I think it is caused by the perspective
The track width at the rear roadwheel, as seen in a picture, would appear narrower than the track width at the front road wheel if measured with a ruler on the surface of the picture (a photograph in a frame is a one dimensional image of whatever was in front of the camera when the picture was snapped).
Tracks narrow to a point in the far distance, the height of the power line poles approach zero, the trees by the horizon are shorter than matchsticks.
Now consider the track from the sprocket down to the first road wheel
Is it totally vertical?
The first road wheel is further away from the camera lens than the sprocket,
the laws of perspective will therefore induce a very slight toe-in in a photo.
Technical drawings add further complications (perspective view, isometric view, oblique view …)
Simple example:
Further complications:
Back to topic.