We have used the relativistic multi-reference Moller-Plesset perturbation theory to calculate the energy levels of Fe XVI, including those of the autoionizing levels with a hole state in the L shell. Comparison of the resulting L-shell transition wavelengths with those from recent laboratory measurements shows remarkable agreement, i.e., agreement within the experimental uncertainties. Our calculation allows us to predict the wavelength of the second strongest 2p-3d Fe XVI line, which has not yet been directly observed in the laboratory, to be 15.266 A. This wavelength is within 0.0042 A of the strong Fe XVII line commonly labeled 3D. Relying on the high accuracy of our calculations, we have reassigned two previously identified lines and predict a different location than previously thought for the strongest Fe XVI magnetic quadrupole transition. Inspection of the spectra of Capella recorded with the transmission grating spectrometers on the Chandra X-ray Observatory yields features corresponding to the predicted location of the innershell excited Fe XVI lines. These features have not been identified before. Our analysis shows that these features are most likely from Fe XVI.