On a case by case basis yes, but in aggregate many people end up staying because they're risk averse. The game the company is playing is trying to walk the lin

On a case by case basis yes, but in aggregate many people end up staying because they're risk averse. The game the company is playing is trying to walk the line where the amount they're underpaying the risk averse who don't leave more than makes up for the turnover cost. From an engineering perspective we know this leads to worse outcomes in terms of velocity and code quality, but those things are not quantifiable in a way that satisfies the bean counters.