I have been told that praising a kid for effort vs. results is much more important. As a child, I was told I was smart and basically tried to coast as much as possible. I still have a lot of trouble learning how to do something new, especially because I have kind of internalized that I'm good at everything.
A bit of both is probably good. For example: in Belgium, in recent years a lot of kids are (often) given points for effort so they can pass a certain grade. More and more the case. I talked to a teacher friend about this, and while he thinks its great, I think it's ridiculous. Do we want a society of people who put in some effort yet achieve nothing? Now, not that I think the educational system shouldn't be flexible…
Employers are obviously interested in grades that reflect the objective skill level of the to-be-employed.
The problem is people subconsciously build up explanations for their observed performance. If a kid routinely has a hard time with maths, its more probable that they develop a causal model that puts themselves in a static role: "I'm just not good at maths". What inter-subjective grading does is strengthening this belief: Imagine this kid learning really hard for a test, then still getting a bad grade. Lesson learned: Effort makes no difference.
What you want school to archieve is the exact opposite. You want to build up a mental model where the personal effort is the differentiator between good and bad results: "I just didn't learn well enough" instead of "I'm just bad at maths."
Grading that way makes it hard to assess objective performance, but it enables the kid to get nearer to his/her optimal performance.
(The opposite is probably more fitting for HN: Kids who didn't develop this model in school because they were "just good in maths" often didn't learn to learn and later have a hard time in college and beyond. Me included :))
Ideally schools would allow for both: Relative grading all the way up to the last school year, then a final test assessing objective performance. That way you maximize the development of a mental model of self-efficiacy while also allowing employers and universities to decide wheter the person is good enough for the position in question.
(I'm an educational scientist by trade)