Why are the estimated bandgaps from electrochemical data using CV larger than the optical bandgaps?

How do you estimate HOMO and LUMO energy levels from CV, using the peak potential or the onset values of the electrochemical peaks?