TAC recommendations for North-East Arctic cod have traditionally been based on Fmax, but Fmed might be preferable to Fmax as a biological reference point. However, the current Fmed estimate of more than twice Fmax appears high, and there is reason to question the basis for the estimate. Furthermore, the robustness of Fmed and its related points Ftow and Fhigh to changes in the parameters used for their estimation has not been adequately investigated. The effect of the various parameters on the estimates of the biological reference points for North-East Arctic cod and haddock is tested. F0·1 and Fmax calculated on a per-recruit basis are strongly dependent on the parameters used to estimate them, especially the natural mortality. Fmcd is generally more robust and the estimates of F t a and Fhing are also relatively stable. A crucial factor in the estimation of Fmed for cod is whether the maturity is density dependent or not. It is suggested that the Fmed -family of reference points provides a better founded basis for management of these stocks than F0.1 and Fmax.