For example, if an athlete back squats (loaded with a barbell) to top of the thigh parallel and can maintain a "neutral" spine, but going any deeper creates lumbar flexion (posterior rotation of the pelvis), then the athlete should squat no deeper than the position in which they can maintain the neutral spine position. The thought is that any lumbar flexion under load may be bad.
In my situation, I have always had athletes squat as deep as possible (maintaining proper foot pressure, and in neutral spine). Many times, as the athlete drops the top of the thighs past parallel, the pelvis basically 'tucks' under, creating some lumbar flexion. I've always coached the athletes to maintain a tall spine, but have never really worried too much about this little bit of lumbar flexion at towards the end range-of-motion... maybe I should, but I have yet to see it cause any issues.
And most of the time, this little bit of lumbar flexion, actually brings the lumbar spine closer to a more true "neutral" position or just slight flexion, as back squatting with load tends put the athlete in more of a lumbar extension position to start with.
Anyway... just the other day, I came across this interesting paragraph in Functional Anatomy of the Spine by Middleditch and Oliver:
"Tension in the thoracolumbar fascia can also be increased by motion of the arms, legs and trunk. Posterior rotation of the pelvis causes an increase in flexion at he lumbosacral junction and hence increases tension in the thoracolumbar fascia. Lumbar spine flexion also passively increases tension in the thoracolumbar fascia. Competitive weightlifters may flex or 'round out' the lumbosacral area and it is suggested that this affords some protection by increasing tension in the thoracolumbar fascia and posterior ligamentous structures, thereby contributing to stability in the low lumbar area (Dolan et al, 1994)."
Maybe that little bit of lumbar flexion under load is ok.