This only measures relative mobility, not absolute mobility. It doesn't show at all that it's harder to move up in the world, but merely that it's harder to move up relative to someone else.
The absolute mobility measure isn't really indicative of what most people think of when they hear "social mobility". For example, in 100% rigid caste society, with any economic growth whatsoever, everyone will make more than their parents.
Relative mobility isn't indicative of people's intuitive notions of mobility either.
Consider an farmer living in a village where incomes for everyone range from 200rs/day to 250rs/day depending solely on random chance (weather, locusts, etc). Relative mobility is high - depending on your crop yield, you could go anywhere from the bottom 1% to the top 1%. And you have no chance to go anywhere besides your farm.
That's both a very short-term (years to days) and localized (in one specific sector of the economy) measurement. If you look at inter-generational mobility (as these studies do), and look at the level of mobility that you care about e.g. over a whole national economy (as these studies do), a son of such farmers who grows up to be in the same job as his father will be shown as moving around solely within the small slice of the class structure (a decile at most?) that makes up the farmers of the village. Even that much movement will only show up if his income is fairly consistently (averaged over at least a year) in a higher percentile than his father's. So yeah, I think relative mobility is a pretty good indicator.
