My simple non-engineer understanding - motor speed is controlled by switching power on and off, not by lowering voltage. The time of on vs off controls the "average voltage" and motor rpm. The ESC doesn't really lower voltage for throttle control, but controls the pulse width. Yep ESC for dummies, by one.
More heat is generated in the ESC at lower rpm, partial throttle.
For the real explanation this may help, ESC portion is toward the bottom.
http://www.rcuniverse.com/articles/uncategorized/greg-coveys-ampd-issue-21-brushless-basics/
Yes, motor speed is controlled by switching power on and off. No, it's not that simple.
The motor has inductance, which means that the current doesn't change immediately in response to voltage changes. ESCs these days switch rapidly enough that the output voltage is somewhat averaged -- so it kinda-sorta acts like a constant voltage.
If you DID switch things slowly enough for the motor to come up to full current on each cycle, then the motor would be the one feeling the pain, not the ESC. ESC's get hot because the transistors act like little resistors when they're on (conduction losses), and because they can't snap from being on to being off instantaneously, meaning that in the middle there's a medium current and a medium voltage across the thing (= lots of power for a little time, known as switching losses). . ESC designers pull their hair out trying to minimize both kinds of loss at the same time.
So you're saying lower rpm makes the ESC work harder, seems I've heard that somewhere before. I could try a 9-4 and 10,000 rpm and see what that draws out.
After all that refutation -- yes, lower RPM will make the ESC work harder (there will be higher switching losses, if nothing else). Checking the ESC temperature every twenty seconds or so for a couple of minutes after a run may tell you something interesting, as the heat migrates out from the inside. If keeping a fan on the ESC (aimed to flow through the heat sink) helps, then it's an ESC issue.