We study the incidence and extent of bribes paid to the doctors in the public health facilities which are cleverly identified using a nationally representative survey. The survey asks households about the fees paid to public doctors, not about the bribe, which makes it less prone to reporting bias. We find that though consultations are free in the public health facilities, 41% of the patients who visited them paid about US$2 as a consultation fee, which is about 16% of their total medical expenditure. Three interesting generalized results that we find on the determinants of incidence and extent of bribe payment are: First, bribe givers and non-givers are different in terms of observed characteristics. Second, the same doctors, when sit in a private facility, charge more, setting an upper limit of bribes at public facilities. Third, travel time of the service seekers is used as a price discriminating device by the public service providers. Results have important implications for combating corruption, especially in developing countries.