In
the Netherlands we have a combined system.
Upon leaving primary school, children take a centralized test, the CITO-toets; both this test and their normal classroom grades, plus the teacher's personal assessment, are used in determining which school type they can progress to.
Final exams for secondary school consist of nationwide exams (the centraal schriftelijk), plus a set of non-centralized exams taken through the year (the schoolonderzoeken). Both account for 50% of the total score. The resulting scores are often looked at in determining continued education for the student, but they are rarely (if ever) used as a formal entry level entrance criterion; instead, access to popular studies such as medicine is determined by chance.
Some years ago, we had a widely publicized case of a student with almost unprecedented exam scores who wanted to go to medical school, but failed the lottery two years in a row. The system didn't make an exception even for her.
Note that practically all of our educational institutions are state owned or state funded and participate in this selection system.
This case brought some discussion on how unfair this system is, leveling the field at the expense of the students with special talent, but few participants tried to make a serious case to have the system abolished. I don't expect to see any changes any time soon. Meanwhile, grades and centralized tests do play an important informal role in school selection, as indicators of the student's ability.