A solid state amp will put out less power maximum power into a 16 ohm load, which of course is an issue if the amp may be pushed into clipping. But if the amp will produce adequate headroom into a high impedance load, in my opinion there's a valid argument in favor of relatively high impedance speakers.
If you look at the distortion curves published by Stereophile at 8, 4 and 2 ohms, you will see that below the onset of clipping the distortion is usually lower into a high impedance load. Due to a psychoacoustic phenomenon called "masking" the distortion at low power levels is of considerably greater subjective consequence than distortion at high power levels, so you want to focus on the low-power end of the distortion curves (again assuming you're not driving the amp into clipping). Personally I'd like to see what's happening down at the milliwatt level but that information isn't given.
Let me just comment that THD measurements are not reliable indicators of relative sound quality from one amp to another, as design choices that minimize THD are often counter-productive from a subjective standpoint (small amounts of high, odd-order distortion are far more audible and objectionable to the ear than are very large amounts of second harmonic distortion; in other words, the industry is meauring with the wrong yardstick). But THD measurements made on the same amplifier under different load conditions are reliable indicators of relative sound quality because they tell us how close to ideal that amp's performance is into that load.