The relatively high rates of bulk carrier casualties in recent years, as well as structural features such as large deck openings, make this vessel type a suitable example for investigating the influence of hydroelastic modelling on predicting wave-induced loads and responses. Two- and three-dimensional fluid–flexible structure interaction models, due to their different degree of complexity and associated data requirements, can be used at different stages of the design process when estimating wave-induced loads, namely preliminary and detailed design stages, respectively. In this paper, therefore, two- and three-dimensional hydroelasticity theories are applied to predict and compare the dynamic behaviour of a bulk carrier hull, based on OBO MV Derbyshire, in waves. Both symmetric and antisymmetric motions and distortions are incorporated in these investigations. The three-dimensional structural model consists entirely of shell finite elements, representing all major external and internal structural components, whilst the two-dimensional model is generated using Timoshenko beam finite element and finite difference discretisations. Issues relevant to the structural modelling stage, for both idealisations, are discussed. The in vacuo dynamic characteristics are compared for all models, with particular emphasis on the influence of hatch openings, shear centre and warping on the antisymmetric dynamics of the structure. For the wet analysis the fluid–flexible structure interaction is carried out using two-dimensional (Timoshenko beam and strip theory) and three-dimensional (beam and shell finite element idealisations combined with potential flow analysis based on pulsating source distribution over the mean wetted surface) analyses. Comparisons are made between steady-state responses predicted by two- and three-dimensional models in bow quartering regular waves. It is shown that whereas the predicted symmetric dynamic responses obtained from two- and three-dimensional models are in good agreement, differences are observed for the antisymmetric dynamic characteristics. It is thought that this may be due to inadequacies in the beam models employed when simulating the global dynamic behaviour of this highly non-prismatic hull girder whilst allowing for the effects of warping.