Methods to determine the hip joint centre (HJC) location are necessary in gait analysis. It has been demonstrated that the methods proposed in the literature involve large mislocation errors. The choice should be made according to the extent by which HJC location errors distort the estimates of angles and resultant moments at the hip and knee joints. This study aimed at quantifying how mislocation errors propagate to these gait analysis results. Angles and moments at the hip and knee joint were calculated for five able-bodied subjects during level walking. The nominal position of the HJC was determined as the position of the pivot point of a 3D movement of the thigh relative to the pelvis. Angles and moments were then re-calculated after having added to HJC co-ordinates errors in the range of ±30 mm. Angles and moments at both hip and knee joints were affected by HJC mislocation. The hip moments showed the largest propagation error:​ a 30 mm HJC anterior mislocation resulted in a propagated error into flexion/extension component of about −22%. The hip abduction/adduction moment was found the second largest affected quantity:​ a 30 mm lateral HJC mislocation produced a propagated error of about −15%. Finally, a 30 mm posterior HJC mislocation produced a delay of the flexion-to-extension timing in the order of 25% of the stride duration. HJC estimation methods with minimum antero-posterior error should therefore be preferred.