How about a year, did you test that?
It's a poorly conveyed study at best. From the article I would imagine the actual results are that the rate of people quitting drops off sharply after 90 days.
The same reasons anyone stays in any job. If the research is accurate, there's a significant correlation between staying 90 days and staying a lot longer. Which is much more significant than the obvious fact that staying 90 days makes it more likely to stay 91, and staying 364 makes it more likely to stay 365.
It is, but correlation is not causation. All other things being equal, the first 90 days is the hardest to learn a new job and become comfortable with it. That's the point of the article. But, when companies manipulate the first 90 days, they're manipulating the outcome. Sure, people are motivated to stay 90 days to get that sign on bonus, but if your pay literally drops after you learn the job, and Wendy's is offering another sign on bonus, it's not that big of a shift to go from McDonald's to Wendy's for another bonus.
Companies are going to keep having high turnover if they treat new employees better (including paying them better with bonuses) than their experienced people that can do the job in their sleep.