Read the working paper
INSEAD Working Paper 2015/73/TOM (revised version of 2015/62/TOM)
We study price competition in markets with a large number (in magnitude of hundreds or thousands) of potential competitors, using the hotel industry as a test bed. We address two methodological challenges: simultaneity bias and high dimensionality. Simultaneity bias arises from joint determination of prices in competitive markets. We propose a new instrumental variable approach to address simultaneity bias in high dimensions. The novelty of the idea is to exploit online search and clickstream data to uncover demand shocks at a granular level, with sufficient variations both over time and across competitors in order to obtain valid instruments at a large scale. We then develop a methodology to identify relevant competitors in high dimensions combining the instrumental variable approach with high dimensional l-1 norm regularization. Our approach is in contrast to many existing applications of high dimensional statistical models that, using either regression or clustering techniques, are primarily concerned with correlation rather than causality. We apply this data-driven approach to identify price competition patterns in the New York City hotel market. We find: 1) engagement in competition-based revenue management is prevalent across branded and non-branded hotels and across all quality tiers. It explains 31.9% of the within-hotel price variation, an additional 7.5% on top of demand-based factors; 2) Branded hotels are both more influential and more influenced on prices; 3) When choosing whose prices to follow, branded hotels are more concerned about quality boundaries but less about geographical boundaries, as compared to independent hotels. Also, budget and luxury hotels are more concerned about quality boundaries but less about geographical boundaries, as compared to economy and up-scale hotels.