As I understand it, this subhorizontal (~ parallel
to backbone) orientation
of the femur in birds is associated with the
position of the center of
mass/weight (CM) in modern birds.
Another important component to this posture is the large antitrochanter of
modern birds that articulates with the femoral head. Most non-maniraptoran
theropods had a well developed supracetabular crest instead, which
articulated with the femur when it was vertical. Archaeopteryx happens to
have among the largest supracetabular crests of all maniraptorans, plus
basically no antitrochanter. So it probably had a theropod-like femoral
orientation.