Deter and Divert: How Incumbents’ Market Structures Guide Hotel Entry Strategies

Location Choice

The mantra “location, location, location” has long underscored the critical role of geographical positioning in tourism and hospitality. Once made, location decisions are difficult to reverse, highlighting their strategic importance given the sunk costs involved (Cabral, 2012; Yang et al., 2014)

Hotel location choice research traditionally emphasizes accessibility factors (S. K. Lee & Jang, 2017), such as proximity to subway stations (Yang et al., 2012), highway exits (Mazzeo, 2002), and tourist attractions (Song & Ko, 2017). An equally important consideration is the degree of agglomeration, which reflects the density of competitors in the area. New entrants are often drawn to these clustered markets, hoping to benefit from positive externalities like demand spillovers and knowledge sharing (Bianco et al., 2025; Fang et al., 2019; Kalnins & Chung, 2004; S. K. Lee & Jang, 2015; Yang et al., 2012, 2014).

Yet, as new entrants move into these markets, incumbents inevitably face potential losses because they cannot alter their positioning ex post (O’Neill & Xiao, 2010). Indeed, a new entrant that strategically diversifies its offering or positions itself at the market’s lower end may reap spillovers and competitive advantage (Bianco et al., 2023; Canina et al., 2005; M. Kim et al., 2020). Conversely, if the newcomers’ offerings are highly similar to incumbents’, it intensifies overall market competition (Mazzeo, 2002) and places downward pressure on prices (Balaguer & Pernías, 2013). Consequently, new market entrants will likely affect incumbents’ performance, which would instead benefit from preventing new hotels from entering the market.

So far, however, current literature has not adequately examined how market structure, beyond basic agglomeration, influences location choice. While some hospitality and tourism studies do address entry barriers, such as increases in casino square footage to deter new entrants (Cookson, 2018) or the use of idle capacity in hotels (Conlin & Kadiyali, 2006; Lado-Sestayo et al., 2017), these works focus on particular tactics rather than a comprehensive view of incumbents’ organizational structure and the market power this can exert. Indeed, although agglomeration addresses the number and class of incumbent properties, the organizational structure of these incumbents may substantially shape new hotels’ entry decisions. Specifically, incumbents’ market organization can establish entry barriers, whether through deliberate defensive strategies or indirectly through competitive market dynamics, ultimately discouraging new hotels from entering those markets.

Entry Barriers

Entry barriers significantly influence market dynamics, helping to mitigate the disruptions often accompanying new market participants (Song & Ko, 2017). Such barriers are particularly prominent in industries with substantial fixed costs, sunk costs, and semi-fixed offerings (Porter, 1980). The hospitality sector exemplifies this scenario because of its high fixed costs leading to inefficiencies (Pérez-Rodríguez & Acosta-González, 2023) and the difficulty of altering key features such as class, location, or the number of rooms after market entry (O’Neill & Xiao, 2010). Moreover, the entrance of innovative startups into a historically less innovative industry intensifies incumbents’ need to strengthen their market positions (Bianco et al., 2024a).

Entry barriers can be classified as legal, technical, strategic, and brand loyalty-related types (McAfee et al., 2004). Legal barriers originate from government statutes, such as patents, licenses, permits, and regulations, that might outright prevent a firm from entering (Shapiro, 1989). Technical barriers arise from industry-specific factors such as significant start-up costs, sunk costs, economies of scale, or the presence of natural monopolies (Sutton, 1991). Strategic barriers refer to incumbents’ actions and specific investments that reshape the competitive environment, such as centralizing market control, preemptively occupying market niches, or signaling competitive intentions to possible entrants (McAfee et al., 2004). Finally, brand loyalty emerges as a formidable barrier when robust consumer allegiance to existing brands deters new competitors from attempting entry (Aaker, 2019; Katz & Shapiro, 1985).

In tourism, even though legal and technical barriers exist, they generally do not constitute insurmountable obstacles for new entrants. For instance, zoning regulations may restrict hotel development to particular areas yet rarely forbid the building of new hotels near incumbents (Suzuki, 2013; Yang et al., 2012). Likewise, while capital-intensive hotel projects involve substantial fixed and sunk costs, these tend to be manageable compared to industries such as manufacturing, where economies of scale play a larger role (Sutton, 1991). Consequently, this study focuses on strategic and brand loyalty barriers as the most relevant and influential factors shaping competition in the tourism sector.

Strategic barriers in the hotel industry can include incumbents’ investments in physical assets or manipulative service offerings intended to deter entrants (Yu & Cannella, 2007). By subtly reshaping competition, these barriers help incumbents to sustain market power and safeguard long-term investments, as they make market entry less appealing for would-be competitors (Conlin & Kadiyali, 2006). Often, such measures work because incumbents leverage their deep understanding of market dynamics to construct conditions that effectively deter new entrants.

Empirical studies within the hospitality and tourism industry underscore how firms undertake strategic asset investments or adjust service offerings to heighten entry risk and complexity for newcomers. Cookson (2018), for example, showed that casinos expand their physical capacity when threatened, signaling an oversupplied market that discourages entrants. Similarly, Conlin and Kadiyali (2006) and Lado-Sestayo et al. (2017) found that hotels created idle (unused) capacity as a deterrent measure. This idle capacity signals to potential entrants that existing firms are prepared to absorb demand fluctuations without significant price increases, reducing the profitability of new market entry. By keeping excess capacity available, incumbents create the perception that the market is already saturated and that any new entrant would struggle to secure a viable customer base.

Such types of strategic entry barriers, known as preemptive barriers, aim to secure a dominant market position via capital investments, which expand incumbents’ capacity, thus curbing demand opportunities for new entrants (Cookson, 2017). Essentially, these methods revolve around market saturation, achieved through market expansion (Conlin & Kadiyali, 2006; Cookson, 2018) or by differentiating product offerings to occupy all market segments (Batsakis et al., 2019; Porter, 1980; Ren et al., 2019).

Conversely, some entry barriers arise indirectly from incumbents’ competitive interactions. In these cases, the impediments are not deliberately created to block market entry but surface naturally from incumbents’ quest for market dominance (Koh & Rojas, 2022). A hotel owner, for instance, might expand the hotel’s portfolio within the same market to contrast with fierce rivals, boosting its own market power and, in turn, discouraging newcomers from entry.

Entry Barriers Among Hotels

Data

This study uses a large dataset comprising 784,895 monthly data points from 4,249 hotels in Texas between January 2001 and December 2023. Following previous research (Bianco et al., 2024b; Zervas et al., 2017), we integrated STR population data with records from the Texas Comptroller of Public Accounts.

STR supplied population data included publicly accessible hotel information such as geolocation (latitude and longitude), brand, parent company, opening date, and class. These were matched with financial and ownership records from the Texas Comptroller of Public Accounts, using geolocation as the primary identifier. The Comptroller’s database, recognized for tracking tax-related information on hotel establishments (Lin & Kim, 2020), contributed monthly revenue data, chain of ownership, and room counts over the observed period.

We relied on the STR population database as the primary dataset, given its extensive use in tourism research and established reliability. After cross-referencing latitude–longitude data, we used hotel names and addresses to resolve borderline cases with nearly identical coordinates. Through this matching process, we successfully linked 85.95% of STR-listed hotels to corresponding Comptroller records. In cases where a single STR hotel appeared under multiple tax records (e.g., due to ownership changes), we consolidated these entries into one record per hotel, noting each ownership change for the exact month and year.

Finally, we validated each variable to ensure it fell within expected bounds (e.g., number of rooms, monthly revenue). The room counts were always valid (i.e., no zeros or negative values), but some monthly revenue data appeared implausible—presumably from typographical mistakes. Since revenue serves solely as a cluster-level control variable rather than a primary outcome of interest, we winsorized the 5th and 95th percentile of revenue values to the average of the respective cluster. Winsorization was preferred over trimming to reduce the undue influence of extreme values while preserving the full dataset for analysis. This approach aligns with prior econometric research demonstrating that Winsorization effectively mitigates outlier impact in economic data without distorting statistical inference (Aguinis et al., 2013). Given that the current study operates at the cluster level, we subsequently aggregated data for each defined cluster.

The decision to use Texas as a sample is twofold. First, Texas has consistently served as a representative sample in various hotel-focused studies (Bianco et al., 2025; Kalnins & Chung, 2004; M. Kim et al., 2020; S. K. Lee & Jang, 2015), partly because its size and demographic diversity produce a broad assortment of hotel markets (e.g., urban, suburban, small cities etc.). Second, the public data available via the Texas Comptroller of Public Accounts is unique in the United States, providing property-level performance metrics and an ownership-change log, both essential for testing the hypotheses.

Measurements and Variables

A key aspect of this study is identifying competitive markets. The study does so by using a Hierarchical Density-Based Algorithm (HDBSCAN), which groups hotels based on their latitude and longitude (Campello et al., 2013). Previously used in hospitality research (Bianco et al., 2025), HDBSCAN permits the creation of differently-sized markets, reflecting the diverse nature of competition across the sample.

In contrast to older approaches that used MSAs (Canina et al., 2005) or fixed radii (Gan et al., 2013), HDBSCAN relies entirely on density calculations by measuring inter-point distances, more accurately mirroring real-world competition by distinguishing dense clusters from sparse noise. Concretely, the mutual distance between points i and j is defined as follows:

Where D m − k ( i , j ) is the computed distance, C o r e k ( i ) is the radius of the cluster under the minimum-point condition (k) and C o r e k ( i , j ) is the haversine distance between point i and point j. We set k = 4 to ensure that each cluster contains at least four hotels, consistent with STR reporting guidelines. Finally, we confirmed the reliability of the resulting clusters using the silhouette score (Rousseeuw, 1987) and the Hopkins statistic (Hopkins & Skellam, 1954), which had a result of 0.73 and 0.99, respectively, thereby indicating satisfactory clustering performance.

For Hypotheses 1a and 2a, which examine the effect of strategic barriers (ownership consolidation and class heterogeneity) on new market entrants, the variable New_Entry measures the monthly count of new hotels whose legal owners did not previously operate in that market. In this study, a hotel owner is defined as the legal entity that owns the property, as recorded in the Texas Comptroller of Public Accounts. Ownership is determined at the property level, meaning that new market entries are identified when a hotel is registered under a legal entity that has not previously owned a hotel in that specific market. This approach allows us to clearly distinguish entries by new market participants from actions taken by owners already operating in the market to solidify their market position, which would represent a preemptive strategy rather than a new market entry (Cookson, 2018; Lado-Sestayo et al., 2017).

Moreover, for hypotheses investigating the effectiveness of strategic and brand loyalty barriers in deterring market entry to nearby competitive markets (diversion), the dependent variables are Competitors_New_Entry, Competitors_Independent_Entry, and Competitors_Loyalty_Beachheads_Entry, which represent the average rate for market entry for nearby competitive markets. In this study, nearby markets are the competitive markets whose centroids reside within a 20 km radius from the focal market.

The 20 km radius was selected as the minimum distance previously identified for defining long-distance competitors (S. K. Lee, 2015). Additional tests were conducted with radii exceeding 20 km, and the results, which involved comparisons of AIC and BIC values, varied between the Gamma and Hurdle components of the model (Appendix A). While the Gamma component suggested that a 40 km radius would be optimal, the Hurdle part indicated that 20 km offered a better fit. Based on alignment with prior literature (S. K. Lee, 2015), the relevance of the Hurdle component to the research objective (i.e., identifying entry diversion into nearby markets), and the weighted within-model measurements (mean AIC and BIC across both components; Feng, 2021), the final decision was to use the 20 km radius.

The study focuses on four main independent variables, each tied to one of two types of entry barriers. For strategic barriers, the degree of ownership consolidation and class heterogeneity are represented by the variables Ownership_consolidation and Class_Heterogeneity, respectively. Because both measure a form of market concentration, they are operationalized using the Herfindahl-Hirschman index (HHI), a common metric for gauging concentration (Gan & Hernandez, 2013). Note that Class_Heterogeneity is inverted such that a higher value corresponds to greater heterogeneity. For loyalty barriers, the study employs Branded_Proportion and Parent_Consolidation. Branded_Proportion reflects the ratio of branded to non-branded hotels in the market, ranging from 0 to 1, while Parent_Consolidation captures the degree to which branded hotels in a market belong to a small set of loyalty programs, also measured via the HHI index.

Additionally, several control variables are included: Mean_Class, denoting the mean class in the market, Mean_RevPAR, the average Revenue Per Available Room; and, Room_Number and Hotel_Number indicating the scale of incumbent competition in terms of total rooms and total hotels, respectively.

For hypotheses investigating whether entry barriers divert hotels to neighboring markets, all dependent and independent variables are computed as the difference between the focal market’s value and the mean value of nearby competitive markets. Hence, these variables are denoted with a “Difference_” prefix. For example, if “Ownership_consolidation” in the focal market is subtracted from the level of Ownership_consolidation in neighboring clusters, the final variable is named “Difference_Ownership_consolidation.”

Modeling Entry Decision

The aim of this study is to model market-entry decisions. Thus, explanatory variables must reflect a rational agent perspective, where new entrants weight the best available opportunities, and acknowledge that the decision to enter a market is taken long before the actual market entry occurs, given the lengthy time required to build a hotel. Even if the owner acquires an existing hotel, it must still conduct feasibility analyses, negotiate, secure financing, and finalize the purchase, which takes a considerable amount of time.

Moreover, the decision to enter a market does not rely solely on the markets’ status at the time of the decision to build, but on anticipated conditions at the opening date, using available data. For instance, if an owner chooses to open a new hotel, the required feasibility necessarily accounts for current projects under construction (although not yet operational). Any discrepancy between these early considerations and the market’s actual state at opening likely arises from construction delays or similar shifts that alter the expected conditions.

To properly represent this decision-making process, the model lags independent and control variables by averaging their values over the 6 months prior to market entry. The decision on the number of months to lag those variables was based on analyzing STR pipeline data, which show an average delay of about 5.8 months between the predicted and actual market entry date. Appendix B describes additional sensitivity tests with longer lag periods (9, 12, and 15 months), confirming that the main findings remain robust.

Appendix C presents descriptive statistics. The dependent variables exhibit marked skewness, with mean values near zero and maximum values relatively small in comparison, which suggests an excess presence of zeroes. This hypothesis is supported by the calculation of skewness for positive values only (Appendix D). Further, due to the substantial size differences among variables and the strong skewness in some control variables, numeric variables not expressed as percentages have been standardized.

Model Selection

The study investigates the effectiveness of preemptive and brand loyalty entry barriers in deterring new entrants and diverting them to nearby markets. To evaluate deterrence, the study utilizes a zero-inflated Poisson regression model. To examine entry diversion into nearby markets, a Hurdle Gamma Model was selected.

Given the idiosyncrasies of markets in terms of location and time trends—which heavily influence entry decisions and can lead to non-independence of observations—both models nest the observations within unique markets using random effects, while also incorporating time fixed-effects. Nesting observations mitigates the risk of inflated Type I error rates and improves model fit, as demonstrated in Appendix E, while allowing us to control for market-level sources of heterogeneity, such as hotel mean hotel occupancy, real estate prices, or specific attractions able to attract tourists. The inclusion of time fixed-effects, meanwhile, allows the models to account for temporal patterns such as seasonality or economic downturns that occurred during the 23-year observation period.

Recognizing the count nature of the dependent variables (New_Entry, New_Independent, and Loyalty_Beachheads), a Poisson regression model was chosen for its suitability in analyzing such data. The similarity between the mean and variance of the dependent variables (Appendix C) excludes overdispersion, a common issue that might otherwise necessitate a binomial regression. Furthermore, Poisson models are widely used in market-entry studies across hospitality and tourism (Fang et al., 2019), and in general management (Gielens & Dekimpe, 2007).

However, given the high frequency of zero entries, a standard Poisson model may not fully capture the underlying distribution (Dotzel et al., 2013). To address this, we follow prior research (Alderighi & Gaggero, 2019; Falk & Hagsten, 2018; Lado-Sestayo et al., 2017) and apply a Zero-Inflated Poisson (ZIP). This model incorporates a separate process for modeling the excessive presence of zeros, ensuring greater accuracy by accounting for the dual nature of the data—instances of zero entries and instances of positive counts—which will then be modeled separately.

To assess the potential for strategic and brand loyalty barriers in focal markets to divert entries to nearby markets, the analysis compares independent and control variables of focal markets with those of nearby markets. The differences are used to evaluate how these variations influence the average rate of entry in the nearby markets. Nearby markets are defined as the mean of the three closest markets whose centroids are within a 20 km radius of the focal market.

By averaging the rate of entries across nearby markets, the dependent variables—Competitors_New_Entry, Competitors_Independent_Entry, and Competitors_Loyalty_Beachheads_Entry—are transformed. While the original count variables for focal markets are now averaged across competitors, the transformed variables retain unique characteristics. On one hand, they become continuous, presenting decimal values, and thereby preventing the use of a Poisson model (Fang et al., 2019). On the other hand, their distributions exhibit strong discrete traits, as the number of entries per month per market is typically low (Appendix F).

Consequently, following recommendations of Chakraborty and Chakravarty (2012) and Bernini and Cracolici (2015), the model selection landed on a Hurdle model with a Gamma distribution, which effectively handles excessive zeroes while also accommodating continuous data with discrete distributional properties. The use of Hurdle models has been established in tourism research (Bernini & Cracolici, 2015; Boto-García et al., 2019), as has the applications of Gamma distribution models (Lado-Sestayo et al., 2017; Santos, 2016). However, their combined use in a single model is novel in hospitality research but has been successfully applied in related social science fields (Rich et al., 2023); indeed, their combined use in a single model is novel in hospitality research but has been successfully applied in related social science fields (Kassahun et al., 2014; Molas & Lesaffre, 2010).

Shared Model Characteristics and Key Differences

Both the zero-inflated Poisson model and the Hurdle Gamma Model share a foundational structure that separately addresses zero and positive outcomes. The part of the model analyzing positive outcomes (Poisson/Gamma) is expressed as follows:

Where ( λ ( j / k ) t ) represents the expected count of entries in market j (focal market) or k (competitive market) during month t. β 0 is the model intercept, μ j denotes the random (nested) effect for market j, ϕ Y e a r ( t ) and ψ Q t r ( t ) are fixed effects for year and month, I j t l a g and C j t l a g represent independent and control variables, respectively, and ε j t λ is the residual error term for the logistic component.

The zero-inflation/hurdle part of the model addresses the probability of observing zero entries (no entry). This probability is modeled through logistic regression, which captures the likelihood of zeroes being influenced by market conditions:

Where ( π ( j / k ) t ) represents the probability of observing zero market entries for market j or k during month t and ε j t π is the error term for the zero/hurdle component.

The primary distinction between the hurdle gamma model and the zero-inflated Poisson regression lies in their treatment of zeros. The Zero-Inflated model estimates the likelihood that an observation falls into the zero category, whereas the hurdle model assesses the probability of observing a non-zero value (Feng, 2021). Consequently, their interpretations differ, as reflected in the reversed binary components of their likelihood functions. Thus, in the sections of the models that account for zero entries, the interpretation is reversed: a positive coefficient in the zero-inflation component of the ZIP model suggests an increased probability of zero market entries, while a positive coefficient in the hurdle component of the Hurdle Gamma model indicates a decreased likelihood of zero market entries.

For the zero-inflated Poisson model:

Where f p o i s s o n ( Y i | Y i > 0 , X i ) is the probability mass function of the Poisson distribution for positive integer values of Y i .

For the Hurdle model:

Where f γ ( Y i | Y i > 0 , X i ) is the density function of the Gamma distribution for positive continuous values of Y i .

Theoretical Implications

This study aimed to examine how incumbent market structures (notably ownership consolidation and class heterogeneity, along with brand loyalty) create entry barriers in the tourism industry. The findings yield several key theoretical contributions: First, by integrating strategic barriers and brand loyalty barriers, this study pushes beyond prior hospitality research that typically focuses on single-issue factors (e.g., purely capacity expansion). This study offers a multi-layered framework showing how these different forms of entry deterrence interact with location choice decisions. Hence, we advance strategic management literature in hospitality and tourism by spotlighting entry barriers specifically tailored to service-dominated, high fixed-cost markets, an area so far underexplored.

Second, while location-choice literature in hospitality often highlights agglomeration and accessibility (e.g., highways, subway stations), results indicate that the loyalty can significantly shape entrants’ decisions. By introducing the term “loyalty beachheads,” we show that branded hotels with established loyalty infrastructures, but no prior presence in a market, behave differently to both independent hotels and general new owners. This brings new insight into how existing brand-wide loyalty programs can affect which markets entrants choose and why some brand expansions succeed faster.

Third, another contribution lies in integrating attributes of neighboring markets. Typically, location-choice studies concentrate on a focal market in isolation. This study demonstrates that nearby markets inevitably influence entrants’ decisions, particularly under high ownership consolidation or brand dominance. By empirically showing how entry diversion occurs, the study confirms that incumbents’ defensive strategies reverberate beyond a single destination, and entry decisions are taken by comparing attributes from multiple markets. This expands the location-choice and entry barrier literatures by underscoring regional or multi-market perspectives on hotel expansion.

Finally, previous agglomeration research suggests that independent hotels often benefit from proximity to branded hotels via demand spillovers (Chung & Kalnins, 2001; Yang & Mao, 2017). Findings refine this view: although independent hotels may reap certain spillover benefits, they actively avoid markets where too many branded hotels co-locate. Thus, this study reconciles resource-seeking agglomeration (where independents hope to gain from brand neighbors) with the reality that excess brand presence can become a barrier, prompting them to search for alternative markets, often neighboring ones where spillover benefits are lower, but so is the competitive pressures posed by branded hotels.

Practical Implications

The results of this study also offer significant insights for practitioners. For incumbent hotels, the findings highlight how managing market structure can help maintain market share and protect competitive advantage. Strategic ownership of multiple properties within the same market, combined with diversification under a select number of major parent companies, can help slow the influx of new competitors, as results show that these strategies are able to deter high volumes of entrants.

At the same time, although offering a diverse range of classes may initially capture multiple market segments, it can also invite new entrants who perceive under-filled niches in that variety. If incumbents intend to limit competition, they should consider how many distinct classes they occupy in a given market. Overextending class diversity might create openings for specialized newcomers to differentiate or target unaddressed sub-segments. On the other hand, incumbents could intentionally maintain a more cohesive brand or class profile, thus sending fewer “signals” of open niches.

For policymakers, these results present a nuanced set of considerations. On one hand, there is a need to foster healthy competition and encourage innovation, which can lead to a more dynamic and diverse tourism sector. Lax regulations, however, may allow dominant incumbents to strengthen their position, potentially discouraging new entrants and limiting consumer choice. On the other hand, regulations that are too restrictive can lead to excessive market fragmentation, diminishing the value of local resources, discouraging significant investments, and making the market less attractive to well-capitalized businesses.

A key policy implication involves refining antitrust and fair-competition guidelines in the local tourism market. Policymakers may consider measures that prevent excessive ownership concentration or limit the influence of loyalty programs, helping to maintain a competitive environment without substantially eroding the economic benefits brought by established incumbents. Additional incentives such as tax breaks, grants, or targeted infrastructure improvements could encourage a broader range of hotel types—particularly mid-range or boutique properties—to enter the market, thus fostering greater diversity and resilience of the touristic accommodation offering.