPXUjYbW91wa5MerL
Antpool blocks pay to this address, it has received 139,652
BTC so far, the research paper considers this address as a single person/miner, which is why the interpretation is wrong, but the math is right.
My issue with this specific argument is how long would it take word to spread of a pool acting maliciously?
People treat hashrate distribution centralization as something new to
BTC, but it's not, it has always been the case, it's always a few pools that control the majority of the hashrate, in fact, it has been worse at times, not too long ago I pulled the hashrate distribution history of 2018 or so, it was nearly 50% Bitmain, Antpool and the other pools they owned, even worse, the majority of mining gears physical location was in 2-3 Chinese provinces, so it's nothing new, these guys make a lot of money by playing honest, makes no sense to rekt billion of dollars of investment to double spend some transactions or attempt to fork the blockchain or anything stupid.