In Germany, package holidays are an important driver of consumer prices. Several challenges arise when measuring the price development of these bundled travel and accommodation services, such as the quality of accommodation and the timing of booking. Statistical practices are currently based on sampling offer prices. By using actual bookings, this paper analyses the possibilities and challenges in compiling a price index out of transaction data for flight package holidays. Our dataset comprises both online bookings and bookings made via stationary travel agencies on a daily basis. The large sample size allows for a disaggregation by individual holiday destination. Several methodological issues such as product definition, the grouping of unstructured text information, and weighting are addressed. Moreover, various index aggregation methods are analysed, which include hedonic regressions, stratification, and also a multilateral index method. Applied to six major holiday destinations for German travellers, all transaction-based methods under consideration exhibit similar price dynamics, pointing to robust results for destinationbased price indicators for package holidays.
