Regional climate models (RCMs) are used to dynamically downscale Global Climate Models (GCMs) to provide high-resolution projections of future climate change to better inform policy and decision making at the regional scale. However, biases from GCMs are transferred to RCMs and this can limit the usefulness of the regional climate projections. This paper investigates the influence of bias correcting 4 GCMs from the Coupled Model Intercomparison Project 3, for regional climate simulations over the CORDEX-Australasia domain using Weather Research and Forecasting. The GCM outputs are bias corrected against ERA-Interim re-analysis as a surrogate truth. Results show that over decadal time scales bias correction removes large systematic precipitation and temperature biases. However, bias correction also introduced biases where there were none, introduced biases of the opposite sign or enhanced existing biases in other regions in some instances. Additionally, several simulations had bimodal temperature distributions irrespective of whether bias correction was used or not and these were not present in the observations, further highlighting limitations of this method and the RCM. The dynamical mechanisms driving the changes in the biases are explored.