The implied modeling assumption in these studies - that governance affects audit fees but
not the reverse - seems unrealistic, however. Company boards often purchase additional or
higher quality auditing services (outside or internal) to effect better corporate governance,
which in turn influences how auditors audit. If governance choices affect auditing and
vice versa, that is, they are co-determined variables, traditional regression methods can
misrepresent what may be jointly determined positive and negative relations. Bedard and
Johnstone (2004), for example, find that auditors’ planned billing rates and chargeable
hours vary positively with assessed earnings management risk and negatively with the
interaction of assessed earnings management risk and reduced governance risk. This is
prima facie evidence of a joint determination. But it is unclear from the models in that
study whether audit fees are determined by auditors’ independent or joint (endogenous)
assessments of earnings management risk and corporate governance. Such assessments
are endogenously determined if, for example, auditors’ views of corporate governance
derive in part from the planning and conduct of the audit and vice versa. Assuming this
is a most likely and logical situation, additional structure is required for a most accurate
representation of the relation between audit fees and governance. This paper offers one
such structure or framework?