For example, why can't I (tenured professor) stop submitting journal papers? In Spain we have a research assessment every six years where basically only indexed journal papers count (even in CS where actually journals are quite irrelevant compared to conferences, but that's a different war of mine...). If you have enough indexed journal papers, you get a pay rise. If you don't, you not only don't get the rise, but you have to teach a punishingly high amount of class hours, basically leaving no time for further research. So for any professor that wants to do research, passing this assessment is pretty much needed. And the overwhelming majority of indexed journal papers are in Elsevier and the like, and it's highly likely that it will be the only thematic fit for a given paper.
Similar mechanisms are in place for tenure requirements, recruiting, grant calls, etc. Changing them is difficult. I do try to fight for it as much as I can, but it won't happen in the short or medium term.
And then there is bibliometrics, the art of lying with publication statistics. Everything humans get wrong is even worse there. A. Aaronson is different from Al. Aaronson, right? No paper has more than 3 authors, right? Journals always count for something, conferences don't, right? etc.
In Spain, gift authorship (adding authors that didn't do anything, typically in a mutual arrangement) is not uncommon. Not predominant, but definitely not uncommon.
And why do publishers like Elsevier hold most of the indexed journals? Pure inertia. We keep sending papers there because we need to for reasons like the ones explained above, then our papers get citations, so those journals keep their impact factor high. And it is extremely difficult for new entrants to break into the rankings, because sending a paper to a new journal that is not indexed yet has a huge opportunity cost (months of work that you aren't going to get recognition out of for tenure, promotion, grants, etc.) so unless they start with really great momentum (support from a powerful scientific society, an open letter by big shots in the field, etc.) they never get the needed papers and citations to climb the rankings. So it's not that publishers do a better job so their journals have higher impact factors, it's that they have a captive market.
Why don't people just do this then? The majority of credible papers still gets submitted to Elsevier and alike publishers (and the fact a paper is published there itself boosts a paper perceived credibility even though many papers they publish happen to be bullshit) and papers published on people's personal websites only are not taken serious. It seems like people still need Elsevier for some reason.