In this paper, we discuss the electronic structure of transition-metal compounds in light of a new theoretical approach using an Anderson impurity Hamiltonian. We arrive at conclusions concerning the magnitude of parameters such as the d–d Coulomb interaction (U) and the charge-transfer energy (Δ) for the transition-metal monoxides, based on theoretical calculations of the band gaps, the optical spectra, photoelectron and inverse-photoelectron spectral shapes, and comparison with experiment. We find that both Δ and U are large but comparable in magnitude.We discuss the implications for superexchange interactions and show that the observed trends in the Néel temperatures are now well described, whereas the traditional Anderson approach gave a qualitatively different trend.