The theoretical model in Section 2 suggested that endogenous quality may
produce bias in the estimated coefficients. It is also possible that the proportion of
GP fundholders may not be exogenous Baines and Whynes, 1996 since GPs Ž .
could choose to become fundholders or not and their decisions may be related to
characteristics of their practices and of the FHSA. We tested for simultaneity by
re-estimating the fixed effect cost function using instrumental variables. The
Davidson and MacKinnon 1993 test did not reject the hypothesis that both Ž .
quality and the proportion of fundholders are exogenously determined