I expect part of the problem is that people who are attracted by clear writing and thinking are turned off by the academic fields that suffer most from overly academic writing.
There are some academics however, who are quite good at conveying their ideas, even in fields that are otherwise noteworthy for their opaqueness to outsiders. The economist Deirdre McCloskey is one example. I've read some systems CS papers that were quite good as well (Google has had several good ones, including this one [1]). Many academic papers are garbage though, and another reason is because they aren't saying anything worth noting. It's a lot easier to hide bullshit via obfuscation than through clarity.