"Why does working longer hours not improve the situation? Because working longer makes you less productive at the same time that it encourages bad practices by your boss. Working fewer hours does the opposite."
What I would like to see is tangible evidence. Perhaps analyzing Jira for several companies to see the effort put into projects vs the overall productivity. Maybe set the baseline at the 40 hour week and analyze a few different chunks like 30, 35, 45, 50, and 60 hours per week. Without some sort of analysis like this all we have pure conjecture, and no way to understand where a reasonable cutoff should be.