Regulating short-term rental platforms: the effects of local regulatory responses on Airbnb’s operations in Europe

Abstract

Many digital platforms offer services that affect real-world socio-economic processes. One example is the impact of short-term rental platforms such as Airbnb or Wimdu on cities and neighborhoods. Because these platforms often operate in a regulatory void characterized by absent, unclear, or poorly enforced laws and regulations, local governments in affected cities have begun experimenting with a variety of instruments to regulate the operations of short-term rental platforms. In this paper, we report how such locally implemented regulatory responses have affected Airbnb’s operations across 13 European cities over the period from 2015 to 2019. Using a difference-in-difference specification with synthetic controls, we assess the impact of different regulatory responses by disaggregating them into motivations, actions, targets, and outcomes. We find that the effectiveness of regulatory responses differs by type of regulation (restricting or clarifying), type of host (professional or private), as well as the enforcement (with or without the cooperation of the platform operator). Through this work, we add to the ongoing debate on the regulation of digital platforms by presenting both empirical evidence as well as an analytical framework.

1. Introduction

Many digital platforms provide services that affect real-world socio-economic processes. Prominent examples include short-term rental (STR) platforms such as Airbnb and Wimdu, which facilitate interactions between hosts offering accommodation and their guests from countries around the world. As is typical for digital platforms, network effects increase the value of these platforms as a function of the number of their users (Parker et al., 2016). However, because STR platforms provide services that affect real-world socio-economic processes, the associated increasing returns to scale can also have unintended consequences. Consider how a platform such as Airbnb becomes increasingly useful the more hosts and guests interact on the platform. While this increases choice for guests, this dynamic has also been associated with negative externalities such as rising housing prices and rents (Garcia-López et al., 2020) as well as the displacement of residents from flashpoint areas in cities (e.g., Wachsmuth and Weisler, 2018). Due to these unintended consequences, STR platforms are increasingly coming under regulatory scrutiny (Edelman and Geradin, 2016; Jacobides and Lianos, 2021; Garud et al., 2022).

The regulation of STR platforms has proven challenging. One reason is that operators seek to avoid being classified under common categorization schemes, arguing that their technological advances and novel ways of operating set them apart from industry incumbents (Kaplan and Nadler, 2015; Mandel, 2016; Webb et al., 2020; Jacobides and Lianos, 2021). For instance, Airbnb fought in court to be considered an information society service, rather than falling under the same regulatory framework as the hotel or real estate industries (McKeen et al., 2018). Furthermore, due to their global operations, STR platforms are simultaneously subject to local, national, and supranational legal frameworks (Parker et al., 2016). As a result, there is often considerable uncertainty about the applicability of existing regulations, allowing platform operators to deliberately evade regulatory scrutiny (Uzunca et al., 2018; Garud et al., 2022; Lucas et al., 2022). Regulators, managers, and researchers are therefore interested in understanding effective regulation of STR platforms (Evans and Schmalensee, 2015; Edelman and Geradin, 2016; Gawer, 2016; Cusumano et al., 2021; Kira et al., 2021).

While the academic literature is far from a consensus on questions of whether, how, and what to regulate with respect to STR platforms, cities around the world bear the costs of STR platform operations. As such, city-level regulators have taken action to regulate STR platforms to address negative externalities (Miller, 2016; Chen et al., 2021). Key to informing policy and research on effective regulatory action is the investigation of platform-level outcomes: It is in the marketplace transactions between supply- and demand-side actors that the economic impact of regulatory action for platform operators can be observed. Therefore, to inform and enable evidence-based policy making and regulation of STR platforms, we address these issues and ask: How do local regulatory responses affect the operations of STR platforms?

We do this by assessing how different local regulatory responses, enacted in major cities across the European Union (EU), have affected the operations of the STR platform Airbnb in terms of supply of and demand for listings. As such, our results provide evidence on the direct economic impact of local regulation for both Airbnb and the hosts operating on the platform. Specifically, we disaggregate regulatory responses into the four tuples: motivations, actions, targets, and outcomes of 34 regulatory measures implemented in 13 European cities from mid-2015 to the end of 2019. We adopt a broad understanding of regulation as any state measure of a “command and control” nature (Black and Kingsford Smith, 2002), enacted and enforced by local authorities against hosts, with or without the cooperation of the STR platform.

Following a sequential, mixed-methods approach with a quantitative dominant design (Morgan, 2007; Venkatesh et al., 2016), we used qualitative data sources such as informant interviews, legal sources, and archival materials to ground the quantitative analysis of Airbnb listing data. In particular, we used qualitative data to establish links between motivations (e.g., housing shortage), actions (e.g., licensing requirements), and targets (e.g., reducing the number of listings) of regulatory responses. Based on a dataset containing longitudinal data on over 960,000 listings from 13 European cities, we apply a difference-in-difference specification with the synthetic control method to assess the effects of local regulatory responses on Airbnb’s operations. To assess the effectiveness of the regulatory responses, we compare the intended and actual effects of local regulation, which we obtain from the qualitative part of our work.

Although the net effect of regulatory responses on Airbnb’s operations is negative (i.e., dampening platform activity), our analysis also shows that this effect is subject to meaningful variation. First, our results show that certain regulatory instruments also have positive effects, such as bringing an influx of new listings to the platform. This suggests that hosts gained clarity about the legality of their market actions based on the regulatory instrument implemented (Deerfield and Elert, 2022). Second, we find that regulatory instruments have different effects depending on the type of host, such that certain instruments affect professional and private hosts differently. Third, we find evidence for the importance of cooperation by the platform operator to effectively enforce regulatory responses. In the absence of a clear path to enforcement, for instance, if the operator refuses to cooperate, local regulatory responses may remain ineffective. With these findings, our study adds to the academic literature on platform regulation and informs regulators, platform operators, and managers when evaluating the effectiveness of regulatory responses.

2. Background

Disagreement abounds in the academic literature over questions of whether, how, and what to regulate when it comes to digital platforms, and STR platforms in particular. A recent special issue of this journal (Jacobides and Lianos, 2021) highlighted that answering these questions is necessary for a baseline understanding among scholars, platform operators, and regulators. To support these efforts, we ground this paper in work on “regulatory voids,” situations in which there is little or no clarity about the precise regulatory state of an industry or market (Aldrich and Fiol, 1994; Short, 2013; Deerfield and Elert, 2022; Gao and McDonald, 2022). Regulatory voids also arise in situations where rules and regulations exist, but are ill-suited for a particular industry (see Uzunca and Borlenghi, 2019). These situations often arise in industries affected by the entry of digital platforms, where existing regulations fail to capture new economic realities. As a result, regulatory voids increase uncertainty about how and which business practices are permissible, and whether operations are subject to regulation at all. Such a lack of consensus among the relevant actors at the national and European level on the rules that should govern digital platforms in Europe has allowed STR platforms to gain greater market traction and evade regulatory scrutiny for as long as possible (Garud et al., 2022; Ozalp et al., 2022). This has contributed to the negative externalities exerted by STR platforms, which, in turn, motivate regulatory action at the local level.

2.1 Motivations for regulating STR platforms

The first question is whether STR platforms should be regulated at all. In practice, regulation in the EU and the United States has historically taken a hands-off approach, relieving platforms of regulatory burdens to minimize the risk of stifling innovation through overregulation (e.g., Shelanski, 2013; Cennamo and Sokol, 2021). More recently, however, the tide is turning, and digital platforms are increasingly subject to regulatory scrutiny to prevent the abuse of market power (Evans and Schmalensee, 2015; Gawer, 2016; Parker et al., 2022; Kira et al., 2021). Arguments in favor of regulating STR platforms emphasize that the absence of regulatory action helps to preserve regulatory voids, thereby protecting the operations of STR platforms, while risking the negative externalities of platform activity. As reflected in previous literature, the primary motivations for regulating STR platforms are to level the playing field with traditional hotel and accommodation providers, and prevent unfair competition, to adequately tax STR platform operations, and to avoid negative externalities on the social fabric of a city through overtourism or housing shortages (e.g., Guttentag, 2017; Nieuwland and Van Melik, 2018; Chen et al., 2021).

There is evidence that the unregulated presence of STR platforms exacerbates housing shortages, which have contributed to price increases in flashpoint cities such as Barcelona or Berlin (Duso et al., 2020; Garcia-López et al., 2020). In the USA, Sheppard and Udell (2016) find that, when the number of Airbnb listings doubled in New York City neighborhoods, housing prices increased by 6–11%. Horn and Merante (2017) find the same effect of an increase in Airbnb listings on long-term rentals in Boston. Barron et al. (2021) provide data for the entire USA and confirm the rent-increasing effect of Airbnb listings in cities, which appears to disproportionately affect low-income tenants (Lee, 2016; Wachsmuth and Weisler, 2018).

Similar negative externalities from STR operations have been documented in the context of urban problems with overtourism. For the city of Sydney, for example, Gurran and Phibbs (2017) present submissions from urban planners to local government inquiries explaining how the increase in STR platform listings has led to noise, nuisance, traffic, parking, and waste management problems in more residential areas of the city. In addition, there is some evidence that STR platforms have contributed to the process of gentrification in trendy European cities by replacing residential life with tourism (Füller and Michel, 2014; Gant, 2016).

The issue of business taxation has also been a motivation for the regulation of STR platforms. Regularly, STR platforms and commercial hosts compete with traditional hotel and accommodation providers without being subject to the same taxation schemes (Thelen, 2018). As a result, STR platforms avoid taxation by virtue of not being categorized as hotel businesses.

Finally, the issue of unfair competition is particularly prevalent in the hospitality industry, which is subject to specific laws that STR platforms have been able to circumvent (McKeen et al., 2018). The main argument for regulation is that STR platforms provide an equivalent service to their incumbent counterparts, and should therefore be considered under the same regulatory framework (Monti and Augenhofer, 2018). There is empirical evidence on the impact of STR platform activities on the hotel industry. For example, Zervas et al. (2017) show for various cities in Texas that the presence of Airbnb has a negative impact on hotel room revenues, especially for hotels in the lower price segment. When STR regulation is enacted, there is some evidence that this trend is reversed: Yeon et al. (2020) show that the introduction of STR regulation in various US cities has a positive impact on low-cost hotels, suggesting that the introduction of regulation has indeed leveled the playing field.¹

2.2 Difficulties with regulation

One of the difficulties in regulating STR platforms is their tendency to evade existing regulation (Kaplan and Nadler, 2015; Mandel, 2016; Webb et al., 2020; Jacobides and Lianos, 2021). Protracted disputes, combined with deliberate actions by platforms to evade scrutiny for as long as possible, often make enforcement of existing legislation difficult (Finck, 2018; McKeen et al., 2018). In addition, informing, implementing, and enforcing regulation is complicated by the fact that the global operations of most STR platforms make them simultaneously subject to local, national, and supranational legal frameworks (Parker et al., 2016). This leaves ample room for platforms to exploit regulatory voids across their geographical footprint. For example, platform operators tend to relocate their main operations to jurisdictions with lower regulatory burdens and oversight, and they respond discriminately to regulatory actions depending on the size of the market in question and the stringency of the regulation adopted (Uzunca and Borlenghi, 2019; Garud et al., 2022).

Another difficulty arises from the role of digital platforms as intermediaries. Platforms, by definition, are not vertically integrated, but rather create value by orchestrating a large number of complementors (e.g., hosts on Airbnb). This raises the question of whether regulation should target the platform organization, which would then act as a private regulator by imposing rules and constraints on complementors to comply with external regulations (Boudreau and Hagiu, 2009; Cutolo and Kenney, 2021). Alternatively, regulation can directly target complementors, whose actions arguably cause the negative externalities that regulation typically seeks to reduce. However, the supply side of most digital platforms, including STR platforms, consists of both non-professional, private individuals who share their underutilized resources to generate additional income (Einav et al., 2016) and professional players who engage in commercial activities on the platforms (Chen et al., 2023). Due to their practice of renting out multiple properties, professional hosts have been accused of causing the lion’s share of harmful effects, leading to increased attention from legislators (Chen et al., 2023). However, regulators often lack the necessary information to clearly distinguish between the two groups without the help of the STR platform. Thus, when regulators lack such information, this limits their ability to regulate effectively (Short, 2013). This reliance on cooperation with digital platforms is another difficulty for regulation in this context (Gao and McDonald, 2022).

STR platforms are for-profit organizations that seek to achieve favorable competitive positioning, increase their market presence, and scale their operations (Edelman and Geradin, 2016; Garud et al., 2022). In the absence of clear regulation, platform operators find themselves in situations of considerable uncertainty (see Uzunca and Borlenghi, 2019; Garud et al., 2022; Lucas et al., 2022). Responses to this uncertainty range from passive to evasive, neither of which are conducive to reducing the difficulties of regulating STR platforms. Operating in regulatory voids can incentivize platform operators to self-regulate. However, without a robust regulatory regime and strict oversight (i.e., monitoring and sanctioning), self-regulation remains symbolic with limited consequences for organizational behavior (Short, 2013). Thus, such strategies often favor minimal action until appropriate regulatory responses clarify legal ramifications (Uzunca and Borlenghi, 2019; Lucas et al., 2022). Perhaps not surprisingly, emerging evidence suggests that self-regulation of digital platforms is less effective than government intervention (e.g., Cusumano et al., 2021). For these reasons, defining an appropriate regulatory framework is challenging.

2.3 Attempts of local regulatory responses

Finally, the question of which level of government is appropriate for the formulation and implementation of regulation remains unresolved. Within the EU’s multi-level governance system, a long period of regulatory inaction at the EU level has left regulatory voids intact, while negative effects continued to be felt at the local level (Finck, 2018). The EU Commission acknowledges that “there is a risk that regulatory grey zones are exploited to circumvent rules designed to preserve the public interest” (European Commission, 2016a).

One of the key principles in determining at which level to regulate is the principle of subsidiarity,² which states that regulatory decisions should be made at the most “appropriate” level of governance (de Búrca, 1999). When it comes to regulating STR platforms, it is argued that cities or regions are the most appropriate level for regulation, given local information and enforcement (Finck and Ranchordás, 2016). In response, national and local policymakers have experimented with different forms of regulation, including co-regulation and state regulation (Finck, 2018). These initiatives are fueled by resistance from incumbent industry coalitions or grassroots movements (Kaplan and Nadler, 2015; Ranchordás and Goanta, 2020; Ricart et al., 2020).

While the EU regularly delegates responsibility to local municipalities, these may not have sufficient resources to enforce regulation and may rely on platforms to support their efforts (Cusumano et al., 2021). Nevertheless, there is some evidence that local regulation is promising. For example, Chen et al. (2021) find that regulatory responses in US cities are generally effective. Chen et al. (2023) examine the so-called “one host, one home” policy, which limits the number of listings a host can offer to one in US cities such as New York City and San Francisco, and find that the policy discourages professional hosts from using Airbnb. Comparing the effects of STR platform regulation across geographies reveals disjointed evidence. For example, in a study of different regulatory responses in cities around the world, Guttentag (2017) finds that taxes, tourism, and local housing are the main motivations, which are addressed through a variety of measures such as rental restrictions, licensing requirements, taxes, and others. Similarly, in a study of European and American cities, Nieuwland and Van Melik (2018) find that even when cities share motivations, the policy instruments chosen for regulation and the effects of regulation vary widely.

In sum, the questions of whether, how, and what to regulate in relation to STR platforms are largely unresolved and empirical evidence on the impact of regulatory responses on the actual operations of STR platforms is scarce, making the formulation of policy responses difficult (cf., Nooren et al., 2018). As a result, the effectiveness of specific regulatory responses remains poorly understood and controversial (Cammaerts and Mansell, 2020). Therefore, evidence on the effectiveness and appropriateness of different options is needed to inform policy decisions (European Commission, 2016b).

3. Methods

We conduct a sequential, mixed-method analysis with a quantitative dominant design (Morgan, 2007; Venkatesh et al., 2016). We draw on qualitative data sources such as informant interviews with policymakers, regulators, and STR platform representatives as well as legal documents and archival materials to guide a quantitative analysis of Airbnb listing data to assess the impact of regulatory responses to Airbnb in Europe using a difference-in-difference model specification with synthetic controls.

3.1 Data

Our analysis focuses on a sample of 34 implemented regulatory responses in the top 13 European cities³ by number of Airbnb listings between July 2015 and December 2019 (54 months). This scope allows us to examine a considerable degree of variation between regulatory responses, as European cities have responded differently.⁴ At the same time, it ensures a meaningful comparison in context: all cities are subject to the supranational EU regulatory framework within which such responses must be articulated. The justification for our focus on city-level regulatory instruments is twofold. First, many European cities have already taken regulatory action in response to problems caused by STR platforms. This allows us to compare the effectiveness of different regulatory options. Second, the negative impact of STR platforms on local economies is particularly acute in large European cities (Füller and Michel, 2014; Gant, 2016). We further disaggregate the data to the districts/neighborhood⁵ level because some measures adopted by cities to regulate STR hosts explicitly target listings in specific neighborhoods (e.g., Barcelona, Paris, and Vienna).

We collected data on all regulatory responses in the cities from three sources (Table 1). First, we consulted archival sources like statutes, judgments, and news items to document the timeline of regulatory responses in each city. The archival data sources allowed us to identify 34 regulatory responses that targeted Airbnb and other STR platforms and that were implemented during the observation period.⁶

Second, we supplemented the archival data with informant interviews with policymakers, industry experts, and local officials as well as Airbnb representatives with knowledge about the regulatory responses to STR platforms in some of the cities. The interviews were semi-structured and asked informants to report on the motivations, policies, and enforcement of the regulatory instruments.

Lastly, we collected monthly listing data for the sampled European cities from the website Inside Airbnb⁷ which reports and visualizes scraped data from Airbnb. The data includes the full details of each listing, including price, location, size of the property listed, as well as additional information specific to Airbnb, such as availability of listings in the coming weeks, type of property (shared or entire home), and user review ratings. Across the 13 cities, our dataset contains 12.7 million observations on over 960,000 listings, which we observe for a maximum of 54 months. We use the Eurostat⁸ database to complement this data with time-varying economic and social indicators that may affect the regulatory framework in the given city.

3.2 Conceptualization of regulatory responses

We used data from the different sources to capture four elements of regulatory responses: the (i) motivation (M), (ii) action (A), and (iii) target (T), of a regulatory response as well as (iv) its outcome (O) (M–A–T–O). This disaggregation of platform regulation is generally consistent with past conceptual work (e.g., Nooren et al., 2018). We used qualitative data sources to construct the motivations and actions, while data on Airbnb listings was used to capture targets and outcomes measured as the impact on Airbnb’s operations. In our investigation, the outcome of each regulatory response is represented by the sign and significance of the average treatment effect on the treated (ATT) from a synthetic control method (described in detail further below). Outcomes thus link motivations, actions, and targets of regulatory responses and allow us to assess desired and actual effects of regulation. Table 2 provides an overview of exemplary regulatory elements.

3.2.1 Motivation

Through our work with informants and archival material we came to understand four main motivations for regulating Airbnb in European cities. Regulatory interventions were motivated by one or more of the following: (i) housing market dynamics (e.g., housing shortage), (ii) (over-)tourism, (iii) taxation, and (iv) competition with incumbent industries such as real estate or hospitality. These general motivations resonate with practitioners as well as previous work on the regulation of Airbnb (e.g., Guttentag, 2017; Nieuwland and Van Melik, 2018; Chen et al., 2021).

Equally consistent across informants and past work was the insight that similar motivations were associated with varying instruments that regulators implemented to put regulations into action. As one policy advisor summarized: “In the case of Airbnb, formulating a regulatory response was motivated by balancing competing interests [...] on the one hand, ensuring prosperity in terms of growth, innovation, and business. […] on the other hand, ensuring fairness for three groups of stakeholders: hoteliers, real-estate owners, and tenants” (interview with policy advisor in Copenhagen, Denmark). See Section S1 of the Supplementary Material for a complete list of interviewees.

3.2.2 Action

Using statutes and administrative acts, news archives, and expert interviews as sources, we derived a timeline of the main regulatory instruments implemented in each city in our sample. The most common include limits on the number of days hosts can rent out a property consecutively or cumulatively in a given period (day caps), licensing schemes to control location and density of hosts (licensing/registration), or systems to report and/or collect taxes on revenues generated through short-term rentals on platforms (tax collection or tax allowance). We provide a complete overview of all 34 identified regulatory responses in the Supplementary Material (Table S3).

3.2.3 Target

Each of the identified actions described above intended to affect a specific measurable objective. Targets are measures of interest to regulators and policy makers. They are relevant because they approximate the desired outcome of regulatory responses. For example, if a local authority wants to protect the city center from overtourism, then one target of regulation can be to limit the number of Airbnb listings in certain neighborhoods. Regulation targets thus represent the operationalization of the dependent variable from our data, and we focus on variables that capture the supply of and demand for listings on the Airbnb platform (Chen et al., 2023). We describe all measures below.

3.2.4 Outcome

To analyze the outcome associated with any given regulatory response, we constructed a panel dataset containing 14,726 neighborhood-month observations across the 13 European cities for the period from July 2015 to December 2019 (i.e., before the start of the COVID-19 pandemic). As described above, we conduct our analysis at the neighborhood level (i.e., collapsing the dataset at the neighborhood-month level), while clustering at the city level, because some regulations only target specific neighborhoods rather than the entire city.

3.3 Measures

We reconstructed the timeline of regulatory instruments implemented across the cities in our sample based on the archival data sources. For each of the regulatory instruments we identified (e.g., day caps, licensing requirements, etc.), we coded a binary variable indicating that the particular instrument was implemented in the given city. The variable is coded as 1 from the month the respective regulation was implemented and retains this value for all post-treatment periods. For all pretreatment periods, or if the respective regulatory instrument was never implemented in a given city, the variable is coded as 0.

We also assessed whether Airbnb cooperated with the local authorities to enforce the respective regulatory instrument. For instance, where the local authorities required hosts to obtain a registration number, we checked whether Airbnb subsequently changed the interface so that providing the registration number became a requirement for prospective and current hosts when creating or modifying a profile. In other instances, Airbnb agreed to share data about hosts’ activities with local authorities to enable proper taxation. The corresponding variable cooperation was coded as 1 if we found evidence of Airbnb’s cooperation and 0 otherwise.

To capture how the implementation of the regulatory instruments affected Airbnb’s operations in each city, we assessed changes in the supply of and demand for listings using four platform measures. These metrics (described below) are of interest to policymakers and regulators because they reflect the direct impact of regulatory responses on Airbnb’s operations.⁹

1. Added listings (supply) measures the number of listings added in a given neighborhood and month. To construct this measure, we check for each listing whether it was available on the platform in the preceding month. In this way, we capture both listings that are completely new (i.e., never listed before) and those that are relisted after being temporarily removed.

2. Removed listings (supply) measures the number of listings removed in a given neighborhood and month. For most regulatory instruments, STR platforms did not have to actively remove listings that did not comply with the new regulation. Rather, they were removed voluntarily by the host (e.g., due to lack of bookings or potential risk of non-compliance with regulation).

3. Availability (supply/demand) measures the number of days listings are available for booking, on average, over the next 30 days. Hosts can mark their listings as available or unavailable for specific dates using a calendar feature provided by Airbnb. Prospective guests can then see which days a particular listing is available and either book directly or ask the host for confirmation. When a listing is booked, that time period is automatically removed from the calendar, reducing the availability of that listing. Thus, if a listing is marked as unavailable in the calendar for a particular period, it means either that the host has not made it available (supply) or that it has already been booked (demand). Therefore, the variable availability captures fluctuations in both supply and demand. However, we are primarily interested in the supply-side effects of regulatory actions on hosts’ decisions to make their listings available for booking on the platform or not, and to a lesser extent in the demand-side effects, which we primarily capture with the variable reviews. Thus, we control for the possible demand component by including the variable air traffic in all calculations. Furthermore, as can be seen in Table 4, the pairwise correlation between availability (potential bookings) and reviews (actual bookings) is low (−0.137). We are therefore confident that availability, as measured in this study, mainly captures a supply-side effect.

4. Reviews (demand) measures the number of new reviews posted for listings in a given neighborhood. As Airbnb does not disclose how often listings have been rented out, we approximate that figure using the above measure. Though review rates on Airbnb have been estimated to range from as low as 18.6% (Ke, 2017) to as high as 72%,¹⁰ we assume that a guest’s propensity to post a review is relatively stable across listings in a given neighborhood and across time. Furthermore, the propensity is also likely to be unaffected by the implemented regulatory responses as these did not target guests directly. As such, the number of bookings is a linear function of the number of reviews posted for a given listing (Ye et al. 2009).

Importantly, for all these measures, we distinguish between commercial listings offered by professional hosts and private listings offered by private individuals. Following Chen et al. (2023), we define professional hosts as those who offer more than one entire home on Airbnb in a given month¹¹ and the listings (entire homes) offered by these hosts as commercial listings. Based on these criteria, an average of 20.71% of hosts were classified as professionals, offering 29.46% of all listings on average (at the city and month level). However, the proportion of commercial listings varies greatly from city to city and fluctuates over time (see Table S4 in the Supplementary Material). Compared to private listings, commercial listings offer a more hotel-like experience (e.g., listings can be booked directly without host confirmation) and are more likely to be specifically targeted at business travelers, but they are also significantly more expensive.¹²

3.4 Controls

To control for alternative explanations of effects, we account for differences within and across cities by including a vector of time-variant economic and social indicators at the city or country level. First, we measure Air Traffic per city as the number of incoming passengers by plane per inhabitant based on data retrieved from Eurostat (air passenger transport by main airports in each reporting country [AVIA_PAOA])¹³ to account for relative popularity of a city as a tourist destination in the respective month. Second, we also obtain data from Eurostat (harmonized unemployment rates [EI_LMHR_M])¹⁴ to control for the Unemployment rate at the country level in the given month as a proxy indicator for individuals’ propensity to rent out property (or parts thereof) on Airbnb. We provide descriptive statistics as well as variable correlations in Tables 3 and 4 below.

3.5 Analysis

Although cities across Europe have implemented similar regulatory measures, the timing of implementation and the geographic reach in terms of neighborhoods varies from city to city. Given the sequential implementation of regulatory responses, we estimate the average treatment effect on the treated (ATT) using the generalized synthetic control (GSC) method as proposed by Xu (2017). Compared to the underlying synthetic control method (Abadie et al., 2010), which has gained popularity for quasi-experimental designs in several fields, the GSC method allows for multiple treatment units with different treatment times. The idea is to model synthetic control units that closely represent the units in the treated group with respect to the outcome of interest to subsequently compare the evolution of the average outcome between the treatment and control groups. In our case, all neighborhoods in which a particular regulatory intervention was implemented serve as treatment units, while all others (within and across cities) are used to construct a weighted synthetic control group based on pretreatment periods. The counterfactual (i.e., control group) represents what would have happened in the absence of the regulatory response. Any differences in outcomes in the post-treatment periods between treated and control units can thus be attributed to the treatment. By constructing a counterfactual outcome that represents what would have happened had the regulatory response not been implemented, we are able to mitigate issues related to endogeneity and reverse causality. In the GSC method, the statistical significance of the estimate is evaluated based on a bootstrapped distribution of the estimate, which is similar to the placebo test in the traditional synthetic control method.¹⁵

4. Results

In the following subsections, we present each regulatory response through the actions taken at the local level. We then report the effects each identified instrument had on the supply of and demand for listings through the platform metrics described above. The results are presented in condensed form for the benefit of the reader, and we provide the full model results in Section S5 of the supplementary material. We include both location and time fixed effects for each neighborhood-month combination.

4.1 Regulation involving temporal or geographical restrictions

Temporal restrictions have been implemented through so-called day caps, which limit the number of days a listing can be rented out through the platform in a calendar year. Instantiations of this regulatory instrument vary in length, ranging from 30 days (Amsterdam) to 120 days (London), and in whether Airbnb supported the regulation through cooperation (e.g., in Amsterdam, Barcelona, and Paris). Informants indicated that this action was motivated by tensions in the housing market and the risk of overtourism.

The results for this regulatory action are shown in Table 5. They indicate that the number of private listings remains unaffected. However, the estimated ATT is positive and significant for both added and removed commercial listings. About 11% more commercial listings are being added in the treated neighborhoods, while almost 29% more are removed. The effect for removed commercial listings is even more pronounced in cities in which Airbnb cooperated in enforcing day caps (compare row cooperation in Table S5 in the Supplementary Material). At the same time, availability decreased for all listings. The number of days listings were available for booking over the next 30 days in treated neighborhoods decreased by 0.4 days for private listings and by over 3 days for commercial listings. Interestingly, when Airbnb cooperated in enforcing the day caps, the availability of commercial listings decreased less severely. Despite these changes in the supply of listings, the demand (i.e., reviews) remained unchanged.

In the following, we provide an initial interpretation of these results. First, private hosts are unlikely to exceed the day caps under the new regulation, so it is not surprising that there is no significant effect on the number of private listings. In contrast, the increase in removed commercial listings suggests that these hosts have concluded that their business model is no longer viable under the new regulation. On the other hand, the increase in newly added commercial listings suggests that the new regulation provides clarity and guidance for other professional hosts. Specifically, the regulatory action resolved gray market rule-breaking into either black (illegal; i.e., exceeding the day cap) or white (legal; i.e., staying below the day cap) market action (Boettke et al., 2004; Lucas et al., 2022). Second, the reduced availability of listings is consistent with our assumption that the variable captures hosts’ deliberate actions to make their listings available or unavailable for booking through the platform (see Section 3.3). Specifically, with temporal restrictions in place, hosts need to plan their activity on the platform more carefully to comply with the new regulation. The stronger effect of the regulation for commercial hosts suggests that they are acting more strategically, carefully planning when to make their listings available for rent (e.g., when demand and prices are high) to generate the highest profit (Wessel et al., 2017). When Airbnb cooperated in enforcing day caps, this typically entailed that the platform would implement technical measures to restrict the number of days a particular listing can be made available for booking. Hosts in these cities could thus be sure not to accidentally violate the new regulation, explaining why the decrease in availability was less severe in the respective cities. Overall, the regulatory action was effective in restricting the supply of, especially, commercial listings with no adverse effects for the demand.

Geographical restrictions, in turn, involve regulatory responses that prohibit (certain) Airbnb listings in specific locations, such as popular neighborhoods near city centers or areas popular with tourists. Thus, hosts who continue to operate in such neighborhoods do so illegally or at least at the fringes of legality, by exploiting poor enforcement (Garud et al., 2022).

The results for geographical restrictions involving local bans of Airbnb are shown in Table 6. In terms of number of listings, only the ATT for removed private listings is negative and significant. This indicates that about 15% fewer listings are being removed in treated neighborhoods. However, in neighborhoods in which Airbnb cooperated in the enforcement of local bans, we observe significant effects for both private and commercial listings (compare row cooperation in Table S6 in the Supplementary Material). Specifically, when Airbnb cooperated, fewer private listings are being added (−6%) and removed (−8%) and more commercial listings are being added (12%) and removed (27%) in the treated neighborhoods. Furthermore, even though the main effect for availability of commercial listings is insignificant, for neighborhoods in which Airbnb cooperated in the enforcement, the average availability of commercial listings decreases by 0.3 days. Finally, the negative ATT for reviews written for private listings indicates that the demand for the listings decrease by 46%, while the demand for commercial listings remained unaffected.

Taken together, the effects of local bans indicate an overall contraction of listing supply, but only in locations in which Airbnb cooperated in the enforcement. The stark increase in removed listings, especially commercial listings, suggests that the regulatory instrument is effective, if technical measures are implemented to enforce it on the platform. However, at the same time, the demand for commercial listings remained unaffected but starkly decreased for private listings.

4.2 Regulation involving registration or licensing requirements

Registration of host activity does not preclude any action by local authority as the registration of hosts alone is without consequence beyond informing authorities. Licensing, on the other hand, requires approval by authorities and hosts typically need to fulfill certain requirements in order to obtain a license. Neither the effect of registration nor licensing is well understood ex ante (cf., Nooren et al., 2018). Registration in particular was often motivated by trying to understand and control host activity (e.g., number and location of listings) to counteract overtourism and tensions on the housing market, as well as for taxation purposes. In Table 7, we report the effects of host activity in Airbnb after hosts were required to register.

The number of added listings on Airbnb is unaffected by the need for hosts to register. However, more private (9%) and commercial (17%) listings were removed from Airbnb in locations in which registration became mandatory, indicating a substantial churn of listings from the platform. This effect was less pronounced if Airbnb cooperated with authorities to enforce the regulation (e.g., by sharing data with authorities or by requiring hosts to enter their registration number before being able to publish a listing; see Table S7 in the Supplementary Material). We also observe a significant decrease in the number of days the remaining private (−0.4 days) and commercial (−2.4 days) listings are available for booking. Further, the requirement to register as a host also had an adverse effect on the demand for both private (−26%) as well as commercial (−32%) listings, indicating a stark reduction of transactions facilitated in regulated locations.

Overall, the observed effects of host registration on supply of and demand for listings suggest an effective regulatory response. While the number of added listings remained on the same level as in untreated neighborhoods, more listings were being removed from the platform after the implementation. Further, the remaining listings were available for fewer days on average and the number of actual transactions (reviews) contracted. Across the board, the effects are more pronounced for commercial listings than for private listings. All this points to an overall correction of market behavior as hosts are required to disclose that and how they are active on STR platforms.

In the following, we report the findings relating to regulatory responses that involved a licensing requirement. This regularly meant that hosts wishing to be active on STR platforms needed to apply for permission at their local authorities, who in turn controlled the permissions by issuing licenses to hosts. In some locations (e.g., Paris), Airbnb cooperated with authorities by requiring hosts to report information that would match them to issued licenses, for instance, by asking hosts to add a license number to their listings. Similar to registrations, licensing was often motivated by finding a response to the housing crisis, to overtourism as well as scrutinizing commercial host activity that might potentially compete with the incumbent hotel industry.

The effects of such responses are shown in Table 8. The outcomes of this action affected supply and demand on Airbnb negatively across the board, with the exception of removed commercial listings. In neighborhoods in which a license became a requirement, 21% fewer private listings and 22% fewer commercial listings were added to the platform compared to untreated neighborhoods. Furthermore, 17% more private listings were being removed, whereas 30% fewer commercial listings were being removed. This effect for commercial listings was even more pronounced in locations were Airbnb cooperated in enforcing the regulation (compare row cooperation in Table S8 in the Supplementary Material). Further, on average, both private as well as commercial listings were available 0.4 fewer days in treated neighborhoods. Finally, results also indicate a significant decline in the demand for both private (−70%) and commercial (−56%) listings.

Assuming that a regulatory response involving licensing host activity is motivated by reducing Airbnb activity in selected locations, this instrument indicates a considerable degree of effectiveness. Across the board, the supply and demand on Airbnb contracted and effectiveness was often increased further in locations in which Airbnb cooperated in enforcing licensing requirements.

4.3 Regulation involving taxation requirements

Taxation requirements oblige hosts and/or Airbnb to report host income and are sometimes combined with tax-free allowances. In some locations, Airbnb cooperated with municipalities and implemented technical features that automatically shared host income information with local authorities (e.g., Copenhagen). According to our informants, the motivations for these actions were to level the playing field for competition as well as ensuring applicability as well as enforcement of tax legislation. This regularly meant that existing tax rules were clarified to be applicable to income generated on STR platforms. In addition, the application of tourism taxes was extended to STR-type lodging.

Modeling these decisions in the same way we treated the instruments above is well in line with our analysis. First, clarifications of the applicability of existing tax rules mean a transition from less regulated (i.e., gray market) to a more regulated status quo (i.e., black vs. white market). Second, the city level remains appropriate since national tax responses were often motivated by the dynamics on STR platforms that particularly affected large cities. Furthermore, many cities extended their tourism taxes to STR platforms in their own capacity, for instance, Lisbon, Berlin, Barcelona, and Amsterdam. Third, while many tax rules often apply at national level, they affect cities such as Madrid, Barcelona, Venice, and Paris where local tax authorities are empowered to request information from Airbnb.

The results corresponding to taxation requirements are shown in Table 9. We observe a significant increase in the number of added private (16%) and commercial (29%) listings, but an even stronger increase in the number of removed private (43%) and commercial (63%) listings. Thus, in terms of number of listings, the net effect of the regulatory response is negative, meaning that more listings are being removed than added in treated neighborhoods. Availability of private listings remained unaffected by the regulatory response but increased by 0.1 days in locations in which Airbnb cooperated in the implementation (see Table S9 in the Supplementary Material). Commercial listings, on the other hand, saw a strong reduction in availability by 2.2 days, which further increased in locations in which Airbnb cooperated. The effect on the demand for listings (i.e., number of reviews) is positive for both private (51%) and commercial (48%) listings.

An initial interpretation of the effectiveness of taxation requirements paints a mixed picture. While more listings are being added to the platform and listings are booked more often in cities in which the regulation on taxation had been overhauled, at the same time, more listings are being removed from the platform and the remaining commercial listings are available for fewer days.

The fact that more listings are being added in the treated neighborhoods could indicate that the regulatory response provided clarity to hosts who were previously concerned about the gray-market nature of STR lodging prior to the implementation. These hosts gained legal certainty about their (prospective) STR activity. Other, especially commercial, hosts may have decided that the new policy challenges the viability of their business model.

4.4 Validity and robustness checks

We conducted two additional sets of tests to further increase the confidence in the reported effects and assess their validity. First, in some of the European cities the implementation of regulatory responses was announced weeks or even months in advance of the actual implementation. In these instances, our estimations of treatment effects might be biased as hosts might exhibit the behavior that we intend to measure as a response to the announcement of the imminent implementation rather than to the implementation itself. For instance, if the implementation of a local ban is announced months in advance, hosts might decide to remove their listings early to avoid repercussions should they forget to remove the listing in due time. For each regulatory response that was implemented during the observed period, we thus did not only search for information on when it took effect, but also traced whether the implementation was publicly announced more than a month beforehand. This was the case for 10 out of the 34 regulatory responses. We subsequently repeated our analysis with the same specification used in the main analysis but considered units to be treated from the month in which the regulatory response was announced rather than enacted. To 95%, the results are consistent with those of our main analysis with regard to the signs and size of coefficients. However, we find that for some of the regulatory instruments the main effects can indeed be expected at the time of the announcement rather than on enactment. For instance, in our main analysis, registration requirements had a negative but insignificant effect on new private and commercial listings being added to Airbnb. In Athens and Paris, the enactment was announced several months in advance. Considering the day of the announcement in our analysis, we find that, as a response to registration requirements, the number of new private listings being added to Airbnb reduced by 13% and the number of commercial listings by 7%.

Second, while all regulatory instruments have specific nuances (e.g., temporal restrictions differ with regard to the maximum number of nights per year), due to our analytical approach, we consider the instruments to be mostly comparable across cities. However, instruments implemented for the purposes of taxation yielded two relatively distinct responses that deserve further investigation: tourist tax and income tax. While cities such as Lisbon and Vienna focused on city-level tourist taxes, the remaining taxation requirements focused on income tax. The tax rate for the tourist tax is typically below 10%, while the tax rate for income tax generated through Airbnb listings can be significantly higher. Further, tourist taxes can be introduced by cities autonomously, while changes to income tax regulations are typically implemented at the national level—though they may be triggered by city-level initiatives. We thus calculate treatment effects separately for subgroups of cities focused on tourist tax and those focused on income tax. We find that the signs and significance levels of the ATT match those of our main analysis for both tourist tax and income tax. However, cumulative treatment effects are generally much stronger for cities in which income taxes had been adjusted.

5. Discussion

We set out to understand the effects of local regulation on Airbnb’s operations in major European cities. This section provides a summary of our main findings and possible explanations for the observed effects. First, we identified different types of negative externalities that motivated municipalities to regulate STR platforms, namely, housing shortage, overtourism, lack of taxation, and unfair competition. We then used qualitative data sources such as interviews, legal sources, and archival material to link these motivations to the particular actions taken by municipalities to regulate STR platforms in the attempt to mitigate the negative externalities without stifling innovation. Interestingly, even when cities shared motivations, this often triggered different actions across the sampled cities (Nieuwland and Van Melik, 2018). For instance, issues on the housing market were mentioned as the motivation for implementing day caps, local bans, registration requirements, or licensing requirements. Yet, the implemented regulatory responses are remarkably similar across European cities. Day caps, for example, only vary in length (30 days in Amsterdam versus 120 days in London) but are otherwise essentially identical. We thus observe a learning effect across jurisdictions, in the sense that municipalities have copied and adapted regulatory instruments within and across national borders, albeit the implementation was triggered based on a variety of motivations. Similar learning effects have been observed in the US context, where several states took inspiration from Indiana’s 2015 ridesharing law and were thus able to pass legislation to regulate Uber and similar platforms more quickly (Deerfield and Elert, 2022; Garud et al., 2022).

Second, we observe both negative as well as positive effects of regulatory responses for Airbnb’s operations. For instance, more private and commercial listings were added to Airbnb in response to taxation requirements. Thus, filling this regulatory void seems to have provided hosts with the necessary clarity to consider joining the platform (Deerfield and Elert, 2022). Yet, the net effect of each of the examined regulatory responses on Airbnb’s operations is negative. For instance, despite more listings being added to Airbnb in cities that implemented taxation requirements, the number of removed listings is considerably higher. The finding of a negative net effect aligns with prior research on STR platform regulation (e.g., Han and Zhang, 2020; Chen et al., 2021). One exception is the so-called “one host, one home” policy that was implemented in New York City and San Francisco in 2016, and later expanded to other cities.¹⁶ Here, the exit of professional hosts from the platform encouraged more non-professional hosts to join, which in turn helped the platform to secure its revenue despite some churn (Chen et al., 2023).

Third, a clear pattern emerges when comparing regulatory responses that restrict the use of Airbnb (i.e., temporal or geographical restrictions) with those that clarify the conditions under which Airbnb can be used in compliance with the law (i.e., registration, licensing, or tax requirements). Specifically, Airbnb’s operations are affected less by restricting policies, both in terms of effect size as well as number of platform metrics affected. For instance, the strongest negative effects for the number of added and removed listings as well as the number of reviews can be observed in response to regulation that was meant to clarify rather than restrict the use of Airbnb. Only listing availability is impacted most negatively by the implementation of temporal restrictions (i.e., day caps), whereas geographical restrictions (i.e., local bans) are notably ineffective. Previous work on regulatory uncertainty provides an initial explanation for this pattern in our findings. For regulation to be effective, there must be some general agreement among the stakeholders about the values underlying the regulatory regime, and shirking must be prevented through surveillance and enforcement (Black, 1997; Haines, 2013; Short, 2013). In this regard, restricting regulatory responses will be highly contested by Airbnb and its hosts. Given that municipalities often had limited resources for surveillance and enforcement, at least initially, and the risk of being caught shirking was therefore low, non-compliance might be considered an acceptable risk by some. Instead of complying with the emerging regulation, which could have threatened Airbnb’s survival, the company and its hosts persisted in the face of regulatory headwinds, thereby “probe[ing] the contours of regulatory boundaries, which may be in flux” (Gao and McDonald, 2022: 951). This is not unusual for platform-based ventures. For instance, Uber repeatedly entered markets despite knowing that their operations would violate existing regulation, and against explicit warnings by regulators, because the company knew that waiting for permission would jeopardize its business model (Uzunca et al., 2018; Deerfield and Elert, 2022; Garud et al., 2022). Such rule-breaking action by platforms may force regulators to clarify their stance or even return to dialogue and negotiation (Gao and McDonald, 2022; Lucas et al., 2022).

Fourth, while non-professional hosts typically use STR platforms to share (parts of) their own homes, professional hosts buy or rent multiple properties (i.e., entire apartments or houses) for the sole purpose of making them available on STR platforms for a profit. Thus, many of the harmful effects attributed to Airbnb’s presence are caused primarily by professional hosts, which have also been shown to drive out non-professional hosts despite operating in different segments of the market (Chen et al., 2023). Though both municipalities and STR platform providers thus have motive to regulate the participation of professional hosts, none of the European regulatory responses we examined explicitly targeted this group or the listings they offer.¹⁷ Nevertheless, our results show that local regulations may affect the listings of professional and non-professional hosts differently. Regulatory responses designed to clarify the conditions under which STR platforms may be used in accordance with the law produced results that were comparable in terms of sign and significance level between private and commercial listings, but most effect sizes were significantly larger for commercial listings. Regulatory responses that restricted the use of Airbnb had more differential effects on either private or commercial listings. For instance, temporal restrictions primarily affected the number of commercial listings, with little effect on private listings.

Finally, our results show that, whether, and to what extent, regulatory responses had an impact on Airbnb’s operations often depended on whether the platform cooperated with municipalities. For instance, registration requirements did not impact the number of commercial listings being added in treated neighborhoods. Yet, in cases where Airbnb cooperated by changing the interface, so that providing the registration number became a requirement, we observe a positive effect for the number of listings added.

5.1 Theoretical implications

Our findings make several contributions to the literature on platform regulation. First, most existing empirical research has been focused on examining the negative effects of STR platforms on the external environment in which they operate. Research has focused either on the disruptive effects of the STR platforms on incumbent businesses, such as hotels, for which unregulated platform activity constitutes unfair competition (e.g., Zervas et al., 2017; Li and Srinivasan, 2019; Yeon et al., 2020), or on the effects on the cities themselves such as the impact on the housing market (e.g., Horn and Merante, 2017; Barron et al., 2021). Our study complements this research by shifting the focus from the external environment to the internal platform operations. Thus, by examining how local regulatory responses affect Airbnb’s operations in major European cities, we provide unique insights into the effectiveness of platform regulation. Here, we pay particular attention to the fact that digital platforms are not vertically integrated, but instead manage a large network of complementors (i.e., hosts in the case of STR platforms), all of whom are autonomous agents (Parker et al., 2017). This makes the regulation of such platforms particularly challenging. For instance, in the case of STR platforms, legislation that directly targets hosts is difficult to enforce because municipalities often lack sufficient data to identify violations and the resources to enforce sanctions, so individual hosts are often considered low-priority offenders (Lucas et al., 2022). On the other hand, the effectiveness of legislation that directly targets the platform as a mega-organization depends on the ability and willingness of the platform provider to enforce it. Here, our study provides novel insights into the impact of regulatory responses on Airbnb’s operations in terms of supply and demand, as well as on the reinforcing or dampening effect of Airbnb’s cooperation.

Second, prior research has shown that regulation is a double-edged sword for entrepreneurship (Friske and Zachary, 2019). Specifically, it is a delicate balance for any governance regime to provide “a functional regulatory environment [that] incentivizes entrepreneurs” (Audretsch et al., 2019: 1150) without surpassing the point at which overregulation discourages entrepreneurship and stifles its welfare effects (Klapper et al., 2006; Braunerhjelm et al., 2015). Thus, attempts to mitigate the social costs of STR platform operations through regulation might come at the expense of entrepreneurial activity. However, if the regulatory voids surrounding the activities of STR platforms are not filled, both the platform and its hosts will experience a significant amount of uncertainty. Thus, a lack of applicable regulation that provides structure, predictability, and protection may discourage them from investing time, effort, and resources to create value (Bjørnskov and Foss, 2013; Lucas et al., 2022). Our findings provide support for this view on the relationship between regulation and entrepreneurship. For instance, regulatory responses that clarified the applicability of existing tax rules to STR platform activity discouraged many, especially professional, hosts from continuing their activity on the platform. However, a significant number of hosts decided to join the platform only after the regulatory response took effect, arguably due to the decrease of uncertainty regarding the applicability of extant rules. Thus, filling the regulatory voids may not only deter platform activity but also encourage it (Uzunca and Borlenghi, 2019). Prior to the implementation of the new regulation, both the platform and its hosts thus exploited “interpretive ambiguity to skirt the potential application of rules” (Lucas et al., 2022: 2). The regulatory response subsequently resolved gray market rule-breaking into either black (illegal) or white (legal) market action (Boettke et al., 2004; Lucas et al., 2022).

Finally, we align with others who contributed useful conceptual language to the literature on platform regulation (e.g., Nooren et al., 2018). Using established conceptual ideas to produce empirical evidence, we emphasize the need for robust yet flexible assessment of regulatory responses. This helps articulate how regulators, platform operators, as well as researchers can frame the many parameters of digital platform regulation to inform their debates. In doing so, our proposed conceptualization also provides a vocabulary to frame topics beyond the context of short-term rental platforms For instance, the terminology of the four-tuple motivation, action, target, outcome (M–A–T–O) could be applied equally well to frame regulation of transportation services (e.g., Uber), or media platforms (e.g., TikTok). Here, our study provides a parsimonious yet comprehensive framework with utility for empirical research.

5.2 Policy and managerial implications

For policymakers and managers, our findings add much-needed nuance to the evidence base used to inform the debate on whether and how to regulate short-term rental platforms. This has implications for policymaking in practice as we show that certain responses are effective. Policymakers can use the findings presented here to think through the link of outcome and action to affect certain targets of interest and in line with their motivation to act. Empirically, our findings suggest that local (i.e., city or neighborhood) regulatory responses can impact the operations of STR platforms significantly.

A recurring theme in the findings is that effects differ contingent on Airbnb’s cooperation. On the one hand, this would support legislation that increases the responsibility of online platforms and obligations in the removal of illegal online content such as unlawful listings on Airbnb. Such legislative package should come with provisions that enable cooperation between local and national authorities in enforcing cooperation obligations imposed on online platforms. This is precisely the approach the EU Digital Services Act is pursuing (European Commission, 2022). The new EU regulation overhauls the liability framework for platforms that will place a larger burden on platforms to police their users and eliminate any illegal user content (European Commission, 2020). A similar reform of federal US law is being debated (Monti and Augenhofer, 2018). This will increase the burden for STR platforms to screen for illegal offers (e.g., because they contravene local regulation). The EU Digital Services Act will empower local municipalities to enforce their local regulations more effectively, both with the help of platforms and the help of other public enforcement bodies in the member states where platforms have their company seats.

On the other hand, if effective regulation hinges on the platform’s cooperation, this may, in turn, allow platform firms to secure a “seat at the table to collaboratively propose, iterate, and shape emerging regulations” (Gao and McDonald, 2022: 951). Such regulatory co-creation represents a novel logic of direct interaction between platforms and regulators (Gao and McDonald, 2022).

5.3 Limitations and future research

The limitations of this paper present opportunities for future research. First, the objective for this research was to study regulatory responses across European cities that have a strong presence of Airbnb. This level of analysis meant that we focused on instruments that were repeatedly seen across cities and we could not consider the nuances of regulation implemented in the respective cities. We thus left out regulatory responses that might have had an impact on Airbnb activity in one specific city while not being replicated in other cities.¹⁸ Future research could thus focus on specific instruments to study how differences in the implementation across cities impact the policy’s effectiveness.

Second, some of the cities in our sample repeatedly changed existing regulations. In most cases, the regulations became more restrictive. For instance, Amsterdam updated its regulation on temporal restrictions (day caps) three times, reducing the number of nights a listing can be rented out from 60 to 30 days. However, in our analytical approach, cities or neighborhoods are considered treated from the month in which the regulatory measure was first enacted. Thus, our analysis does not explicitly capture the effects of these revisions. While we expect the initial implementation to have the largest impact on Airbnb’s operations, future research should consider the effects of revisions of regulation.

Third, for the regulatory responses we examine, enactment and enforcement almost always coincide (i.e., begin on the same day). However, in some cases, the enforcement was intensified by city officials (e.g., more policing, or harsher penalties) months or years after enactment to improve compliance. Although a more intense enforcement is likely to have a stronger impact on Airbnb’s operations (van Holm, 2020), we were unable to reliably determine for all cities and neighborhoods when and to what extent the enforcement of the studied regulatory instruments changed. Future research could therefore make an important contribution by studying how regulatory responses can be most effectively enforced.

Lastly, the timeline of implemented regulatory instruments that we compiled includes instruments that were implemented as early as 2013 (see Table S3 in the Supplementary Material). However, we lack the corresponding quantitative Airbnb listing data for the period before mid-2015 to be able to examine the effectiveness of these early regulatory interventions. As a result, any revision of these early regulatory interventions during our observation period had to be ignored in our analysis. Thus, examining the effectiveness of early attempts to regulate STR platforms is an important complement to our analysis.

6. Conclusion

Digital platforms offer opportunities for innovative business and local policymakers, incumbents, and platform operators are striving to find a way to accommodate digital platforms in a sustainable and responsible way. To that end, the effectiveness of regulatory responses is a key piece of the puzzle. As Brian Chesky (CEO and co-founder of Airbnb) notes, “Government regulation used to be an existential crisis [for Airbnb]. We used to debate whether Airbnb will exist in the future. We now no longer debate it. We are debating ‘what is the tax rate?’, ‘what’s the nights cap?’,‘what’s the registration scheme?’ We are debating how we exist” (Swisher, 2020).

Supplementary Data

Supplementary materials are available at Industrial and Corporate Change online.

Footnotes

References