It's not all about price. The determinants of occupancy rates in peer-to-peer accommodation A methodological contribution

Purpose This paper aims to identify the key drivers of occupancy rates in peer-to-peer accommodation. Design/methodology/approach The applied methodology fits the specific characteristics of this market segment: the peculiar distribution of the occupancy rate (a ratio characterised by a large share of zeros) requires the adoption of a mixed discrete-continuous model; the insidious issue of price endogeneity is dealt with a control function approach; the econometric specification takes into account the monopolistic competition, the relevant market regime in the hospitality industry. The model is tested on Airbnb listings in the Balearic Islands (Spain). Findings The occupancy rate of peer-to-peer properties in the Balearic Islands strongly depends on their geographical location and online reputation. There is a qualitative difference between two groups: listings with positive occupancy rates, which demand tends to be inelastic, and listings with zero occupancy. The authors found that the price is a not a statistically significant determinant of the latter group membership. Originality/value This paper applies a zero-inflated beta model, never used in previous analyses of occupancy rates, to provide a benchmark for future studies. This procedure allows the estimation of unbiased marginal effects. It, thus, offers important technical and managerial implications, as a wrong understanding of how occupancy depends on price would deliver ineffective managerial decisions. This paper highlights the importance of methodological choices, as coefficients are highly sensitive to misspecifications of the model.