Empirical studies on vertical price transmission in retail fuel markets commonly use average group data of petrol stations. In this paper a simulation approach is employed to illustrate that, in this case, the persistence of price responses tends to be overestimated. To explore the real extent of the problem, we apply the mean group (MG) and the mean group with common correlated effects (MG-CCE) estimators to individual data from petrol stations. When heterogeneity in the pricing dynamics is captured by MG and MG-CCE estimators, persistence of retail price responses becomes considerably lower than the typical OLS estimations from average group data would suggest.