Can somebody show me some data or a paper to support this?
It's simple. When a miner mines a block, they are the first to have verified that block; it happens during mining to check for the validity of the transactions. When they broadcast their block, the rest of the miners must firstly verify it before mining on top. The bigger the block size, the more time it takes to verify the block, and hence, the more the time the lucky miner gains to continuing mining alone.
And it's not just verification, it's propagation as well.
FYI, Compact Blocks (
https://bitcoincore.org/en/2016/06/07/compact-blocks-faq/) partially mitigate this problem as you don't have to propagate whole block and the TX already verified when it arrive on your node.
On one hand, bitcoin remaining to be in this 1 mb size limit ensures that decentralization is taken care of
Actual block size limit isn't 1MB, but rather 4 kWU (4 million weight unit) or 1 vMB (1 virtual mega byte) where actual size of block can reach up to 4MB.
and that no one has to pay so much fees in the future.
How can you be sure when in this week you need to spend few dollars to ensure your TX included in next few blocks?