Purpose: The purpose of this paper is to provide a robust version of the Malmquist Productivity Index (MPI) in order to evaluate hotels' productivity levels during the Great Recession.Design/methodology/approach: Based on the order-m frontiers, we apply a robust version of an MPI. We decompose the productivity into three robust components. We use a sample of hotels operating in the Balearic Islands and Canary Islands, and we decompose and evaluate their productivity levels during the period 2004–2013. Moreover, we evaluate hotels' productivity performance during the pre-crisis period, the US subprime crisis period, the global financial crisis (GFC), the sovereign debt crisis period and the after-crisis period.Findings: Our findings show that productivity did not deteriorate due to the adverse effects of economic crisis. This is mainly driven by increased technical change and the ability to operate at optimal scales. The long-term investment in innovation policies which are related to services and processes, appear to be the dominating feature behind hotels' productivity levels, which helped the hotel industry to recover quickly from the Great Recession.Originality/value: The vast majority of empirical studies evaluating the productivity change in the hotel industry apply MPIs which are based on data envelopment analysis (DEA). However, the productivity measurement which is based on the nonparametric framework is sensitive to sample characteristics. In order to avoid such shortcomings, we apply a robust version of the MPI.