I've read several times that box speakers have much higher distortion than panels (ESL's)? The box itself is apparently the culprit with measured distortion of 4% or more not being uncommon. In comparison distortion on panels has been found in some instances to be almost unmeasurable.
For sure many speakers will struggle to perform as low as 4% distortion, expecially in the bass (where 20%+ distortion is pretty much standard). A lot depends on volume level, and as you lower the volume level requirements then many designs can perform adequately.
However, I think you are mistaken about box speakers being systematically a problem due to the "box". For example, top of the line ATC speakers give you 121 db SPL at 1 meter (with 10 db headroom) with a THD of less than 0.3%. Do you have any examples of panels that meet or exceed this kind of performance?