# RECORD fb8eee688ea40496609ebe82ad7180db4acffe2c # TITLE: The impact of short-term rental activity on house prices: evidence from coastal Virginia # YEAR: 2024.0 VENUE: The annals of regional science DOI: 10.1007/s00168-024-01328-4 # AUTHORS: Navid Safari; Lei Zhang; Timothy Komarek # PAGES: 33 CANDIDATE TABLE PAGES (1-based): [12, 14, 16, 17, 18, 25, 26, 27, 28, 29, 30, 31] # ABSTRACT: Short-term rentals such as Airbnb have grown substantially in recent years. In this study, we investigated the effect of Airbnb rentals on neighboring house prices in Hampton Roads, Virginia. We focus on three different measures of Airbnb activity to obtain more robust results. Our measures include active Airbnb listings, active listings with at least one reservation, and the sum of reservations over three different periods. Consistent with previous literature, our main results show that prior Airbnb rental activity has a positive impact on the surrounding housing sales prices. Furthermore, the localized impact diminishes over geographic space (i.e., distances greater than 500 m). We also examine how Airbnb rental activity impacts house prices across the price distribution of home sales, using quantile regression. We find that the effect is strongest for higher-priced homes, while lower-priced houses are less impacted by Airbnb rentals. ===== PAGE 1 (score 5, numeric-paren 0, table-words 0, var-words 5) ===== Vol.:(0123456789) The Annals of Regional Science (2025) 74:6 https://doi.org/10.1007/s00168-024-01328-4 ORIGINAL PAPER The impact of short‑term rental activity on house prices: evidence from coastal Virginia Navid Safari1 · Lei Zhang2   · Timothy M. Komarek3 Received: 7 December 2022 / Accepted: 5 December 2024 / Published online: 31 December 2024 © The Author(s) 2024 Abstract Short-term rentals such as Airbnb have grown substantially in recent years. In this study, we investigated the effect of Airbnb rentals on neighboring house prices in Hampton Roads, Virginia. We focus on three different measures of Airbnb activity to obtain more robust results. Our measures include active Airbnb listings, active listings with at least one reservation, and the sum of reservations over three differ- ent periods. Consistent with previous literature, our main results show that prior Airbnb rental activity has a positive impact on the surrounding housing sales prices. Furthermore, the localized impact diminishes over geographic space (i.e., distances greater than 500  m). We also examine how Airbnb rental activity impacts house prices across the price distribution of home sales, using quantile regression. We find that the effect is strongest for higher-priced homes, while lower-priced houses are less impacted by Airbnb rentals. Keywords  Short-term rental · Airbnb · House price · Hedonic model · Quantile regression * Lei Zhang l2zhang@odu.edu Navid Safari navid.safari@gcsu.edu Timothy M. Komarek Tim.Komarek@colostate.edu 1 J. Whitney Bunting College of Business and Technology, Georgia College & State University, Milledgeville, GA, USA 2 Department of Finance, Strome College of Business, Old Dominion University, Norfolk, VA, USA 3 Department of Economics, Colorado State University, Fort Collins, CO, USA ===== PAGE 12 (score 15, numeric-paren 0, table-words 4, var-words 3) ===== N. Safari et al. 6  Page 12 of 33 included in the regression analyses as a hedonic attribute. We have three sets of models with different assumptions for distance and time to explore the impact of Airbnb listings on housing prices. Table 2   Description and summary statistics for Airbnb densities (independent variables) and ln(price) (dependent variable). Airbnb properties and housing sales in Hampton Roads, Virginia (January 2015– December 2019) For each house sold, we drew three concentric rings with different radii: Ring 1, 0–300 m; Ring 2, 300– 500 m; and Ring 3, 500–1000 m Phase 1 includes the month of sale and the three months before that. For example, if a house is sold in August; thus, Phase 1 includes May, June, July, and August. Phase 2 includes the fourth month before the sale and three months before that. Suppose a house is sold in August; thus, Phase 2 includes Janu- ary, February, March, and April. Phase 3 includes the eighth month before the sale and three the months before that. Suppose that a house is sold in August; thus, Phase 2 includes September, October, Novem- ber, and December The mean of the inflation-adjusted house sale price (unlogged) is $242,111.90 Variable Description Mean Std. Dev Min Max Independent variables Measure 1: “Active” Airbnb Airbnb_Phase1 _Ring1 Number of Airbnb in phase1 and ring 1 0.803 4.15 0 138 Airbnb_Phase1 _Ring2 Number of Airbnb in phase1 and ring 2 1.028 4.229 0 143 Airbnb_Phase1 _Ring3 Number of Airbnb in phase1 and ring 3 3.797 12.036 0 267 Airbnb_Phase2 _Ring1 Number of Airbnb in phase2 and ring 1 0.645 3.537 0 136 Airbnb_Phase3 _Ring1 Number of Airbnb in phase3 and ring 1 0.669 3.727 0 144 Measure 2: “Positive Occupancy Rate” Airbnb_Phase1 _Ring1 Number of Airbnb in phase1 and ring 1 0.678 3.588 0 138 Airbnb_Phase1 _Ring2 Number of Airbnb in phase1 and ring 2 0.875 3.737 0 130 Airbnb_Phase1 _Ring3 Number of Airbnb in phase1 and ring 3 3.211 10.66 0 255 Airbnb_Phase2 _Ring1 Number of Airbnb in phase2 and ring 1 0.516 2.923 0 132 Airbnb_Phase3 _Ring1 Number of Airbnb in phase3 and ring 1 0.563 3.209 0 133 Measure 3: Sum of Reservations Airbnb_Phase1 _Ring1 Number of reservations in hundreds in phase1 and ring 1 0.080 0.481 0 31.13 Airbnb_Phase1 _Ring2 Number of reservations in hundreds in phase1 and ring 2 0.102 0.487 0 16.05 Airbnb_Phase1 _Ring3 Number of reservations in hundreds in phase1 and ring 3 0.368 1.372 0 37.27 Airbnb_Phase2 _Ring1 Number of reservations in hundreds in phase2 and ring 1 0.061 0.381 0 21.83 Airbnb_Phase3 _Ring1 Number of reservations in hundreds in phase3 and ring 1 0.052 0.341 0 24.05 Dependent variable P Natural log of the inflation-adjusted house sale price 12.25 0.558 6.767 15.21 ===== PAGE 13 (score 8, numeric-paren 0, table-words 1, var-words 6) ===== The impact of short‑term rental activity on house prices:… Page 13 of 33  6 4.1  Model 1 We start with a pooled cross-sectional model using Airbnb density in the clos- est ring (ring 1: 0 to 300 m) and in the closest phase (phase 1: months S to S-3) before house sale ( Airbnb_Phase1_Ring1). P is the natural log of the inflation-adjusted sale price for all the houses sold during the study period. The house characteristic variables contained in X are those described in Table 1. T are time fixed effects including sales year dummy variables to capture macroeconomic changes within a given year and sales month dummy variables to remove seasonality effects. M contains location fixed effects, including public school and zip-code-level fixed effects (M), to control for unob- served time-invariant characteristics that may jointly affect housing prices and Airbnb activities, such as commercial activities, infrastructure, and public facili- ties. Finally, ϵ is the error term. 4.2  Model 2—time effect In this model, we used Airbnb densities in the closest ring (ring 1) but for differ- ent time periods before the sale, namely, Airbnb_Phase1_Ring1, Airbnb_Phase2_ Ring1, and Airbnb_Phase3_Ring1, to investigate the impact of Airbnb on house prices over time. 4.3  Model 3—distance effect We further used Airbnb densities in the closest period prior to the sale (Phase 1: months S to S-3), but for different Rings, namely, Airbnb_Phase1_Ring1, Airbnb_Phase1_Ring2, and Airbnb_Phase1_Ring3, to examine if Airbnb impact on house prices varies by distance between house sales and Airbnb properties. (1) P = 훼+ (Airbnb_Phase1_Ring1)훽1 + 훾X + 훿T + 휃M + 휖 (2) P = 훼+ ( AirbnbPhase1Ring1 ) 훽1 + ( AirbnbPhase2Ring1 ) 훽2 + ( AirbnbPhase3Ring1 ) 훽3 + 훾X + 훿T + 휃M + 휖 (3) P = 훼+ ( AirbnbPhase1Ring1 ) 훽1 + ( AirbnbPhase1Ring2 ) 훽2 + Airbnb_Phase1_Ring3)훽3 + (+훾X + 훿T + 휃M + 휖 ===== PAGE 14 (score 29, numeric-paren 6, table-words 7, var-words 8) ===== N. Safari et al. 6  Page 14 of 33 4.4  Unconditional quantile regression Another goal of this study is to estimate the potential Airbnb effect across the complete distribution of house prices. To do so, we used the unconditional quantile regression (UQR) method proposed by Firpo et al. (2009). Compared to the (conditional) quantile regression method developed by Koenker and Bassett (1978), this methodology allows us to estimate the effects on an outcome variable that is not conditioned by the set of covariates included in the model (Fortin et al. 2011). UQR is based on extending the concept of the influence function to what has been termed the recentered influence function (RIF). The UQR model is estimated by regressing RIF on the covariates. where Pq휏 is the natural log of the real sale price, P, at the 휏 th quantile (q). 1 is an indicator function that takes a value of 1 for P below Pq휏 ; and fy(Pq휏) is the density of P at the 휏 th quantile. 5  Results We begin by presenting estimates from the hedonic model in Eq. (1), with results shown in Table 3. Columns (1) to (3) report estimates for three different measures of Airbnb activity in the nearest ring (ring 1: 0 to 300 m) and the closest phase (phase (4) RIF ( Y;Pq휏, FPq휏 ) = Pq휏+ (휏−1{P ≤Pq휏 }) fy(Pq휏) = (Airbnb_Phase1_Ring1)훽1 + 훾X + 훿T + 휃M + 휖 Table 3   Model 1: OLS regression estimates, all three measures Standard errors are reported in parentheses and were obtained using 500 bootstrap replicates. ⁎ p < 0.1; ⁎⁎ p < 0.05; ⁎⁎⁎ p < 0.01 The estimations correspond to Eq. (1), with the natural log of the inflation-adjusted house sale price as the dependent variable The coefficients of the control variables introduced in Table 1, are omitted from this table. A complete version of the table is available upon request Measure (1) Active Measure (2) Positive Occupancy Measure (3) Sum of reservations Airbnb_Phase1 _Ring1 0.0034*** (0.0002) 0.0035*** (0.0002) 0.0169*** (0.0018) Zip code dummies Yes Yes Yes School district dummies Yes Yes Yes Year of sale dummies Yes Yes Yes Month of sale dummies Yes Yes Yes N 102,142 102,142 102,142 adj. R-sq 0.789 0.789 0.789 ===== PAGE 15 (score 11, numeric-paren 0, table-words 4, var-words 2) ===== The impact of short‑term rental activity on house prices:… Page 15 of 33  6 1: months S to S-3) before the house sale, with standard errors in parentheses. To address potential heteroscedasticity and autocorrelation, we primarily use bootstrap standard errors for their flexibility and robustness. We also conducted additional regressions using clustered standard errors by ZIP code and month of sale, which showed only a minimal increase compared to our main bootstrap standard errors. Drawing on guidance from Cameron and Miller (2015) and Abadie et al. (2023), which recommend adopting clustered standard errors only if they significantly exceed robust ones, we evaluated the minor differences and decided to retain our main results. For comparison, Table 7 in the appendix includes Table 3 results with both clustered and bootstrap standard errors.5 The estimate in column (1) indicates that having one additional active Airbnb property within 300 m of a house in phase 1 increases the sale price by 0.34%.6 Considering that the average number of Airbnb listings within 300 m of a house in phase 1 is 0.803, as shown in Table 2, the average magnitude of the impact for the first measure is approximately 0.27%. Moving to column (2), we examine a similar measure of Airbnb activity—positive occupancy, where an additional Airbnb rental increases nearby housing sales by 0.35%. The mean number of Airbnb listings with a positive occupancy rate for phase 1 and ring 1 is 0.678, resulting in an average impact of approximately 0.24% for the second measure. In accordance with our findings, our results align with previous studies such as Barron et al. (2021), who reported a 1% increase in Airbnb listings resulting in a 0.026% increase in house prices across the entire USA. Koster et al. (2021) spe- cifically observed analogous outcomes in Los Angeles, indicating that a 1 percent- age point increase in the Airbnb listings rate within 100 m led to a 2.51% rise in property prices, while a similar increase within 500 m resulted in a 3.72% increase. Further reinforcing our observations, Garcia-López et al. (2020) conducted a com- parable investigation and found average increases of 4.59% in transaction prices and 3.67% in posted prices attributed to Airbnb activity in Barcelona. These consistent trends across different regions and methodologies contribute to the broader under- standing of the impact of Airbnb activity on housing markets. Finally, in column (3), we present the impact of the sum of reservations for all Airbnb listings over all four months in phase 1 (months S to S-3). Here, an increase of 100 more Airbnb reservations within 300 m of a house is associated with a 1.69% higher home sales price. The mean for the number of Airbnb reservations for phase 1 and ring 1 is 0.08 (scaled by a hundred), resulting in an average impact of approxi- mately 0.13%. This measure differs from the previous two because it assesses the intensity of guest traffic. We included this measure with the expectation that the number of reservations would better capture any negative externalities associated with short- term rental (STR) activity. While the average impact magnitude for this measure is comparatively lower than the effects observed for the first two measures, the overall positive trend suggests that the net effect of STR activity on house prices remains positive, indicating that the income stream received by short-term rental activities 5  Results for other tables using clustered standard errors are available upon request. 6  Percentage change is calculated using exp(β)-1. ===== PAGE 16 (score 42, numeric-paren 15, table-words 9, var-words 7) ===== N. Safari et al. 6  Page 16 of 33 is more valued by homeowners than potential negative externalities in the neighbor- hood. This observation aligns with the conclusions drawn by Kim et al. (2017), who utilized a differences-in-differences approach to examine the impact of restrictive regulations on rental rights for property owners in coastal areas of Florida. Their study revealed a negative effect on house prices following the implementation of such regulations. Their findings suggested that property owners are likely to prior- itize their right to rent houses without restrictions over potential negative externali- ties associated with renters in the neighborhood. We recognize the inherent challenge of endogeneity in estimating the impact of short-term rentals on house prices. To mitigate this concern, our analysis incorpo- rates zip code × calendar month fixed effects to account for potential time-varying amenities that could be correlated with both house prices and Airbnb activity. Additionally, Year × Month fixed effects are included to fully capture the time-var- ying trend in house prices. These results combined with our main result discussed above are illustrated in Table 8 in Appendix. In each 3-column set for each meas- ure, the first column is our main result from Table 3 which does not incorporate zip code × month and Year × Month fixed effects. The second column includes Year × Month but omits zip code × calendar month fixed effects, and the third col- umn includes zip code × calendar month but excludes Year × Month fixed effects. As illustrated in the table, the results from each 3-column set corresponding to each measure of Airbnb activity are very close and similar. This pattern holds true for the results in our other tables as well. To maintain simplicity and avoid confu- sion, we present only the results which does not include zip code × month and Year Table 4   Model 2: Time Effect, all three measures Standard errors are reported in parentheses and were obtained using 500 bootstrap replicates. ⁎ p < 0.1; ⁎⁎ p < 0.05; ⁎⁎⁎ p < 0.01 The estimations correspond to Eq. (2), with the natural log of the inflation-adjusted house sale price as the dependent variable The coefficients of the control variables introduced in Table 1, are omitted from this table. A complete version of the table is available upon request Measure (1) Active Measure (2) Positive Occupancy Measure (3) Sum of reservations Airbnb_Phase1_Ring1 0.0033*** (0.0006) 0.0019*** (0.0005) 0.0101*** (0.0030) Airbnb_Phase2_Ring1 − 0.0006 (0.0008) 0.0011* (0.0006) 0.0067* (0.0040) Airbnb_Phase3_Ring1 0.0008 (0.0007) 0.0012** (0.0005) 0.0055 (0.0038) Zip code dummies Yes Yes Yes School district dummies Yes Yes Yes Year of sale dummies Yes Yes Yes Month of sale dummies Yes Yes Yes N 102,142 102,142 102,142 adj. R-sq 0.789 0.789 0.789 ===== PAGE 17 (score 45, numeric-paren 18, table-words 10, var-words 5) ===== The impact of short‑term rental activity on house prices:… Page 17 of 33  6 × Month fixed effects. The complete results for Tables 4, 5, and 6 also are shown in appendix as Tables 9 to 11. The results in Table 4 explore how Airbnb activity in different periods before house sales affects its price. The estimation results for the three measures are reported in columns (1) to (3). Consistent with our expectations, the period closer to the sale has a stronger effect that diminishes over time. For all three measures, we find significant and higher positive coefficients for Airbnb listings in the months of S to S-3 (phase 1) compared to Airbnb listings in the months of S-4 to S-7 (phase 2) and S-8 to S-11 (phase 3). For instance, the estimation results for our third measure in column (3) show that 100 additional Airbnb reservations within 300 m of a house are associated with 1.01%, 0.67%, and 0.55% higher home sale prices in phases 1, 2, and 3, respectively. The results in Table 5 show the effect of distance. We expected that closer Airbnb properties would have a greater impact on house sale prices. All results in Table 5 are significant, and the direction of the coefficients is consistent for each measure of Airbnb activity. An interesting result here is that all three measures show a negative effect of Airbnb listings in ring 3 (Airbnb listings at 500 to 1000 m distance from houses) on house prices. Although these negative results are statistically different from zero, their magnitudes are very small, indicating that Airbnb listings far from a property (i.e., higher than 500 m) do not have a strong impact on house prices. In particular, Table  5, Column (1) shows that for phase1, the month of the house sale and the 3-month period prior to it, one additional Airbnb listing in ring 1 has a 0.36% impact on house prices. The average magnitude of this impact Table 5   Model 3: Distance Effect, all three measures Standard errors are reported in parentheses and were obtained using 500 bootstrap replicates. ⁎ p < 0.1; ⁎⁎ p < 0.05; ⁎⁎⁎ p < 0.01 The estimations correspond to Eq. (3), with the natural log of the inflation-adjusted house sale price as the dependent variable The coefficients of the control variables introduced in Table 1, are omitted from this table. A complete version of the table is available upon request Measure (1) Active Measure (2) Positive Occupancy Measure (3) Sum of reservations Airbnb_Phase1 _Ring1 0.0036*** (0.0002) 0.0038*** (0.0003) 0.0183*** (0.0022) Airbnb_Phase1 _Ring2 0.0008*** (0.0003) 0.0008** (0.0003) 0.0112*** (0.0027) Airbnb_Phase1 _Ring3 − 0.0007*** (0.0001) − 0.0008*** (0.0001) − 0.0069*** (0.0009) Zip code dummies Yes Yes Yes School district dummies Yes Yes Yes Year of sale dummies Yes Yes Yes Month of sale dummies Yes Yes Yes N 102,142 102,142 102,142 adj. R-sq 0.789 0.789 0.789 ===== PAGE 18 (score 78, numeric-paren 52, table-words 7, var-words 6) ===== N. Safari et al. 6  Page 18 of 33 Table 6   Unconditional Quantile Regression Result, all three measures The results for each of our three measures are presented in each row, with each measure individually analyzed through a distinct quantile regression. In the table, each col- umn represents the results corresponding to specific quantiles The estimations correspond to Eq. (4), with the natural log of the inflation-adjusted house sale price as the dependent variable Each regression includes Zip code, School district, Year of sale, and Month of sale dummies as fixed effects. The total number of observations for each regression is 102,142 Standard errors are in parentheses. Standard errors for the quantile regression were obtained using 500 bootstrap replicates. ⁎ p < 0.1. ⁎⁎ p < 0.05. ⁎⁎⁎ p < 0.01 The coefficients for the control variables introduced in Table 1, are not included in this table. The complete version of this table is available upon request 0.10 Quantile 0.20 Quantile 0.30 Quantile 0.40 Quantile 0.50 Quantile 0.60 Quantile 0.70 Quantile 0.80 Quantile 0.90 Quantile Measure (1): Airbnb _Phase1_Ring1 0.0030*** (0.0007) 0.0018*** (0.0005) 0.0017*** (0.0004) 0.0023*** (0.0003) 0.0030*** (0.0003) 0.0041*** (0.0003) 0.0059*** (0.0003) 0.0053*** (0.0004) 0.0040*** (0.0005) Measure (2): Airbnb _Phase1_Ring1 0.0031*** (0.0009) 0.0017*** (0.0005) 0.0015*** (0.0005) 0.0022*** (0.0004) 0.0029*** (0.0003) 0.0041*** (0.0004) 0.0060*** (0.0004) 0.0058*** (0.0004) 0.0045*** (0.0005) Measure (3): Airbnb _Phase1_Ring1 0.0123* (0.0063) 0.0020 (0.0039) 0.0002 (0.0034) 0.0077*** (0.0028) 0.0139*** (0.0025) 0.0228*** (0.0026) 0.0356*** (0.0030) 0.0361*** (0.0032) 0.0312*** (0.0040) ===== PAGE 19 (score 12, numeric-paren 0, table-words 3, var-words 5) ===== The impact of short‑term rental activity on house prices:… Page 19 of 33  6 is approximately 0.29%, while an additional Airbnb listing in ring 2 increases the sale price of nearby houses by 0.8%. Finally, one additional Airbnb listing in ring 3 decreases the neighborhood house sale price by 0.07%. The results of the first two measures are similar and consistent. The results in Column (3) reveal that the third measure shows a positive effect of 1.83% on house price for each 100 more reser- vations in ring 1, with an average magnitude of approximately 0.146%. 100 more reservations in ring 2 are associated with a 1.12% positive effect, and the average magnitude of this effect is 0.114%. Finally, 100 more reservations in ring 3 are asso- ciated with a 0.69% negative effect on house sales. Ordinary least-squares regressions estimate the conditional mean impact of the STR on housing prices. To estimate the impact of Airbnb on higher versus lower house prices, we proceed to the unconditional quantile regression (UQR) model shown in Eq. (4). Because the Airbnb impact is concentrated in ring 1 and phase 1, we apply the UQR model to our baseline Eq. (1). The results are presented in Table 6. For the first two measures, the coefficients for all quantiles (first and second rows) are significant at the 1% level. According to our third measure (third row), most quantiles have significant coefficients. The unconditional quantile regression results are consistent for all the three measures. The results reveal that the positive impact of Airbnb density on house prices is not consistent across the entire house price distribution. This relationship is notably stronger for the 70th and 80th quantiles, with the 90th quantile also showing a relatively strong impact from short-term rentals. The 10th quantile exhibits a higher impact relative to its neighboring quantiles. However, the literature suggests that elevated tails are often linked to data sparsity, potentially introducing noise in the tails of quantile regression estimation (specifically the 10th and 90th quantiles), especially for heavy-tailed distributions (Wang, Deyuan, and Xuming, 2012). Conversely, it appears that house prices in the middle range receive an impact close to the average, very similar to the OLS regression result. Lower quantiles, namely the 20th, 30th, and 40th, show the lowest impact. This result is better illus- trated in Fig. 5, where each panel plots the explanatory variable’s UQR coefficient estimates and their associated 95% confidence intervals. Considering that the quantile result for the tails could be noisy, the interpretation might be biased for the 10th and 90th quantiles. However, there is a clear upward trend in the impact of short-term rental activity from the 20th to the 80th quantile. For instance, in Table 6, the quantile regression result for the second measure of Airbnb activity (second row) shows that both the 70th and 80th quantiles have coef- ficients of 0.0060 and 0.0058, respectively, noticeably higher than those of the other quantiles. Furthermore, houses in the middle price range receive an impact close to the average result coming from OLS regression. For example, the coefficients of the second measure of Airbnb activity for the 50th and 60th quantiles are 0.0029 and 0.0041, respectively, compared to 0.0035 of OLS regression for our second measure. Additionally, lower-priced houses are less impacted by short-term rental activity. For instance, the coefficients of the second measure of Airbnb activity for the 20th, 30th, and 40th quantiles are 0.0017, 0.0015, and 0.0022, respectively. ===== PAGE 25 (score 28, numeric-paren 12, table-words 5, var-words 4) ===== The impact of short‑term rental activity on house prices:… Page 25 of 33  6 Table 7   Model 1: OLS regression estimates, all three measures (Clustered Standard Errors vs, Bootstrap Standard Errors) Standard errors for regression results in columns (1), (3), and (5) were computed using 500 bootstrap replicates and are presented in parentheses. For columns (2), (4), and (6), robust-clustered standard errors clustered by zip code and month are reported in parentheses ⁎ p < 0.1. ⁎⁎ p < 0.05. ⁎⁎⁎ p < 0.01 The estimations correspond to Eq. (1), with the natural log of the inflation-adjusted house sale price as the dependent variable The coefficients of the control variables introduced in Table 1, are not included in this table Measure (1)-Active Measure (2)-Positive Occupancy Measure (3)-Sum of reservations (1) (2) (3) (4) (5) (6) Airbnb_Phase1 _Ring1 0.0034*** (0.0002) 0.0034*** (0.0003) 0.0035*** (0.0002) 0.0035*** (0.0003) 0.0169*** (0.0018) 0.0169*** (0.0022) Zip code dummies Yes Yes Yes Yes Yes Yes School district dummies Yes Yes Yes Yes Yes Yes Year of sale dum- mies Yes Yes Yes Yes Yes Yes Month of sale dummies Yes Yes Yes Yes Yes Yes N 102,142 102,142 102,142 102,142 102,142 102,142 adj. R-sq 0.789 0.789 0.789 0.789 0.789 0.789 ===== PAGE 26 (score 38, numeric-paren 18, table-words 7, var-words 4) ===== N. Safari et al. 6  Page 26 of 33 Table 8   Model 1: OLS regression estimates, all three measures (Including Year × Month and Zip Code × Calendar Month Interactions) The table displays results for each measure under three different scenarios: without including Year * Month or Zip code * calendar Month (Columns 1, 4, and 7), including Year * Month (Columns 2, 5, and 8), and including Zip code * calendar Month (Columns 3, 6, and 9) Standard errors are in parentheses. Standard errors for the regression were obtained using 500 bootstrap replicates. ⁎ p < 0.1. ⁎⁎ p < 0.05. ⁎⁎⁎ p < 0.01 The estimations correspond to Eq. (1), with the natural log of the inflation-adjusted house sale price as the dependent variable The coefficients of the control variables introduced in Table 1, are not included in this table. The complete version of this table is available upon request Measure (1)—Active Measure (2)—Positive Occupancy Measure (3)—Sum of reservations (1) (2) (3) (4) (5) (6) (7) (8) (9) Airbnb_Phase1 _Ring1 0.0034*** (0.0002) 0.0034*** (0.0002) 0.0036*** (0.0002) 0.0035*** (0.0002) 0.0035*** (0.0002) 0.0038*** (0.0003) 0.0169*** (0.0018) 0.0170*** (0.0018) 0.0206*** (0.0019) Zip code dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes School district dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes Year of sale dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes Month of sale dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes Year * Month No Yes No No Yes No No Yes No Zip code * calendar Month No No Yes No No Yes No No Yes N 102,142 102,142 102,142 102,142 102,142 102,142 102,142 102,142 102,142 adj. R-sq 0.789 0.789 0.794 0.789 0.789 0.794 0.789 0.789 0.794 ===== PAGE 27 (score 70, numeric-paren 46, table-words 7, var-words 4) ===== The impact of short‑term rental activity on house prices:… Page 27 of 33  6 Table 9   Model 2: Time Effect, all three measures (Including Year × Month and Zip Code × Calendar Month Interactions) The table displays results for each measure under three different scenarios: without including Year * Month or Zip code * calendar Month (Columns 1, 4, and 7), including Year * Month (Columns 2, 5, and 8), and including Zip code * calendar Month (Columns 3, 6, and 9) Standard errors are in parentheses. Standard errors for the regression were obtained using 500 bootstrap replicates. ⁎ p < 0.1. ⁎⁎ p < 0.05. ⁎⁎⁎ p < 0.01 The estimations correspond to Eq. (2), with the natural log of the inflation-adjusted house sale price as the dependent variable The coefficients of the control variables introduced in Table 1, are not included in this table. The complete version of this table is available upon request Measure (1)—Active Measure (2)—Positive Occupancy Measure (3)—Sum of reservations (1) (2) (3) (4) (5) (6) (7) (8) (9) Airbnb_Phase1 _Ring1 0.0033*** (0.0006) 0.0033*** (0.0006) 0.0033*** (0.0006) 0.0019*** (0.0005) 0.0019*** (0.0005) 0.0018*** (0.0005) 0.0101*** (0.0030) 0.0103*** (0.0030) 0.0108*** (0.0033) Airbnb_Phase2_Ring1 − 0.0006 (0.0008) − 0.0007 (0.0008) − 0.0007 (0.0008) 0.0011* (0.0006) 0.0010* (0.0006) 0.0013** (0.0006) 0.0067* (0.0040) 0.0061 (0.0040) 0.0079* (0.0045) Airbnb_Phase3_Ring1 0.0008 (0.0007) 0.0009 (0.0007) 0.0012* (0.0007) 0.0012** (0.0005) 0.0012** (0.0005) 0.0016*** (0.0006) 0.0055 (0.0038) 0.0059 (0.0038) 0.0094** (0.0043) Zip code dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes School district dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes Year of sale dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes Month of sale dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes Year * Month No Yes No No Yes No No Yes No Zip code * calendar Month No No Yes No No Yes No No Yes N 102,142 102,142 102,142 102,142 102,142 102,142 102,142 102,142 102,142 adj. R-sq 0.789 0.789 0.794 0.789 0.789 0.794 0.789 0.789 0.794 ===== PAGE 28 (score 78, numeric-paren 54, table-words 7, var-words 4) ===== N. Safari et al. 6  Page 28 of 33 Table 10   Model 3: Distance Effect, all three measures (Including Year × Month and Zip Code × Calendar Month Interactions) The table displays results for each measure under three different scenarios: without including Year * Month or Zip code * Month (Columns 1, 4, and 7), including Year * Month (Columns 2, 5, and 8), and including Zip code * Month (Columns 3, 6, and 9) Standard errors are in parentheses. Standard errors for the regression were obtained using 500 bootstrap replicates. ⁎ p < 0.1. ⁎⁎ p < 0.05. ⁎⁎⁎ p < 0.01 The estimations correspond to Eq. (3), with the natural log of the inflation-adjusted house sale price as the dependent variable The coefficients of the control variables introduced in Table 1, are not included in this table. The complete version of this table is available upon request Measure (1)—Active Measure (2)—Positive Occupancy Measure (3)—Sum of reservations (1) (2) (3) (4) (5) (6) (7) (8) (9) Airbnb_Phase1 _Ring1 0.0036*** (0.0002) 0.0036*** (0.0002) 0.0036*** (0.0002) 0.0038*** (0.0003) 0.0038*** (0.0003) 0.0038*** (0.0003) 0.0183*** (0.0022) 0.0184*** (0.0022) 0.0193*** (0.0022) Airbnb_Phase1_Ring2 0.0008*** (0.0003) 0.0008*** (0.0003) 0.0008*** (0.0003) 0.0008** (0.0003) 0.0008** (0.0003) 0.0007** (0.0003) 0.0112*** (0.0027) 0.0112*** (0.0027) 0.0115*** (0.0027) Airbnb_Phase1_Ring3 − 0.0007*** (0.0001) − 0.0007*** (0.0001) − 0.0005*** (0.0001) − 0.0008*** (0.0001) − 0.0007*** (0.0001) − 0.0005*** (0.0001) − 0.0069*** (0.0009) − 0.0068*** (0.0009) − 0.0051*** (0.0010) Zip code dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes School district dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes Year of sale dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes Month of sale dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes Year * Month No Yes No No Yes No No Yes No Zip code * calendar Month No No Yes No No Yes No No Yes N 102,142 102,142 102,142 102,142 102,142 102,142 102,142 102,142 102,142 adj. R-sq 0.789 0.789 0.794 0.789 0.789 0. 0.794 0.789 0.789 0.794 ===== PAGE 29 (score 192, numeric-paren 157, table-words 6, var-words 6) ===== The impact of short‑term rental activity on house prices:… Page 29 of 33  6 Table 11   Unconditional Quantile Regression Result, all three measures (Including Year × Month and Zip Code × Calendar Month Interactions) The results for each of our three measures are presented in three rows under three different scenarios: without including Year * Month or Zip code * Month (Rows 1, 4, and 7), including Year * Month (Rows 2, 5, and 8), and including Zip code * Month (Rows 3, 6, and 9). Each measure is analyzed separately using quantile regression. Columns represent the results corresponding to specific quantiles Each regression includes Zip code, School district, Year of sale, and Month of sale dummies as fixed effects. The total number of observations for each regression is 102,142 Standard errors are in parentheses. Standard errors for the quantile regression were obtained using 500 bootstrap replicates. ⁎ p < 0.1. ⁎⁎ p < 0.05. ⁎⁎⁎ p < 0.01 The estimations correspond to Eq. (4), with the natural log of the inflation-adjusted house sale price as the dependent variable The coefficients for the control variables introduced in Table 1, are not included in this table. The complete version of this table is available upon request Additional Fixed Effect 0.10 Quantile 0.20 Quantile 0.30 Quantile 0.40 Quantile 0.50 Quantile 0.60 Quantile 0.70 Quantile 0.80 Quantile 0.90 Quantile Measure (1) (1) – 0.0030*** (0.0007) 0.0018*** (0.0005) 0.0017*** (0.0004) 0.0023*** (0.0003) 0.0030*** (0.0003) 0.0041*** (0.0003) 0.0059*** (0.0003) 0.0053*** (0.0004) 0.0040*** (0.0005) (2) Year * Month 0.0030*** (0.0007) 0.0018*** (0.0005) 0.0017*** (0.0004) 0.0022*** (0.0003) 0.0029*** (0.0003) 0.0041*** (0.0003) 0.0059*** (0.0003) 0.0052*** (0.0004) 0.0040*** (0.0005) (3) Zip code * Calendar month 0.0032*** (0.0008) 0.0022*** (0.0005) 0.0020*** (0.0004) 0.0025*** (0.0003) 0.0032*** (0.0003) 0.0046*** (0.0003) 0.0063*** (0.0003) 0.0056*** (0.0004) 0.0041*** (0.0005) Measure (2) (4) – 0.0031*** (0.0009) 0.0017*** (0.0005) 0.0015*** (0.0005) 0.0022*** (0.0004) 0.0029*** (0.0003) 0.0041*** (0.0004) 0.0060*** (0.0004) 0.0058*** (0.0004) 0.0045*** (0.0006) (5) Year * Month 0.0031*** (0.0009) 0.0017*** (0.0005) 0.0015*** (0.0005) 0.0021*** (0.0004) 0.0029*** (0.0003) 0.0041*** (0.0004) 0.0060*** (0.0004) 0.0058*** (0.0004) 0.0045*** (0.0005) (6) Zip code * Calendar month 0.0034*** (0.0009) 0.0023** (0.0006) 0.0019*** (0.0005) 0.0025*** (0.0004) 0.0034*** (0.0004) 0.0048*** (0.0004) 0.0065*** (0.0004) 0.0062*** (0.0005) 0.0045*** (0.0006) Measure (3) (7) – 0.0123* (0.0063) 0.0020 (0.0039) 0.0002 (0.0034) 0.0077*** (0.0028) 0.0139*** (0.0025) 0.0228*** (0.0026) 0.0356*** (0.0030) 0.0361*** (0.0032) 0.0312*** (0.0040) (8) Year * Month 0.0126** (0.0063) 0.0024 (0.0039) 0.0002 (0.0034) 0.0075*** (0.0028) 0.0138*** (0.0025) 0.0227*** (0.0026) 0.0357*** (0.0030) 0.0361*** (0.0032) 0.0312*** (0.0040) (9) Zip code * Calendar month 0.0165** (0.0067) 0.0076* (0.0041) 0.0042 (0.0036) 0.0110*** (0.0030) 0.0187*** (0.0026) 0.0290*** (0.0028) 0.0415*** (0.0032) 0.0403*** (0.0034) 0.0313*** (0.0043) ===== PAGE 30 (score 135, numeric-paren 102, table-words 8, var-words 6) ===== N. Safari et al. 6  Page 30 of 33 Table 12   Unconditional Quantile Regression Result, all three measures (Census Block Dummies vs. Zip Code Dummies) The table displays results for each measure under two different scenarios for comparison: our main result from Table 5 (Rows 1, 3, and 5) and new results with Census block dummies replacing Zip code dummies (Rows 2, 4, and 6). Each measure is analyzed separately using quantile regression. Columns represent the results correspond- ing to specific quantiles Each regression includes Zip code, School district, Year of sale, and Month of sale dummies as fixed effects. The total number of observations for each regression is 102,142 Standard errors are in parentheses. Standard errors for the quantile regression were obtained using 500 bootstrap replicates. ⁎ p < 0.1. ⁎⁎ p < 0.05. ⁎⁎⁎ p < 0.01 The estimations correspond to Eq. (4), with the natural log of the inflation-adjusted house sale price as the dependent variable The coefficients for the control variables introduced in Table 1, are not included in this table. The complete version of this table is available upon request Additional Fixed Effect 0.10 Quantile 0.20 Quantile 0.30 Quantile 0.40 Quantile 0.50 Quantile 0.60 Quantile 0.70 Quantile 0.80 Quantile 0.90 Quantile Measure (1): Airbnb _Phase1_ Ring1 (1) – 0.0030*** (0.0007) 0.0018*** (0.0005) 0.0017*** (0.0004) 0.0023*** (0.0003) 0.0030*** (0.0003) 0.0041*** (0.0003) 0.0059*** (0.0003) 0.0053*** (0.0004) 0.0040*** (0.0005) (2) Census block dummies 0.0027*** (0.0008) 0.0014*** (0.0005) 0.0015*** (0.0004) 0.0019*** (0.0003) 0.0025*** (0.0003) 0.0036*** (0.0003) 0.0054*** (0.0004) 0.0048*** (0.0004) 0.0037*** (0.0005) Measure (2): Airbnb _Phase1_ Ring1 (3) – 0.0031*** (0.0009) 0.0017*** (0.0005) 0.0015*** (0.0005) 0.0022*** (0.0004) 0.0029*** (0.0003) 0.0041*** (0.0004) 0.0060*** (0.0004) 0.0058*** (0.0004) 0.0045*** (0.0006) (4) Census block dummies 0.0027*** (0.0009) 0.0013** (0.0006) 0.0013*** (0.0005) 0.0016*** (0.0004) 0.0023*** (0.0003) 0.0034*** (0.0004) 0.0053*** (0.0004) 0.0052*** (0.0005) 0.0042*** (0.0006) Measure (3): Airbnb _Phase1_ Ring1 (5) – 0.0123* (0.0063) 0.0020 (0.0039) 0.0002 (0.0034) 0.0077*** (0.0028) 0.0139*** (0.0025) 0.0228*** (0.0026) 0.0356*** (0.0030) 0.0361*** (0.0032) 0.0312*** (0.0040) (6) Census block dummies 0.0091 (0.0065) 0.0001 (0.0040) − 0.0007 (0.0035) 0.0040 (0.0028) 0.0086*** (0.0025) 0.0161*** (0.0026) 0.0280*** (0.0030) 0.0300*** (0.0033) 0.0299*** (0.0041) ===== PAGE 31 (score 123, numeric-paren 91, table-words 7, var-words 7) ===== The impact of short‑term rental activity on house prices:… Page 31 of 33  6 Table 13   Unconditional Quantile Regression Result, all three measures (Private Room vs. Entire Place) The table displays results for each measure in two rows under two different scenarios: including only “Entire place” listings (rows 1, 3, and 5) and including only “Private room” listings (Rows 2, 4, and 6). Each measure is analyzed separately using quantile regression. Columns represent the results corresponding to specific quantiles Each regression includes Zip code, School district, Year of sale, and Month of sale dummies as fixed effects. The total number of observations for each regression is 102,142 Standard errors are in parentheses. Standard errors for the quantile regression were obtained using 500 bootstrap replicates. ⁎ p < 0.1. ⁎⁎ p < 0.05. ⁎⁎⁎ p < 0.01 The estimations correspond to Eq. (4), with the natural log of the inflation-adjusted house sale price as the dependent variable The coefficients for the control variables introduced in Table 1, are not included in this table. The complete version of this table is available upon request Style: 0.10 Quantile 0.20 Quantile 0.30 Quantile 0.40 Quantile 0.50 Quantile 0.60 Quantile 0.70 Quantile 0.80 Quantile 0.90 Quantile Measure (1): (1) Entire place 0.0028*** (0.0008) 0.0020*** (0.0005) 0.0021*** (0.0004) 0.0028*** (0.0004) 0.0035*** (0.0003) 0.0049*** (0.0003) 0.0067*** (0.0004) 0.0060*** (0.0004) 0.0051*** (0.0005) (2) Private room 0.0152*** (0.0034) 0.0044** (0.0021) − 0.0008 (0.0019) − 0.0020 (0.0015) 0.0004 (0.0014) 0.0010 (0.0014) 0.0061*** (0.0016) 0.0036** (0.0018) − 0.0069*** (0.0022) Measure (2): (3) Entire place 0.0030*** (0.0009) 0.0019*** (0.0006) 0.0020*** (0.0005) 0.0029*** (0.0004) 0.0036*** (0.0004) 0.0050*** (0.0004) 0.0069*** (0.0004) 0.0067*** (0.0005) 0.0058*** (0.0006) (4) Private room 0.0130*** (0.0038) 0.0038 (0.0024) − 0.0026 (0.0021) − 0.0034** (0.0017) − 0.0007 (0.0015) 0.0001 (0.0016) 0.0053*** (0.0018) 0.0040** (0.0020) − 0.0062** (0.0024) Measure (3): (5) Entire place 0.0128* (0.0068) 0.0034 (0.0042) 0.0046 (0.0037) 0.0120*** (0.0030) 0.0177*** (0.0027) 0.0276*** (0.0028) 0.0415*** (0.0032) 0.0406*** (0.0035) 0.0374*** (0.0044) (6) Private room 0.0298 (0.0265) − 0.0100 (0.0163) − 0.0539 (0.0143) − 0.0342*** (0.0117) − 0.0123 (0.0104) − 0.0018 (0.0108) 0.0161 (0.0125) 0.0326** (0.0135) − 0.0106 (0.0169) ===== PAGE 32 (score 9, numeric-paren 0, table-words 0, var-words 9) ===== N. Safari et al. 6  Page 32 of 33 Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permis- sion directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/. References Abadie A, Athey S, Imbens GW, Wooldridge JM (2023) When should you adjust standard errors for clus- tering? Q J Econ 138(1):1–35 Alyakoob M, Rahman M (2018) Airbnb: Is it a curse or a blessing for restaurant employment Airbnb Newsroom (2022) Airbnb first quarter 2022 financial results. https://​news.​airbnb.​com/​airbnb-​first-​ quart​er-​2022-​finan​cial-​resul​ts/ Ayouba K, Breuillé ML, Grivault C, Le Gallo J (2020) Does Airbnb disrupt the private rental market? An empirical analysis for French cities. Int Reg Sci Rev 43(1–2):76–104 Barrios JM, Hochberg YV, Yi H (2023) The cost of convenience: Ridehailing and traffic fatalities. J Oper Manage 69(5):823–855. https://​doi.​org/​10.​1002/​joom.​1221 Barron K, Kung E, Proserpio D (2021) The effect of home-sharing on house prices and rents: evidence from Airbnb. Mark Sci 40(1):23–47 Benitez-Aurioles B, Tussyadiah I (2020) What Airbnb does to the housing market. Ann Tour Res 90(1):103108 Cameron AC, Miller DL (2015) A practitioner’s guide to cluster-robust inference. J Hum Resour 50(2):317–372 Daniels K, Grinstein-Weiss M (2019) The impact of the gig-economy on financial hardship among low- income families. Available at SSRN 3293988 Elíasson L, Ragnarsson ÖP (2018) Short-term renting of residential apartments. Effects of Airbnb in the Icelandic housing market (No. wp76). Department of Economics, Central bank of Iceland Erhardt GD, Roy S, Cooper D, Sana B, Chen M, Castiglione J (2019) Do transportation network compa- nies decrease or increase congestion? Sci Adv 5(5):eaau2670 Farronato C, Fradkin A (2022) The welfare effects of peer entry: the case of airbnb and the accommoda- tion industry. Am Econ Rev 112(6):1782–1817 Filippas A, Horton JJ, Zeckhauser RJ (2020) Owning, using, and renting: some simple economics of the “sharing economy.” Manage Sci 66(9):4152–4172 Firpo S, Fortin NM, Lemieux T (2009) Unconditional quantile regressions. Econometrica 77(3):953–973 Fortin N, Lemieux T, Firpo S (2011). Decomposition methods in economics. In: Handbook of labor eco- nomics (Vol 4, pp 1-102). Elsevier Franco SF, Santos CD (2021) The impact of Airbnb on residential property values and rents: evidence from Portugal. Reg Sci Urban Econ 88:103667 Gant AC (2016) Holiday rentals: the new gentrification battlefront. Sociol Res Online 21(3):112–120 Garcia-López MÀ, Jofre-Monseny J, Martínez-Mazza R, Segú M (2020) Do short-term rental platforms affect housing markets? Evidence from Airbnb in Barcelona. J Urban Econ 119:103278 Jiao J, Hansen K, Azimian A (2021) Land value impacts of Airbnb listings on single-family homes in Austin, Texas, USA. International Journal of Housing Markets and Analysis Ke L, O’Brien DT, Heydari B (2021) Airbnb and neighborhood crime: The incursion of tourists or the erosion of local social dynamics? PLoS ONE 16(7):e0253315 Kim JH, Leung TC, Wagman L (2017) Can restricting property use be value enhancing? Evidence from short-term rental regulation. J Law Econ 60(2):309–334 Koenker R, Bassett Jr G (1978) Regression quantiles. Econometrica: Journal of the Econometric Society pp 33–50 Koster HR, van Ommeren J, Volkhausen N (2021) Short-term rentals and the housing market: quasi- experimental evidence from Airbnb in Los Angeles. J Urban Econ 124:103356