Also if it's an internal focus lenses the effective focal length (angle of view) will be less than 105mm
Dave
The CIPA standard is that the difference between nominated and actual focal length at infinity must be within 5%. If you look at Bill Claff's Optical Bench Hub you will see that the wide ends of zooms are consistently less wide and the long ends less long than the nominal focal lengths. The 24-120/4, eg, is actually 24.7 (= +3% discrepancy) and 116.5 (= -3% discrepancy); the 70-200/2.8 is actually 71.5 to 196. The 24-105 is not on the site (and may not be, because the data comes from published patents) but it would be remarkable if the pattern was not maintained.
Shortening of focal lengths at short focus distances = focus breathing is a different matter.
The angle of view is the angle the chief ray (the off-axis ray that crosses the aperture on the axis) makes with the axis in
object space. The formula that the tangent of the half angle of view = half image diameter / focal length applies only if the pupil magnification is unity, so IRL lenses with the same focal length can have different angles of view, the angle of view can vary with focus distance, and the relationship between focal length and angle of view will usually vary across a zoom range.