Identifying price-demand elasticity for air travel using empirical data is fraught with issues. The largest of which is the problem of endogeneity. In this paper, we introduce instrumental variables derived from flow traffic passenger numbers to overcome endogeneity. When analyzing the price-demand relationship using flight date-point of sale and booking date-days to departure level data, flow traffic has the ideal property of influencing ticket prices via an airline's inventory control function yet is uncorrelated with demand shocks in the origin and destination market of interest. Ordinary least square (OLS) regression models report that the demand of the given market is highly inelastic at −0.148. Implementing the 2-Stage Least Squares (2SLS) model with our proposed instruments, we find that demand is in fact elastic at −1.157 which is consistent with industry observations. The proposed model is the first to estimate price elasticity using granular level data combining revenue accounting coupons, booking and ticketing data. The elasticity estimates account for endogeneity and granular characteristics such as days to departure, booking and travel day of the week, point of sale, holidays and special events.