Various models have been proposed to explain the formation of the Ibero-Armorican Arc (IAA) as either primary or secondary (rotational). Models for secondary arc formation require significant rotation around a vertical axis at the end of the Variscan orogeny. Estimates for the amount of north–south (present coordinates) contraction (horizontal shortening) required for these rotations range from 54 at the core to 91% at the SE tip of the arc, near to the present day Mediterranean coast. These estimates are compared with coeval deformational structures developed in three areas of the orogen: (1) in the autochthonous hinterland in the southern branch of the arc; (2) in the Cantabrian Zone foreland fold–thrust belt in the core of the arc; and (3) in the SE continuation of the southern branch of the IAA in the present day Iberian Chain. From this analysis it follows that the contraction associated with late Variscan deformation is much less than that needed to explain the formation of the IAA from a formerly linear orogen via rotation around a vertical axis. We therefore propose that the geometry of the arc should be re-evaluated. We discuss the palaeomagnetic data purported to support the secondary origin, as well as the role of late Variscan regional strike-slip faults in modifying the geometry of the Iberian and Armorican massifs. Our analysis suggests that the IAA formed mainly from a primary non-rotational margin-controlled curvature, slightly modified by c. 5–20% of superposed contraction during late Carboniferous and/or Alpine times.