Skip to main content
. 2023 Feb 6;97:104738. doi: 10.1016/j.tourman.2023.104738

Table 3A.

First stage regression results: Full model with r = 500 m [Model 1].

Spec. 2
r = 500m
ln COVID-19 Stringency Index (g,t)*POST63 −0.149***
(0.004)
ln Average Market Price (i,500 m,t) 0.628***
(0.002)
ln Number of Competitors (i,500 m)*Season(t) Yes
ln RGDP (g,t) 2.082***
(0.018)
ln CPI(g,t) −1.020***
(0.070)
ln Beds(i)*Season(t) Yes
ln Photos(i)*Season(t) Yes
ln Host's Properties(i)*Season(t) Yes
Multiplatform(i)*Season(t) Yes
ln Experience (i,t) Yes
Cancellation Policy(i)*Season(t) Yes
Instantbook(i)*Season(t) Yes
MidLong Rent(i)*Season(t) Yes
Constant −15.045***
(0.294)
Time Fixed Effects Yes
Individual Fixed Effects Yes
Number of Observations 2,205,347
Robust Standard Errors Yes
F 3419.176
R2 Adjusted 0.983
R2 Within 0.085
R2 Within Adjusted 0.085
AIC −340,204.744
BIC −339,776.127

Notes: a) Subscripts: i = Property, t = Time (Year-Week), g = Country, 500 m = Radius r defining the relevant market. Individual Fixed Effects stands for property-level (i) fixed effects included in the model. b) Dependent variable = ln Local Airbnb Demand(i,r,t). c) Estimates were generated by the means of Stata 17 (command reghdfe). d) ***p < 0.001, **p < 0.01, *p < 0.5, + p < 0.1. e) The reference baseline of time-invariant variables interacted with Season(t) is autumn. f) Coefficients of the full model have been excluded due to space constraints, yet are available from the authors upon request.