They do and they don't...
First off, I do quite well in exams, yet I didn't revise for my GCSE's that much... I understood what was being taught, and therefore I just figured out everything in the actual exam rather than learn it. This is all well and good for GCSE's, but this does not work for A levels, which requires you to learn some basic facts that require too much outside understanding to understand the principals behind them, which would take me more than 2 years. Therefore, some basic learning is required.
Does this prove your intelligence though? If exams were all based upon skills testing, and thinking skills, yes, it does test your actual intellect. However the exam system we have now does not require you to actually understand anything for GCSE and maybe AS level, and you can get almost full marks by just learning absolutely everything. However, using chemistry as an example, A2 level requires full understanding and full knowledge to get close to 100%, therefore I would argue that A2 chemistry does test your intelligence.
However, upon saying this, you do not need to understand everything to get an A grade at A2, only the A* boundary really tests intelligence.
This begs the question "What is intelligence?". Is it knowledge, or understanding, or both? I would argue that one can appear intelligent by acquiring a vast amount of knowledge, yet one can only usefully utilise that knowledge if they understand it to the full extent, thereby knowledge without understanding is not true intelligence. Understanding however, automatically proceeds to knowledge, whereas knowledge does not necessarily proceed to understanding, unless the person with the knowledge has the ability to understand.
Due to this, I conclude that understanding is the basis for intelligence, and only upon the acquisition of understanding can one proceed to begin the hard work required to master their topic to make further use out of it.
Does the current exam system test this? I would argue not, as one can pass well without understanding. I would therefore argue theta exams should be set upon two stages. The first of which is a skills based understanding paper, alike to general studies, but subject focused, where a understanding of the subject and mechanisms is essential- a online interactive platform may be the best way to do this. I would then prepose a second knowledge based test, which it is possible to revise for.
I would take the fraction of correct answers from the understanding test and multiply it by the raw marks, or even a UMS equivalent for the knowledge test, giving a result in which higher scores are better. I would then rank these results, and take percentages of the ranks for the grades, to assess intelligence in that subject.
If anyone disagrees with any of my logic, I am willing to consider any amendments or corrections.
I hope my ideas on this matter have been interesting.