Impact of COVID-19 Pandemic on Airbnb Listings in New York City: Challenges and Opportunities for Urban Housing Sustainability

Abstract

Short-term rental (STR) platforms like Airbnb have significantly impacted urban housing sustainability, particularly in cities like New York City. The COVID-19 pandemic disrupted the STR market, raising questions about its resilience and effects on sustainable urban housing. This study addresses the following research questions: (1) How did unit and neighborhood characteristics influence the survival of Airbnb listings during the pandemic? (2) What changes occurred in the factors determining the emergence of new listings during the pandemic? Using data from Inside Airbnb, we applied Cox proportional hazard models and negative binomial regression to analyze changes before and after the pandemic. We found that during the pandemic, price discounts became crucial for listing survival, while traditional quality indicators like superhost status and high ratings lost significance. The importance of subway accessibility decreased, reflecting shifts in traveler preferences. Additionally, new listings were less likely to emerge in high-density Airbnb areas and more likely in neighborhoods with higher crime rates. These findings highlight the need for sustainable regulatory approaches that balance the benefits of STR platforms with protecting housing affordability and community well-being. Our study provides insights for policymakers aiming to promote sustainable urban housing during global crises.

1. Introduction

The rapid expansion of short-term rental (STR) platforms, such as Airbnb, has significantly transformed urban housing markets worldwide over the past decade. These platforms have introduced a new dynamic to the accommodation industry, offering a decentralized marketplace where hosts and guests can connect directly. Unlike traditional hotels, which are typically concentrated in specific districts, STRs are dispersed across urban areas, providing more diverse options in terms of location and pricing [1,2]. By 2019, STRs had grown at a much faster rate than the hotel industry in the United States, accounting for 9% of all accommodation in the 10 largest cities [3]. However, this growth has not been without controversy. The proliferation of STRs has been linked to various negative externalities, including noise pollution, parking shortages, and increased pressure on public services [4,5,6]. More critically, the rise of STRs has exacerbated housing affordability issues in many urban areas, as the conversion of long-term rental units to STRs reduces the supply of available housing and drives up rent and property values [7,8,9,10]. STRs increase rent and house prices through two primary mechanisms: the potential for STR revenue, which raises house prices as it is factored into property values [11,12], and the reduction in long-term rental housing supply, which drives up rent due to decreased availability [7,13,14]. In response to these challenges, local governments have enacted regulations, such as owner-occupancy requirements, licensing, and multiple taxes, to mitigate the negative impacts on housing affordability. Recent studies indicate that stricter regulations have effectively reduced the number of STRs, house prices, and rent [15].

The COVID-19 pandemic, which began in early 2020, introduced a significant and unprecedented disruption to the global economy, including the STR market. With travel restrictions, social distancing measures, and a sharp decline in tourism, the demand for short-term rentals plummeted. As a result, many STRs were either temporarily or permanently removed from the market [16]. This decline in STR activity raises important questions about the resilience of the STR market and the broader implications for urban housing markets.

This study seeks to investigate the resilience of STRs in New York City during the COVID-19 pandemic by addressing the following research questions:

• What unit and neighborhood characteristics were associated with the survival of STRs during the pandemic? Existing studies suggest that STRs are more likely to survive in neighborhoods with high tourist appeal and robust infrastructure [17,18]. This study will explore whether these factors continued to play a role during the pandemic or if new determinants emerged.
• How do these associations differ before and after the pandemic? The pandemic may have altered the traditional drivers of STR success, such as proximity to transportation and tourist attractions [19]. This research will compare the pre-pandemic and post-pandemic periods to identify shifts in these relationships.
• What factors determine the location of new STRs during the pandemic? New STRs may have emerged in response to shifting demand patterns, such as a preference for less densely populated areas or locations with lower COVID-19 infection rates. This study will examine these trends in detail.
• How have these determinants changed compared to the pre-pandemic period? By comparing the factors influencing STR locations before and after the pandemic, this research aims to uncover changes in urban land use patterns related to STRs.

We use a comprehensive dataset from Inside Airbnb, which provides monthly measurements of Airbnb listings across New York City, further supplemented by neighborhood-level data from the American Community Survey (ACS), crime statistics from the New York Police Department (NYPD), and employment data from the LEHD Origin-Destination Employment Statistics (LODES) dataset. By integrating these extensive data sources, the study examines how external shocks like the COVID-19 pandemic impact the STR market and, in turn, urban housing dynamics. The integration of big data allows for a detailed examination of the spatial and temporal patterns in STR resilience. This research contributes to the broader discourse on urban land use by highlighting how STRs, as a form of peer-to-peer accommodation, interact with traditional housing markets and how these interactions are reshaped by global crises. The findings are expected to offer valuable insights for urban planners and policymakers, particularly in the context of balancing the needs for tourism development and housing affordability in large metropolitan areas like New York City.

2. Literature Review

The rise of short-term rental (STR) platforms such as Airbnb has revolutionized the accommodation sector, introducing significant shifts in urban housing markets globally. As a form of disruptive innovation, Airbnb and similar platforms have expanded the traditional boundaries of tourism accommodation, allowing everyday homeowners to become hosts and offering travelers alternatives to conventional hotels. This transformation has sparked considerable academic interest, leading to a robust body of literature examining the multifaceted impacts of STRs on urban environments.

2.1. Disruptive Innovation and the Sharing Economy

The concept of Airbnb as a disruptive innovation is well established in the literature. Guttentag [20] highlighted how Airbnb represents a shift from formal to informal accommodation sectors, challenging established hotel industries and altering consumer expectations. Subsequent studies have explored the trust mechanisms that underpin the sharing economy, with Wang, Asaad, and Filieri [21] investigating how hosts develop trust in Airbnb as a platform, which in turn influences their continued participation. Similarly, Yang et al. [22] examined the factors that build consumer trust and attachment to Airbnb, underscoring the importance of platform trust in sustaining the STR market.

Airbnb’s impact on traditional hotel markets has been a focal point of research, with studies like that of Dogru, Mody, and Suess [23] quantifying the platform’s disruptive influence on key hotel markets. Their findings demonstrate that Airbnb’s growth correlates with a significant decrease in hotel revenues, particularly in markets where Airbnb has a strong presence. This disruption is not limited to the hotel industry but extends to broader urban economies, influencing housing markets, local businesses, and employment patterns [14].

2.2. STRs and Urban Housing Markets

The expansion of STR platforms has had profound effects on urban housing markets, particularly concerning housing affordability and availability. Research has consistently shown that the growth of STRs contributes to rising rent and property values. For instance, Horn and Merante [24] provided empirical evidence from Boston, demonstrating that Airbnb’s presence drives up local rent by reducing the supply of long-term rental housing. Similarly, Garcia-López et al. [8] found that in Barcelona, the proliferation of STRs leads to higher rent and house prices, exacerbating the affordability crisis in the city.

The mechanisms through which STRs impact housing markets are multifaceted. One primary factor is the “rent gap” theory, where the potential income from STRs increases property values, as homeowners and investors capitalize on the opportunity to generate higher returns from short-term rentals [11,12]. Moreover, the conversion of long-term rental units into short-term rentals reduces the overall supply of housing, leading to increased competition and higher rent [25].

Studies have also explored the broader socio-economic implications of STR growth, including its role in driving gentrification. Wachsmuth and Weisler [14] argued that STRs contribute to the gentrification of urban neighborhoods by attracting higher-income tourists and displacing long-term residents, particularly in areas with significant rent gaps. This process not only alters the demographic composition of neighborhoods but also intensifies social inequalities within cities.

2.3. Regulatory Responses

In response to the challenges posed by STRs, cities worldwide have implemented various regulatory measures aimed at mitigating their negative impacts on housing markets and communities. These regulations often include licensing requirements, occupancy limits, zoning restrictions, taxation, and enforcement mechanisms designed to control the proliferation of STRs and protect housing affordability. Ferreri and Sanyal [26] examined how cities like London have adopted “regulated deregulation” strategies, where platforms like Airbnb are subject to specific rules that aim to balance economic benefits with housing needs. These strategies involve allowing STRs to operate but within a framework of regulations that address issues such as maximum allowable rental days, mandatory registration, and compliance with safety standards. The effectiveness of these regulatory approaches, however, has been mixed. For example, Jin, Wagman, and Zhong [27], in their study of short-term rental regulations in Chicago, found that implementing a middle-ground ordinance—which included limitations, registration requirements, and a new tax—led to a 16.4% decline in active Airbnb listings over two years. Notably, this effect was only significant after the city began receiving detailed data feeds from STR platforms, highlighting the crucial role of enforcement mechanisms. They further demonstrated that such regulations produced nuanced effects on different stakeholders, including localized reductions in burglaries near buildings that prohibited STR listings and varied impacts on Airbnb revenues across different areas. Their findings suggest that the success of STR regulations depends not only on the design of the rules but also on effective enforcement. Similarly, Gurran, Zhang, and Shrestha [28] analyzed the impact of Airbnb regulations in coastal Australia, finding that well-enforced regulations can mitigate some of the adverse effects on local housing markets. They noted that regulations need to be tailored to the specific contexts of different regions to be effective. These studies collectively underscore the complexity of regulating STRs and highlight that a one-size-fits-all approach is often ineffective. Effective regulation requires a deep understanding of local market conditions and the specific challenges faced by different communities. Robust enforcement mechanisms are also essential to ensure compliance and achieve the intended outcomes of the regulations.

2.4. STR Resilience During the COVID-19 Pandemic

The COVID-19 pandemic has presented an unprecedented shock to the STR market, dramatically altering patterns of demand and supply. With travel restrictions and public health measures in place, many STRs experienced sharp declines in bookings, leading to a significant reduction in active listings [16]. Recent studies have begun to examine the resilience of STRs during this period, focusing on how different unit and neighborhood characteristics influenced their survival and adaptation [19]. Choi and Won [29] further explored this topic by comparing the resilience of STRs in rural versus non-rural markets in Virginia. Their findings highlight how the STR market’s response to the pandemic can vary significantly based on the geographic and socio-economic context. While rural areas demonstrated a different pattern of resilience compared to urban environments, this study underscores the importance of context when analyzing STR resilience. However, the environmental and market conditions in New York City—a densely populated, urban setting—differ markedly from those in Virginia, necessitating a more localized examination of STR resilience in large metropolitan areas.

Research on the effects of the pandemic on STRs continues to grow, primarily focusing on the heterogeneous impact across different markets and types of hosts. Zhang et al. [30] identified three types of STR hosts and demonstrated that each type employed different strategies to respond to decreased demand. Meanwhile, Huang and Chen [31] highlighted the various stressors that hosts faced during the pandemic and the different coping strategies they adopted. Additionally, Jang and Kim [32] investigated the spatially heterogeneous effects of COVID-19 on Airbnb performance in Florida, revealing that leisure clusters negatively impacted Airbnb revenue, while rural counties with a strong focus on leisure and hospitality were less affected than urban areas. These studies collectively underscore the complexity of STR resilience and the varying factors that influence market responses during a global crisis.

While the existing literature provides a robust foundation for understanding the impact of STRs on urban housing markets and the effectiveness of various regulatory responses, significant gaps remain. Notably, the resilience of STRs during unprecedented global shocks, such as the COVID-19 pandemic, has not been fully explored. Although recent studies have begun to address how STRs have adapted to these disruptions, there is still much to learn about the specific unit and neighborhood characteristics that contribute to STR survival in such volatile environments. Furthermore, the implications of these shifts for urban housing markets, particularly in terms of affordability and availability, are not yet fully understood. The pandemic has also highlighted the need for a re-evaluation of traditional success factors for STRs, as changes in traveler behavior and safety concerns may have permanently altered the landscape of the sharing economy. Given these gaps, this study seeks to provide a more nuanced understanding of how external shocks like the COVID-19 pandemic impact the STR market.

3. Data and Methodology

This study examines the relationship between COVID-19 and the STR market by comparing the survival and generation mechanisms of New York City Airbnb listings before and after the COVID-19 pandemic. The primary data source is Inside Airbnb (IA), which provides a large open dataset regarding Airbnb listings for major cities. Research using web-scraped data obtained from third-party sources such as Inside Airbnb or AirDNA has become increasingly popular due to the granularity and timeliness of the data provided [14]. IA founder Murray Cox releases web-scraped data monthly, including detailed information on the location, pricing, and characteristics of Airbnb listings. The data provided by IA include unit-specific attributes (e.g., room type, price, rating). This study further enriches the Airbnb data by incorporating neighborhood characteristics from various sources, including the American Community Survey (ACS), the New York Police Department (NYPD), and the LEHD Origin-Destination Employment Statistics (LODES) dataset. These additional datasets provide a comprehensive view of the socio-economic and infrastructural context within which Airbnb listings operate.

3.1. Variable List

The variables used in this study are divided into two categories: unit characteristics and neighborhood characteristics.

Table 1. Variables used in the survival analysis.

Unit of Analysis			Airbnb Unit
Dependent Variable			Survival Time
Independent Variables	Category		Variables
	Unit	Room	Room type (shared or private room = 0, entire house = 1) # of beds (0/1/2/3+) Price per night Whether the price was dropped (No = 0, Yes = 1) Minimum stay: 30 nights or more (No = 0, Yes = 1) Whether host is a superhost (No = 0, Yes = 1) Rating (≤90 = 0, 91–95 = 1, 96–100 = 2)
		Location	Distance to nearest airport Distance to nearest hospital
	Neighborhood (census tract)		# of subway stations within 1000 feet Ratio of Airbnb listings to housing units Ratio of the number of crimes to population Population density Ln(median income) Homeownership rate Rental vacancy rate Unemployment rate % non-Hispanic Black Location quotient (LQ) of leisure industries (NAICS 71: Arts, Entertainment, and Recreation) Location quotient (LQ) of hospitality industries (NACIS 72: Accommodation and Food Services)

presents a detailed list of the variables employed in the analysis.

Unit Characteristics:

• Room Type: Whether the listing is an entire house/apartment or a shared/private room.
• Number of Beds: The total number of beds available in the listing.
• Price per Night: The nightly rate charged for the listing.
• Price Discount: A binary variable indicating whether the host reduced the price during the analysis period.
• Minimum Stay: Whether the listing requires a minimum stay of 30 days or more.
• Superhost Status: Whether the host is designated as a “superhost” by Airbnb, a status that indicates a high level of service and guest satisfaction.
• Rating: The average rating of the listing, categorized into three groups (0–90, 91–95, 96–100).
• Distance to Nearest Airport: The Euclidean distance from the listing to the nearest airport, calculated using ArcGIS Pro 3.2.0.
• Distance to Nearest Hospital: The Euclidean distance from the listing to the nearest hospital, also calculated using ArcGIS Pro.
• Neighborhood Characteristics:
• Subway Accessibility: The number of subway stations within 1000 feet of each census tract, reflecting the ease of access to public transportation.
• Ratio of Listings to Housing Units: The density of Airbnb listings in a neighborhood, measured as the ratio of listings to total housing units.
• Crime Rate: The number of crimes per 1000 persons within a 1000-foot radius of each census tract, sourced from the NYPD.
• Population Density: The number of residents per 1000 square feet in each census tract, providing a measure of neighborhood density.
• Median Income: The median household income in each census tract, obtained from the ACS.
• Homeownership Rate: The percentage of housing units occupied by owners, rather than renters.
• Rental Vacancy Rate: The percentage of rental units that are vacant, indicating the availability of rental housing in each neighborhood.
• Unemployment Rate: The percentage of the labor force that is unemployed in each neighborhood.
• Percentage of Non-Hispanic Black Residents: The proportion of the population in each census tract that identifies as non-Hispanic Black, included to explore potential racial and socio-economic dynamics in the STR market.
• Location Quotient (LQ) of Leisure Industries: A measure of the relative concentration of leisure-related industries (NAICS 71: Arts, Entertainment, and Recreation) in a neighborhood.
• Location Quotient (LQ) of Hospitality Industries: A measure of the relative concentration of hospitality-related industries (NAICS 72: Accommodation and Food Services) in a neighborhood.

Samples are restricted to those located in tracts with at least 200 housing units. To ensure data accuracy, outlier prices were removed using the same method as Jiao and Bai [17]. As of March 2019, there were 46,146 Airbnb listings in New York City.

3.2. Empirical Strategy

This study employs two primary analytical approaches: the Cox proportional hazard model for survival analysis and the negative binomial regression model for generation analysis. Each approach is designed to address different aspects of the STR market during the COVID-19 pandemic.

3.2.1. Survival Analysis

The survival analysis aims to explore the factors that influence whether an Airbnb listing remains active over time. This analysis is conducted using the Cox proportional hazard model, which is a type of survival analysis commonly used in both medical and social sciences to examine the time until the occurrence of a specific event (e.g., the deactivation of an Airbnb listing) [33]. The dependent variable in this model is the time until a listing is deactivated, while the independent variables include both unit-specific characteristics (e.g., room type, price, rating) and neighborhood-level factors (e.g., crime rates, proximity to hospitals). The Cox model is particularly useful in this context because it does not assume a specific baseline hazard function, allowing for greater flexibility in analyzing how different factors influence the likelihood of survival over time [34].

The analysis is structured to compare two distinct periods: pre-COVID-19 (March 2019 to February 2020) and post-COVID-19 (March 2020 to February 2021). This comparison allows for the examination of how the pandemic has altered the survival mechanisms of STRs in New York City. A joint significance test is performed as a post-estimation procedure to determine whether the survival mechanisms have changed significantly between the two periods.

3.2.2. Generation Analysis

The generation (or creation) analysis investigates the factors that contribute to the emergence of new Airbnb listings during the pandemic. This analysis is conducted using a negative binomial regression model, which is appropriate for count data, particularly when the variance exceeds the mean [35]. The dependent variable in this model is the number of new Airbnb listings in a given neighborhood, and the independent variables are similar to those used in the survival analysis, focusing on neighborhood characteristics that might attract new STRs.

The negative binomial regression model is used here because it accounts for overdispersion in the data, where the variance is greater than the mean—a common occurrence in count data, especially in urban studies involving the number of occurrences of specific events [36]. This analysis is also conducted for both the pre-COVID-19 and post-COVID-19 periods to identify shifts in the factors driving the creation of new Airbnb units. By comparing these two periods, the study aims to uncover how the pandemic has influenced the patterns of STR emergence in different neighborhoods of New York City (

Table 2. Analytic Strategy.

Methodology	Unit of Analysis	Pre-COVID-19	Post-COVID-19	Difference
		March 2019–February 2020	March 2020–February 2021
Survival Analysis (Cox Proportional Hazard Model)	Airbnb Units	Unit Character (×COVID-19 period) Neighborhood Character (×COVID-19 period) PUMA Fixed Effect (×COVID-19 period)		Joint Significance Test (Post-estimation)
Negative Binomial Regression	Neighborhood (Census Tract)	Neighborhood Character (×COVID-19 period) PUMA Fixed Effect (×COVID-19 period)

).

3.3. Study Area: New York City

New York City serves as the case study for this research due to its significant STR market and the diverse socio-economic landscape across its neighborhoods. The city has been at the forefront of debates surrounding STRs, particularly regarding their impact on housing affordability and neighborhood dynamics. With over 1% of its housing units listed on Airbnb, NYC provides a robust context for examining how STR markets respond to external shocks like the COVID-19 pandemic.

In New York City, the proliferation of STRs has been a focal point of concern for both residents and policymakers. The city, known for its housing affordability crisis, has witnessed significant pressure from STRs, particularly in neighborhoods with already high housing costs. Research has consistently shown that an increase in STRs contributes to rising rent and house prices, further exacerbating affordability issues [4,5,6,7,8,9,10].

STRs have been identified as drivers of gentrification in NYC, particularly in neighborhoods with a high density of Black residents, by reducing the stock of affordable rental units [14]. In response, the NYC government has implemented several regulations aimed at mitigating these negative impacts. These regulations include requirements for hosts to obtain special licenses, limits on how much hosts can charge in rent-stabilized properties, and multiple taxes applicable to transient occupancy. The COVID-19 pandemic, however, has significantly impacted the STR market in NYC. Monthly changes in Airbnb listings show a 25% decrease in active units during the pandemic compared to pre-pandemic levels. The introduction of new regulations, such as the data-sharing requirement in January 2021, which mandates automatic reporting of STR activity to the city government, has further contributed to this decline. In particular, the number of new Airbnb listings has decreased by 41% during the post-COVID-19 period compared to the pre-COVID-19 period, indicating a shift in the creation dynamics of STRs. The survival rate of existing listings has also been affected, with different patterns emerging across various neighborhoods in the city.

This study seeks to explore these shifts in greater detail, focusing on the survival and generation mechanisms of Airbnb units in NYC during the COVID-19 pandemic. By doing so, it aims to provide insights into how external shocks, like the pandemic, impact the STR market and the broader urban housing landscape in one of the world’s most dynamic cities.

4. Results

4.1. Descriptive Statistics

Table 3. Failure rate of each analysis period.

	Number of Units in Analysis	Failure Event in 1 Year	Failure Rate
Pre-COVID-19	46,146	18,286	39.6%
Post-COVID-19	27,860	11,424	41.0%

provides an overview of the survival records for each analysis period, highlighting the differences in failure events between the pre-COVID-19 and post-COVID-19 periods. A failure event is defined as the removal of a listing from the platform. During the pre-COVID-19 period, there were 18,286 failure events out of 46,146 units, resulting in a failure rate of 39.6%. In the post-COVID-19 period, 27,860 units were analyzed, with 11,424 failure events, leading to a slightly higher failure rate of 41.0%. This indicates that the pandemic had a noticeable impact on the survival of Airbnb listings in New York City.

Table 4. Descriptive statistics.

	Pre-COVID-19		Post-COVID-19
	Mar 2019–Feb 2020		Mar 2020–Feb 2021
	Mean	SD	Mean	SD
Unit
Room type (entire house = 1)	0.501	0.500	0.510	0.500	*
Number of beds	1.118	0.654	1.093	0.618	***
Price per night (USD)	120	68	118	65	***
Price discount (yes = 1)	0.075	0.263	0.376	0.484	***
Minimum stay (30 days or more = 1)	0.097	0.295	0.076	0.265	***
Superhost (yes = 1)	0.170	0.376	0.184	0.387	***
Rating
0–90 (%)	40		36
91–95 (%)	18		18
96–100 (%)	42		45
Distance to nearest airport (miles)	6.901	2.224	6.905	2.199
Distance to hospital (miles)	0.795	0.522	0.800	0.523
Neighborhood
Subway accessibility	2.106	2.086	2.042	1.965	***
Ratio of listings to housing units	0.022	0.015	0.021	0.015	***
Crimes per 1000 persons	0.159	0.141	0.155	0.133	***
Pop density (per 1000 square feet)	2.585	1.317	2.579	1.311
Median income (USD)	84,741	39,836	84,626	39,392
Homeownership rate (%)	23.65	16.45	23.77	16.44
Rental vacancy rate (%)	4.15	3.61	4.06	3.54	***
Unemployment rate (%)	5.40	3.25	5.40	3.22
LQ of leisure	1.35	5.63	1.33	3.75
LQ of hospitality	2.70	3.70	2.76	3.74
Number of listings	46,146		27,860

Notes: *: p < 0.05; ***: p < 0.001.

summarizes the descriptive statistics of the variables used in the survival analysis, providing a detailed comparison between the pre-COVID-19 and post-COVID-19 periods. Significant changes are evident in several key variables, reflecting the impact of the pandemic on Airbnb listings in New York City. The proportion of units offering discounted prices rose dramatically from 7.5% in the pre-COVID-19 period to 37.6% in the post-COVID-19 period, which is statistically significant. This sharp increase suggests that hosts widely adopted pricing discounts as a strategy to attract bookings amid the reduced demand during the pandemic.

The average number of beds per unit decreased slightly, indicating a shift towards smaller units, which might have been perceived as more flexible or safer options during the pandemic. Similarly, the average price per night dropped slightly. The proportion of units with a minimum stay requirement of 30 days or more decreased from 9.7% to 7.6%, indicating that hosts were more inclined to accept shorter-term bookings during the pandemic, possibly due to the decrease in demand for long-term stays. Interestingly, the proportion of superhost-designated listings increased from 17.0% to 18.4%. This suggests that experienced hosts, who typically maintain higher ratings and better reviews, were better positioned to retain their status or were more likely to meet the criteria for superhost designation during the pandemic. Regarding neighborhood characteristics, there was a slight but statistically significant decrease in subway accessibility. This could reflect a geographical shift in listing locations, potentially moving away from densely populated or highly accessible areas during the pandemic. Meanwhile, crime rates per 1000 persons also experienced a minor decrease, which might reflect broader social changes during the pandemic, such as decreased urban activity or stricter law enforcement. Lastly, the rental vacancy rate decreased from 4.15% to 4.06%, pointing to a tightening rental market during the pandemic, which could be attributed to increased demand for long-term rentals as short-term travel declined. In contrast, variables such as the distance to the nearest airport and distance to the nearest hospital remained stable, with no significant differences observed between the two periods. This stability suggests that these factors were less affected by the pandemic and maintained their importance as constant characteristics of the listings.

4.2. Results of Survival Analysis

The results, summarized in

Table 5. Results of survival analysis.

	Unit Characteristics						Neighborhood Characteristics						Full Model
	(1)			(2)			(1)			(2)			(1)			(2)
	Pre-COVID-19			Post-COVID-19			Pre-COVID-19			Post-COVID-19			Pre-COVID-19			Post-COVID-19
	Mar 2019–Feb 2020			Mar 2020–Feb 2021			Mar 2019–Feb 2020			Mar 2020–Feb 2021			Mar 2019–Feb 2020			Mar 2020–Feb 2021
	Coef.		HR	Coef.		HR	Coef.		HR	Coef.		HR	Coef.		HR	Coef.		HR
Unit
Room type (entire house = 1)	−0.2767	***	0.7583	−0.2525	***	0.7769							−0.2722	***	0.7617	−0.2511	***	0.7780
Number of beds (Ref. 0–1)
2	0.1008	***	1.1061	0.2305	***	1.2592							0.0970	***	1.1019	0.2302	***	1.2588
3+	0.2709	***	1.3111	0.2772	***	1.3195							0.2632	***	1.3010	0.2744	***	1.3158
Price per night (USD)	0.0014	***	1.0014	0.0027	***	1.0027							0.0015	***	1.0015	0.0028	***	1.0028
Price discount (yes = 1)	−0.5563	***	0.5733	−0.9104	***	0.4023							−0.5597	***	0.5714	−0.9130	***	0.4013
Minimum stay (30 days or more = 1)	0.4078	***	1.5035	0.6324	***	1.8821							0.4114	***	1.5090	0.6372	***	1.8912
Superhost (yes = 1)	−0.0880	***	0.9157	0.6241	***	1.8666							−0.0884	***	0.9154	0.6259	***	1.8698
Rating (Ref. 0–90)
91–95	−0.1148	***	0.8915	0.4137	***	1.5125							−0.1126	***	0.8935	0.4186	***	1.5198
96–100	−0.2501	***	0.7787	0.0919	**	1.0962							−0.2468	***	0.7813	0.0943	**	1.0989
Distance to nearest airport (miles)	0.0242	+	1.0245	−0.0067		0.9933							0.0283	+	1.0287	−0.0202		0.9800
Distance to hospital (miles)	0.0091		1.0092	0.0584	+	1.0602							0.0113		1.0114	0.0662	*	1.0685
Neighborhood
# of subway station within 1000 feet							0.0128	*	1.0129	0.0268	***	1.0271	0.0074		1.0074	0.0336	***	1.0341
Ratio of Airbnb listings to housing units							−0.0208		0.9794	−0.1948		0.8230	0.0149		1.0150	0.2090		1.2324
Crimes per 1000 persons							0.0036		1.0036	0.0120		1.0120	0.0119		1.0119	−0.1505		0.8603
Pop density (per 1000 square feet)							−0.0056		0.9944	0.0182		1.0184	0.0000		1.0000	0.0170		1.0171
Median income (in USD 1000s)							−0.0013	***	0.9987	0.0004		1.0004	−0.0014	***	0.9986	−0.0007		0.9993
Homeownership rate (%)							−0.0017	*	0.9983	0.0019	+	1.0019	−0.0012	+	0.9988	0.0007		1.0007
Rental vacancy rate (%)							0.0060	**	1.0060	0.0057		1.0057	0.0049	*	1.0049	0.0023		1.0023
Unemployment rate (%)							−0.0008		0.9992	−0.0003		0.9997	−0.0023		0.9977	0.0000		1.0000
LQ of leisure							−0.0025	+	0.9975	−0.0020		0.9980	−0.0027	*	0.9973	−0.0023		0.9977
LQ of hospitality							−0.0104	***	0.9896	−0.0023		0.9977	−0.0110	***	0.9891	−0.0060		0.9940
N	74,006						74,006						74,006
LR chi2	4353.010						790.560						4461.770
Wald χ2	Prob > chi2 = 0.0000						Prob > chi2 = 0.0000						Prob > chi2 = 0.0000
Region controls	Yes						Yes						Yes
Event-related sharing data controls	Yes						Yes						Yes

Notes: +: p < 0.1; *: p < 0.05; **: p < 0.01; ***: p < 0.001.

and

Table 6. Joint significance test of survival analysis.

	(1)			(2)			(3)
	Pre-COVID-19			Post-COVID-19			Difference
	Mar 2019–Feb 2020			Mar 2020–Feb 2021
	Coef.		HR	Coef.		HR	Coef.		HR
Unit
Room type (entire house = 1)	−0.2722	***	0.7617	−0.2511	***	0.7780	0.0211		1.0213
Number of beds (Ref. 0–1)
2	0.0970	***	1.1019	0.2302	***	1.2588	0.1331	***	1.1424
3+	0.2632	***	1.3010	0.2744	***	1.3158	0.0113		1.0114
Price per night (USD)	0.0015	***	1.0015	0.0028	***	1.0028	0.0013	***	1.0013
Price discount (yes = 1)	−0.5597	***	0.5714	−0.9130	***	0.4013	−0.3533	***	0.7024
Minimum stay (30 days or more = 1)	0.4114	***	1.5090	0.6372	***	1.8912	0.2258	***	1.2533
Superhost (yes = 1)	−0.0884	***	0.9154	0.6259	***	1.8698	0.7142	***	2.0426
Rating (Ref. 0–90)
91–95	−0.1126	***	0.8935	0.4186	***	1.5198	0.5312	***	1.7010
96–100	−0.2468	***	0.7813	0.0943	**	1.0989	0.3411	***	1.4065
Distance to nearest airport (miles)	0.0283	+	1.0287	−0.0202		0.9800	−0.0485	*	0.9526
Distance to hospital (miles)	0.0113		1.0114	0.0662	*	1.0685	0.0549		1.0564
Neighborhood
# of subway station within 1000 feet	0.0074		1.0074	0.0336	***	1.0341	0.0262	**	1.0265
Ratio of Airbnb listings to housing units	0.0149		1.0150	0.2090		1.2324	0.1941		1.2142
Crimes per 1000 persons	0.0119		1.0119	−0.1505		0.8603	−0.1623		0.8502
Pop density (per 1000 square feet)	0.0000		1.0000	0.0170		1.0171	0.0170		1.0172
Median income (in USD 1000s)	−0.0014	***	0.9986	−0.0007		0.9993	0.0007		1.0007
Homeownership rate (%)	−0.0012	+	0.9988	0.0007		1.0007	0.0020		1.0020
Rental vacancy rate (%)	0.0049	*	1.0049	0.0023		1.0023	−0.0026		0.9974
Unemployment rate (%)	−0.0023		0.9977	0.0000		1.0000	0.0023		1.0023
LQ of leisure	−0.0027	*	0.9973	−0.0023		0.9977	0.0004		1.0004
LQ of hospitality	−0.0110	***	0.9891	−0.0060		0.9940	0.0050		1.0050
N	74,006
LR chi2	4461.770
Wald χ2	Prob > chi2 = 0.0000
Region controls	Yes
Event-related sharing data controls	Yes

Notes: +: p < 0.1; *: p < 0.05; **: p < 0.01; ***: p < 0.001.

, indicate that several unit characteristics—such as price per night, price discounts, minimum stay requirements, superhost status, and ratings—significantly affect the survival of Airbnb listings, with some of these effects intensifying after the pandemic began. For instance, listings with higher prices per night were consistently less likely to survive, and this effect became more pronounced post-COVID-19. This suggests that during the pandemic, there was heightened price sensitivity among consumers, leading to a competitive advantage for lower-priced listings. The coefficient for price per night increased from 0.0015 pre-COVID-19 to 0.0028 post-COVID-19, showing a stronger negative impact on survival in the latter period. The impact of price discounts on survival was particularly notable. In the pre-COVID-19 period, listings that offered discounts had a 43% lower chance of being removed from the platform, and this protective effect increased to 60% post-COVID-19. This finding aligns with the argument by Leoni [37] that flexible pricing strategies are crucial for enhancing the survival chances of STRs during economic downturns. The stronger effect of price discounts in the post-COVID-19 period underscores the importance of adaptive pricing strategies in response to decreased demand. Minimum stay requirements also played a significant role in survival. Listings requiring stays of 30 days or more were less likely to survive in both periods, but the effect was stronger post-COVID-19. This suggests that during the pandemic, short-term stays became even more preferred, possibly due to the uncertainties surrounding travel restrictions and consumer preferences for flexibility. The coefficient for minimum stay increased from 0.4114 pre-COVID-19 to 0.6372 post-COVID-19, indicating a growing disadvantage for long-term stay units. Interestingly, superhost status and ratings, which are typically associated with higher service levels and better guest experiences, exhibited a reversal in their effects. Before the pandemic, superhosts and higher-rated listings had better survival rates. However, in the post-COVID-19 period, these factors were associated with a higher likelihood of failure. This reversal may reflect a shift in market dynamics where the benefits of being a superhost or having a high rating are outweighed by other pressures, such as reduced demand or the financial strain of maintaining high service levels during the pandemic.

Among the neighborhood characteristics, subway accessibility showed a significant shift in its impact on survival. In the pre-COVID-19 period, proximity to subway stations had little effect on survival. However, in the post-COVID-19 period, it became negatively associated with survival, indicating that listings in areas with better subway access were more likely to fail. This could be due to a reduced reliance on public transportation during the pandemic, as travelers and residents alike avoided crowded spaces. The location quotient (LQ) of leisure and hospitality industries did not show significant changes in their effects on survival between the two periods. This suggests that while these factors were expected to play a role due to the pandemic’s impact on tourism-related industries, their influence on Airbnb survival remained relatively stable.

In summary, the survival analysis reveals that the COVID-19 pandemic significantly altered the dynamics of the STR market in New York City. Hosts with adaptive pricing strategies, such as offering discounts, were better able to navigate the challenges of the pandemic. Additionally, changes in neighborhood characteristics, such as the decreased importance of subway accessibility, highlight shifts in consumer behavior during the pandemic.

4.3. Results of Negative Binomial Regression

The negative binomial regression analysis focuses on the factors influencing the generation of new Airbnb listings. This analysis reveals that several neighborhood characteristics significantly impact where new Airbnb units are likely to emerge, with notable changes observed between the pre- and post-COVID-19 periods.

During the pre-COVID-19 period, several neighborhood characteristics played a significant role in determining where new Airbnb units were likely to emerge. One of the most notable factors was the density of existing Airbnb listings in a neighborhood. Areas with a higher concentration of existing Airbnb units were more likely to see new listings, indicating a clustering effect that is typical in STR markets. This clustering suggests that certain neighborhoods had become established hubs for Airbnb activity, likely due to their popularity among tourists and the success of existing hosts in those areas. Another important factor was the crime rate within a neighborhood. In the pre-COVID-19 period, neighborhoods with higher crime rates were less likely to see the generation of new Airbnb units. This is consistent with the general understanding that safety is a key consideration for both hosts and guests. High crime rates can deter both hosts from listing their properties and guests from choosing those areas for their stay.

The post-COVID-19 period, however, saw notable shifts in these dynamics. The effect of Airbnb density on the generation of new listings diminished significantly. While clustering still occurred, the strength of this effect was weaker, suggesting that hosts may have been less inclined to enter already saturated markets during the pandemic. This could be due to increased competition among hosts, a reduction in overall demand, or a strategic shift towards less crowded areas where new listings might stand out more or cater to new traveler preferences.

Interestingly, the negative effect of crime rates on the generation of new Airbnb units also weakened post-COVID-19. While areas with higher crime rates still saw fewer new Airbnb units being generated, the negative impact was less pronounced than before. This shift might reflect changing perceptions of safety, possibly influenced by broader societal changes during the pandemic, or it could indicate a different profile of Airbnb guests and hosts who are less sensitive to crime statistics in their decision-making. It is also possible that the distribution of crime changed due to the pandemic, affecting how crime data from 2019 relate to new Airbnb listings.

The distance to key landmarks such as the Empire State Building continued to influence the likelihood of new listings, with properties closer to such landmarks being more likely to be listed on Airbnb. However, this relationship also showed a slight shift post-COVID-19, with the effect becoming more pronounced. This could suggest that during the pandemic, proximity to well-known, iconic locations became even more critical for attracting guests, possibly due to a reduced number of travelers who prioritized staying near major attractions during shorter or more focused visits.

Overall, the findings from the negative binomial regression highlight significant shifts in the generation of new Airbnb listings in response to the COVID-19 pandemic (

Table 7. Results of negative binomial regression.

	(1)			(2)			(3)
	Pre-COVID-19			Post-COVID-19			Difference
	Mar 2019–Feb 2020			Mar 2020–Feb 2021
	Coef.		IRR	Coef.		IRR	Coef.		IRR
Neighborhood
Distance to Empire State Building (miles)	−0.1788	***	0.8363	−0.2156	***	0.8061	−0.0368		0.9639
Distance to nearest airport (miles)	−0.1258	***	0.8818	−0.1618	***	0.8506	−0.0359		0.9647
# of subway station within 1000 feet	0.0868	***	1.0907	0.0787	***	1.0818	−0.0081		0.9919
Ratio of Airbnb listings to housing units (%)	0.3876	***	1.4735	0.3301	***	1.3911	−0.0575	*	0.9441
Crimes per 1000 persons	−0.8125	***	0.4437	−0.5509	***	0.5764	0.2616	*	1.2989
Pop density (per 1000 square feet)	0.1169	***	1.1240	0.1247	***	1.1328	0.0078		1.0078
Median income (in USD 1000s)	−0.0008		0.9992	−0.0006		0.9994	0.0002		1.0002
Homeownership rate (%)	−0.0067	***	0.9934	−0.0087	***	0.9913	−0.0021		0.9979
Rental vacancy rate (%)	0.0129	**	1.0129	0.0107	*	1.0108	−0.0021		0.9979
Unemployment rate (%)	−0.0210	*	0.9792	−0.0208	*	0.9794	0.0002		1.0002
LQ of leisure	0.0061	+	1.0062	0.0033		1.0034	−0.0028		0.9972
LQ of hospitality	0.0030		1.0030	0.0056		1.0056	0.0026		1.0026
Constant	3.3952 ***			3.2372 ***
/ln alpha	−0.8583
alpha	0.4238
Region controls	Yes

Notes: +: p < 0.1; *: p < 0.05; **: p < 0.01; ***: p < 0.001.

). The reduction in the clustering effect and the diminished influence of crime rates suggest that the market dynamics for STRs have become more complex and less predictable in the post-COVID-19 environment. These shifts reflect broader changes in traveler behavior, safety perceptions, and market conditions, which urban planners and policymakers need to consider when addressing the future of the STR market in cities like New York.

5. Discussion and Conclusions

5.1. Discussion

This study provides a comprehensive analysis of the shifts in survival and generation mechanisms of Airbnb listings in New York City during the COVID-19 pandemic. The findings reveal significant changes in how various unit and neighborhood characteristics influenced the likelihood of listings remaining active and the emergence of new listings before and after the pandemic.

One of the most notable findings is the increased importance of price discounts as a survival strategy during the pandemic. Before COVID-19, offering a price discount reduced the hazard of a listing being deactivated by 43%. This protective effect became even more pronounced post-COVID-19, with a 60% reduction in the hazard. This finding underscores the heightened price sensitivity among consumers during the pandemic and aligns with the results of Leoni [37], who emphasized the importance of dynamic pricing strategies in enhancing STR resilience during economic downturns.

Conversely, traditional quality indicators such as superhost status and high ratings, which were positively associated with listing survival before the pandemic, exhibited a reversal in their effects post-COVID-19. Superhost status, which previously reduced the hazard of deactivation, became associated with an increased hazard in the post-pandemic period. Similarly, high ratings no longer provided the same protective effect. This shift suggests that during the pandemic, travelers prioritized different factors when choosing accommodation. Concerns about health and safety, changes in travel purposes (e.g., from leisure to necessity travel), or a focus on cost-saving may have diminished the importance of these traditional quality markers.

The decreased significance of subway accessibility in the survival of listings post-COVID-19 further underscores the changes in traveler behavior and preferences. Prior to the pandemic, proximity to subway stations was an asset, facilitating easy access to attractions and amenities. However, with health concerns about crowded public transportation and a rise in remote work reducing the need for commuting, listings near subway stations were more likely to be deactivated during the pandemic. This shift may reflect a preference for accommodation in less densely populated areas or those perceived as safer from a health perspective.

In examining the generation of new Airbnb listings, the study found that the clustering effect observed pre-pandemic weakened during the pandemic. Before COVID-19, new listings were more likely to emerge in neighborhoods with a high density of existing Airbnb units, possibly due to established demand and host networks. Post-COVID-19, this effect diminished, suggesting that hosts were less inclined to enter already saturated markets during a period of reduced demand. Instead, there was an unexpected increase in new listings in neighborhoods with higher crime rates. This could be attributed to hosts in these areas seeking alternative income sources due to economic pressures or changes in the types of travelers using Airbnb during the pandemic, such as frontline workers or individuals needing temporary housing.

These findings have significant implications for urban housing sustainability and policy. The decrease in active Airbnb listings, particularly in high-density areas with good subway access, may have temporarily increased the availability of long-term rental housing in these neighborhoods. This could alleviate some pressure on housing affordability in the short term. However, as tourism demand recovers, there is potential for a resurgence of STR activity, which could once again strain housing markets and exacerbate affordability issues. Sustainable regulatory approaches are needed to balance the benefits of STR platforms with the imperative to protect housing affordability, community well-being, and sustainable urban development.

The shifts observed in traveler preferences and host strategies during the pandemic highlight the need for adaptive and nuanced regulatory approaches. Policies that encourage the conversion of STRs back to long-term rentals in high-demand neighborhoods could be beneficial for housing affordability and community stability. Additionally, enforcement mechanisms, such as data-sharing requirements from platforms, are essential to monitor STR activity effectively. The changes in the importance of factors like price discounts and quality indicators suggest that regulations should consider not only the quantity of STRs but also the dynamics of pricing and service quality in the market.

This study also underscores the complexity of the STR market and the challenges in regulating it effectively. The divergent impacts of unit and neighborhood characteristics on listing survival and generation before and after the pandemic indicate that static regulatory measures may not be sufficient. Policymakers need to understand the evolving dynamics of the STR market, especially in response to external shocks like the pandemic, to design effective interventions.

5.2. Conclusions

The COVID-19 pandemic has significantly altered the landscape of the short-term rental market in New York City, with important implications for urban housing sustainability. This study reveals that during the pandemic, price discounts became a crucial strategy for Airbnb hosts to keep their listings active, while traditional quality markers like superhost status and high ratings lost their pre-pandemic significance. The decreased importance of subway accessibility reflects a shift in traveler preferences, possibly due to health concerns and changes in mobility patterns.

The generation mechanisms of new Airbnb listings also shifted, with new listings less likely to emerge in high-density Airbnb areas and more likely in neighborhoods with higher crime rates. These changes suggest that hosts adapted to the new market conditions, possibly seeking opportunities in less saturated or lower-cost areas.

These findings have important implications for urban housing policy and the regulation of STRs. Policymakers must consider the dynamic nature of the STR market and the factors that influence host and traveler behavior, especially during crises. Adaptive regulatory approaches that can respond to changing market conditions are necessary to balance the benefits of STR platforms with the need to protect housing affordability, community well-being, and sustainable urban development.

As the tourism industry recovers from the pandemic, there is potential for the STR market to rebound strongly. Policymakers should proactively address the challenges posed by STRs to prevent negative impacts on housing markets. This includes implementing effective enforcement mechanisms, encouraging the return of STRs to the long-term rental market where appropriate, and considering the diverse factors that influence STR dynamics.

Understanding these shifts is crucial for developing sustainable urban housing policies that can withstand future shocks. By addressing the limitations identified and pursuing the suggested avenues for future research, scholars and policymakers can better navigate the complex dynamics of the STR market and promote sustainability in urban housing.

5.3. Limitations and Future Research

While this study provides valuable insights into the impacts of the COVID-19 pandemic on the Airbnb market in New York City, several limitations should be acknowledged. First, the analysis is specific to New York City, which has unique market characteristics and regulatory environments. Therefore, the findings may not be generalizable to other cities with different contexts. Future research could extend this analysis to other urban areas to examine whether similar patterns emerge.

Second, the study relies on data from Inside Airbnb, which, while comprehensive, may not capture all aspects of the STR market, such as listings on other platforms or unregistered rentals. Additionally, the data do not provide information on hosts’ motivations or travelers’ preferences beyond what can be inferred from the available variables. Qualitative studies or surveys could complement this research by providing deeper insights into the behaviors and perceptions of hosts and guests during the pandemic.

Third, the study covers a specific time frame during the pandemic. As the situation evolves, long-term effects may differ from the patterns observed in this period. Continuous monitoring of the STR market is necessary to understand the lasting impacts of the pandemic on urban housing sustainability.

Building upon the limitations identified, future research should explore several key areas to deepen the understanding of the impacts of the COVID-19 pandemic on the short-term rental (STR) market and urban housing sustainability.

Firstly, extending the analysis to other cities and regions is essential to determine the generalizability of the findings. Comparative studies across different urban contexts, both domestically and internationally, could reveal whether similar patterns in the survival and generation mechanisms of Airbnb listings are observed elsewhere. Such research would help identify the influence of local market characteristics, cultural factors, and regulatory environments on STR dynamics during global crises.

Secondly, incorporating data from multiple STR platforms, such as Vrbo, Booking.com, and HomeAway, would offer a more comprehensive view of the STR market. This approach would address the limitation of relying solely on Inside Airbnb data and capture a broader spectrum of listings, including those on less prominent platforms or operating unofficially.

Thirdly, future research should employ qualitative methodologies to delve deeper into the motivations and behaviors of hosts and guests during the pandemic. Interviews, focus groups, and surveys could uncover the underlying reasons behind hosts’ decisions to adjust pricing, deactivate listings, or enter new markets. Similarly, understanding travelers’ preferences, concerns, and decision-making processes would shed light on the demand side of the STR market.

Fourthly, longitudinal studies are needed to assess the long-term effects of the pandemic on the STR market and urban housing sustainability. As the situation evolves and the tourism industry recovers, it is crucial to monitor whether the observed shifts are temporary or indicative of lasting changes in host and traveler behaviors. Such studies could track the resurgence of STR activity, changes in regulatory responses, and the impact on housing affordability over an extended period.

Fifthly, exploring the effectiveness of different regulatory approaches in managing the STR market during and after the pandemic is vital. Comparative analyses of cities that adopted varying regulatory strategies—ranging from strict enforcement to more flexible, adaptive policies—could provide valuable insights into best practices. Understanding how regulations impact the balance between promoting tourism, protecting housing affordability, and ensuring community well-being would inform policymakers in crafting effective interventions.

By pursuing these avenues, future research can contribute to a more comprehensive understanding of the STR market’s dynamics, inform more effective and adaptive policy measures, and ultimately promote sustainability in urban housing.