Content uploaded by Yang Yang
All content in this area was uploaded by Yang Yang on Oct 03, 2017
Content may be subject to copyright.
Intra-metropolitan location choices for star-rated and non-rated budget hotels: The role of
1. Hao Luo, 2 Yang Yang*
1 Business School, Sun Yat-sen University,
Guangzhou, Guangdong, 510275, China.
Phone: (86) 20-8411-2561; Fax: (86) 20-8411-3687
2 School of Tourism and Hospitality Management, Temple University,
Philadelphia, PA 19122, United States
Phone: (01) 215-204-8701; Fax: (01) 215-204-8705
* Corresponding author (email@example.com)
Luo, H. and Yang, Y. (2016). Intra-metropolitan location choices for star-rated
and non-rated budget hotels: The role of agglomeration economies. International
Journal of Hospitality Management, 59: 72-83.
The authors would like to thank the anonymous reviewers for their constructive comments on
improving an early version of this paper. This project is partly supported by a research grant funded by
National Science Foundation of China (to Hao Luo) (No. 41171112).
Intra-metropolitan location choices for star-rated and non-rated budget hotels: The role of
In this study, we aim to investigate the role of agglomeration economies, which include both
urbanization economies and localization economies, in determining the hotel location choices
within a metropolitan setting. Using a sample of 110 star-rated hotels and 535 non-rated budget
hotels established in the urbanized area of Beijing over the period 2005–2012, we estimate a
mixed logit model of hotel location selection. The results suggest that both high star-rated and
non-rated budget hotels are enticed by externalities rising from localization economies to choose
locations with a high concentration of hotel incumbents. Also, an inverted-U-shaped relationship
exists between hotel incumbent concentration and location probability for both types of hotels.
Furthermore, budget hotels prefer locations that offer advantages of urbanization economies.
Lastly, our results suggest that both star-rated and budget hotels are more likely to choose
locations around Beijing Olympic Park, especially in and before 2008. We discuss several policy
implications of these findings.
Keywords: hotel location; budget hotels; chain-affiliated hotel location selection; agglomeration
economies; mixed logit model; Beijing Olympics
In the hotel industry, the importance of a well-formulated location strategy cannot be
overemphasized. A superior location is a significant competitive advantage, because it helps a
hotel attract sufficient and sustained patronage, which alleviates potential business risk and
supports a high level of profitability over the long term. From the demand side, location factors
are among the most important selection criteria (Tsai, Wu, and Zhou, 2011), and hotel guests
generally are willing to pay higher prices for more convenient locations and locations with
superior views (Rigall-I-Torrent, et al., 2011). From the supply side, location determines the
current and future profitability and productivity of hotels (Marco-Lajara, Claver-Cortés, and
Úbeda-García, 2014), and once located, hotels are likely to be influenced by both positive and
negative externalities from their neighbors (Chung and Kalnins, 2001). Hotels are extremely
expensive to build and are typically viewed as long-term investments; since relocating a hotel is
not a realistic option, a flawed location strategy generally cannot be rectified in any meaningful
way (Yang, Luo, and Law, 2014). Furthermore, an understanding of how hotels make location
choices helps government and tourism organizations better formulate strategies to attract new
hotels and develop the local tourism industry.
Hotel entrants’ location decisions are complicated by the multi-dimensional nature of locations
(Adam and Amuquandoh, 2014). As a long-term investment, a location should be evaluated
based on predictions of location-related environmental change and potential demand shifts.
Moreover, a wide variety of factors should be scrutinized for each potential site, such as
environmental, economic, cultural, political, and marketing variables (Roehl and Van Doren,
1990; Yang, et al., 2014). In most existing empirical studies, researchers investigated a single or
a very limited set of location factors, such as agglomeration (Freedman and Kosová, 2012; Yang,
Wong, and Wang, 2012), transportation networks (Li, Fang, Huang, and Goh, 2015), market
positioning (Shoval, 2006), and market entry time (Lee and Jang, 2013). Although the
importance of hotel location decisions has been recognized by members of both academia and
industry, to our knowledge only few empirical analyses have been performed under the rigorous
industrial location modeling framework with a particular focus on various factors related to
agglomeration economies covering both localization and urbanization economies (Arauzo-
Carod, Liviano-Solis, and Manjón-Antolín, 2010).
The Chinese hotel industry has attracted substantial research attention over the past decade due
to its burgeoning growth (Gross, Gao, and Huang, 2013). In China, two types of hotels co-exist:
star-rated hotels and non-rated budget hotels. National and local tourism bureaus evaluate full-
service hotels using the star-rating system, which was implemented to improve service quality
and build market confidence (Liu and Liu, 1993). Non-rated budget hotels, on the other hand,
refer to those non-rated hotels that are not administered by tourism bureaus in China, and in
general, they provide only limited services (Huang, Liu, and Hsu, 2014). One decade ago, star-
rated hotels dominated the Chinese lodging market, especially in major metropolises. However,
along with improvement in citizens’ disposable income and the growth of a burgeoning middle
class in China, the number of budget hotels has increased dramatically over the last decade to
satisfy the accommodation needs of leisure and business travelers (Chan and Ni, 2011). The
expansion of various budget hotel chains in China is also linked to entrepreneurship in the new
Chinese institutional environment (Wang and Sun, 2014). Hua, Chan, and Mao (2009) and Ren,
Qiu, Wang, and Lin (2016) found that location is one of the critical factors associated with guest
satisfaction of Chinese budget hotels. With the rapid expansion of budget hotel chains in China,
a better understanding of location selection strategy becomes important—and even imperative—
to identifying superior location sites. To our knowledge, a qualitative evaluation of location
choices of budget hotels in China has been performed in just one study (Huang, et al., 2014).
To fill the research gap in hotel location analysis, using a dataset of star-rated hotels and non-
rated budget hotels in Beijing over the period 2005–2012, we use a mixed logit model that fuses
the empirical traditions of service location with various determinants of hotel location. We
contribute to the existing literature on hotel location in at least three ways. First, unlike
researchers in previous empirical studies, we investigate the effects of agglomeration economies
by analyzing the separate effects of urbanization economies and localization economies
(Moomaw, 1988). Second, we compare the location differences between star-rated hotels and
non-rated budget hotels. Because they follow different business models and are subject to
different governmental regulations, their location choice models might also be different. Since
most budget hotels in China are operated by chains, the influence of chain-level activities must
be taken into account when modeling their location selection behavior (Yang, et al., 2014).
Lastly, we specifically look into the effect of 2008 Beijing Summer Olympics, as a type of mega-
events, on hotel location patterns. Although past literature considered the effect of mega-events
on inter-regional hotel location selection (Zhang, Guillet, and Gao, 2012), none has investigated
this effect at an intra-regional/metropolitan scale for hotel location site choice.
The paper is organized as follows. In section 2, we review previous literature on hotel location
choice. In section 3, we explain our empirical strategy for hotel specification and describe our
data sources. In section 4, we present and discuss the estimation results from the specified
econometric model. Finally, in section 5, we discuss relevant policy implications before offering
some concluding remarks.
Site characteristics and hotel location
Several theoretical hotel location models have been proposed to better understand hotel location
choices. One of the most popular models is the tourist-historic model (Ashworth and Tunbridge,
1990), which demonstrates a comprehensive typology of hotel locations within a specific
geographic area. The six types of location zones identified by this model (e.g., city gates,
transportation hubs and interchanges, transition zones) were later confirmed in different
geographies around the world (Aliagaoglu and Ugur, 2008; Rogerson, 2012; Shoval and Cohen-
Hattab, 2001). Several researchers extended the classic monocentric location model in business
geography to explain hotel location patterns using the bid-rent approach. For example, Egan and
Nield (2000) explained the spatial hierarchy of hotels in terms of distance to city center, and
Shoval (2006) identified two monocentric geographies of hotels based on individual or organized
tourism markets. Inspired by the Diamond model of competitiveness analysis, Juan and Lin
(2011) proposed a checklist for resort location covering six categories of location factors. A
panel of experts assigned the highest scores to endowment and government-related factors.
Based on a survey of hotel owners, Adam and Amuquandoh (2014) extracted six dimensions of
hotel location factors: economic, neighborhood, transportation, laws and regulations, physical
site, and socio-cultural characteristics. Yang, et al. (2012) also empirically identified several
hotel location factors, such as agglomeration effect, public service infrastructure, road
accessibility, subway accessibility and accessibility to tourism sites.
In the existing empirical literature, a large variety of site characteristics have been used to
explain various business indicators of hotels, such as hotel price (Lee and Jang, 2011),
productivity (Barros, 2005), and failure rate (Ingram and Baum, 1997). Under the general
hedonic pricing framework, several location attributes have been highlighted as having
significant hedonic value: CBD proximity, distance to airport (Lee and Jang, 2011), interstate
location (White and Mulligan, 2002), and beach accessibility (Rigall-I-Torrent, et al., 2011).
While Barros (2005) found that superior market accessibility contributes to higher hotel
efficiency, Tsang and Yip (2009) indicated that hotels close to high-end joint-venture hotels tend
to generate higher revenue per available room. Moreover, Baum and Mezias (1992) showed that
being located closer to other hotels increased a hotel’s survival rate in Manhattan.
Property characteristics and location choice
As suggested in previous studies, hotels demonstrate remarkable heterogeneity in site selection,
even when confronted with similar sets of options. For example, Yang, et al. (2012) pinpointed
location preference heterogeneity across hotels operated under different ownership types,
designated with different star ratings, established in different years, and offering different
services. Likewise, Adam and Amuquandoh (2014) found that hotel location factors are rated
with different levels of importance by owners of hotels of different classes, ownership types,
ages and sizes.
Hotels of different classes cater to different potential markets and therefore, share different
location strategies. Li, et al. (2015) found that compared to high-end hotels, low-end
establishments in Hong Kong are more likely to choose locations in affluent shopping districts.
Some studies highlighted several unique location patterns for budget hotels. Martin and Rod
(1990) highlighted a locational shift of budget hotels from near highways to downtown sites in
the United States. According to the monocentric hotel location model proposed by Egan and
Nield (2000), budget hotels choose to either locate at the edge of a city, or select converted
buildings at the edge of the city center. Huang, et al. (2014) examined the location patterns of
budget hotels in China and found that they prefer convenient metropolitan locations and typically
choose locations adjacent to upscale hotels.
With regard to hotel age, older hotels have competitive advantages associated with occupying
desirable micro-locations (Lee and Jang, 2013). Furthermore, hotels with different ownership
types have different criteria in selecting location sites. The owners of individually-owned hotels
are more focused on economic factors due to heavy concerns about the viability of their
investments (Adam and Amuquandoh, 2014). Lastly, a hotel’s scale plays an important role in
location preference because different scales represent different costs for land use. In order to
reduce the land cost, large hotels generally tend to locate far from the center of a city, as less rent
is typically required in peripheral locations (Egan and Nield, 2000).
Agglomeration economies and hotel co-location
Substantial research efforts have been directed toward hotel agglomeration in the literature.
Agglomeration refers to hotel co-location patterns (i.e., choosing locations near other hotels and
other firms/businesses) that potentially lead to competitive advantages (Alcácer and Chung,
2014). In the economics literature, Fujita, Krugman, and Venables (2001) suggested that the
spatial concentration of productive activities is largely shaped by the presence of increasing
returns and local externalities. Also, in the sociology literature, it has been argued that social
communication networks and concomitant information exchanges greatly shape the benefits
associated with agglomeration (Granovetter, 1992). On one hand, Chung and Kalnins (2001)
argued that the co-location of hotels leads to production enhancements and heightened demand.
Demand-heightened agglomeration reduces guests’ search costs through greater information
exposure to the potential market. Adam and Mensah (2014) evaluated the perceived spatial
agglomeration effects of hotel owners, and found that the availability of tourism business and the
benefits from other proximate hotels are highly rated as typical agglomeration effects. In
particular, hotels prefer locations with a high concentration of mid- and up-scale neighboring
hotels (Freedman and Kosová, 2012).
Two interaction forces, agglomeration and competition, are intertwined with hotel co-location
(Freedman and Kosová, 2012). Choosing a location with a large number of other hotels nearby
does not always lead to beneficial outcomes. For example, some empirical evidence suggests that
a high concentration of hotels leads to a heightened level of competition, which outweighs the
benefits of agglomeration; as a result, co-location results in a loss of profits (Marco-Lajara,
Claver-Cortés, and Úbeda-García, 2014). A higher level of local competition with more
neighboring hotels leads to a lower room rate (Balaguer and Pernías, 2013; Becerra, Santaló, and
Silva, 2013; Falk and Hagsten, 2015). Moreover, several scholars pinpointed the non-linear
effect of co-location. Baum and Haveman (1997) showed that entrants are enticed to locations
with more neighboring hotels when the concentration is moderate; after a certain point, an
increased number of hotels in an area tends deter new entrants. Marco-Lajara, Claver-Cortés,
Úbeda-García, and Zaragoza-Sáez (2014, 2015) discovered an inverted-U-shaped relationship
between the level of agglomeration and hotel cost: below a particular agglomeration level, co-
location boosts a hotel’s operating costs.
In analyses of agglomeration economies in the hospitality literature, the role of urbanization
economies has been largely overlooked. The benefits derived from intra-industrial clustering, as
a type of agglomeration economies, are attributed to localization economies, whereas the benefits
associated with inter-industrial clustering are associated with urbanization economies (Melo,
Graham, and Noland, 2009). Previously, researchers only focused on localization economies,
leaving urbanization economies open for investigation. According to literature in economic
geography and regional sciences, agglomeration economies are not solely driven by a single
industry that a firm falls into, but also a variety of different industries. The geographical
concentration of diversified economic activities covering a large range of sectors can result in a
snowball effect, which ultimately nurtures new entrants from forward and backward inter-
sectoral linkages (Lall, Shalizi, and Deichmann, 2004). Therefore, urbanization economies are
available to all local firms irrespective of sector (Frenken, Van Oort, and Verburg, 2007).
Moreover, these externalities arising from urbanization economies in densely populated areas
consist of pooling of labor with multiple specialization, inter-industry knowledge spillovers,
availability of general infrastructure, cross-fertilization of ideas, and cross-sectoral input sharing
(Barrios, Görg, and Strobl, 2006; Morikawa, 2011). Compared to localization economies,
urbanization economies are found to be more important for knowledge-intensive services
(Duranton and Puga, 2000).
In the context of hotel industry, externalities stemming from urbanization economies can be
particularly attractive for new hotel entrants. Zhang, Luo, Xiao, and Denizci Guillet (2013)
unveiled the positive relationship between the degree of urbanization and hotel development at
Guangdong Province, China. Walsh, Enz, and Canina (2004) found that population density is an
important determinant of lodging demand in the U.S. Recognizing the importance of regional
life, Aliagaoglu and Ugur (2008) found that a diversified business mix with bus stations and
coffee houses help explain the distribution of hotels in Erzurum, Turkey. Moreover, Rigall-I-
Torrent and Fluvià (2011) indicated that hotels charge price premiums with the presence of
public goods including cultural, sport and restaurant facilities in the justification, which are
associated with urbanization economies.
The agglomeration effect is heterogeneous to different types of new hotel entrants, and it has
been found to depend heavily on product heterogeneity between entrants and incumbents nearby.
Freedman and Kosová (2012) indicated that agglomerative effects are positive across
differentiated hotels, and hotels are more likely to co-locate with other hotels serving different
segments. Kalnins and Chung (2004) showed that economy and unbranded hotels choose to
cluster with upscale establishments, but upscale hotels locate away from other types of hotels.
This is because upscale hotels are more likely to generate spillovers to their neighbors based on
their affluent resources. Canina, Enz, and Harrison (2005) also found that lower-end hotels are
more likely to receive positive spillover effects by co-locating in a cluster with upscale hotels.
Enz, Canina, and Liu (2008) obtained similar results, and they demonstrated that low-end hotels
gain price premiums by locating close to upscale hotels. The agglomeration effect is also not
uniformly distributed across hotels with different ownership types and different sizes. For
example, Chung and Kalnins (2001) discovered that independent hotels in rural markets are more
likely to obtain agglomeration benefits from co-locating with chain-affiliated hotels. Lee and
Jang (2015) found that undifferentiated agglomeration externalities can boost hotel operating
In today’s hotel industry, many hotel properties are affiliated with international or domestic hotel
chains, and the strategic location plans of chain affiliates require tactical considerations
regarding the spatial allocation of member hotels (Yang, et al., 2014). Ingram and Baum (1997)
found that intra-brand competition within an area leads to a low hotel survival rate. Kalnins
(2004) investigated encroachment between hotel entrants and incumbents affiliated with the
same chain. The results suggest that a new franchised hotel tends to cannibalize the revenue of
same-brand hotels nearby, whereas a new company-owned hotel leads to an increase in revenue
for same-brand hotels in the vicinity. To the best of our knowledge, no researchers have explored
the sequential location choices of chain-affiliated/franchised hotels yet. Investigating this topic
will yield important insights into hotel chain expansion by revealing factors that determine the
optimal location for new chain affiliates.
Empirical hotel location models
To empirically investigate the location choice decisions of hotel entrants, two types of
econometric models can be used: the count data model (CDM) and the discrete choice model
(DCM) (Yang, et al., 2014). Both models specify location alternatives as being in certain zones.
The CDM is used to analyze which characteristics of a zone will affect hotels based on the
number of new entrants established there. Specifically in the context of hotel location analysis,
CDM specifies the number of hotels or new entrants in each zone during a given period of time
as a function of a set of explanatory variables (Holl, 2004; Ingram and Inman, 1996). In the
context of business location choice, the DCM, which is based on the economic principle of profit
maximization, is used to examine which characteristics make a zone more attractive than others
as a location for a new hotel entrant to maximize profits (Arauzo-Carod, et al., 2010; Guimarães,
Figueiredo, and Woodward, 2004). The profit that the new hotel obtains from a certain location
(akin to site attractiveness), can be specified as a function of both hotel characteristics (such as
star rating, hotel size, and hotel function) and site attributes (including accessibility,
agglomeration, and environment) (Yang, et al., 2012).
In modeling the location decisions of new firms, the DCM possesses several advantages over the
CDM. First, the DCM results can be interpreted as a structural model of location choice whereas
the CDM results can only be interpreted as a reduced-form model (Arauzo-Carod, et al., 2010).
Second, after data is aggregated for each zone, the DCM can be used to incorporate individual
characteristics of hotels, which cannot be fully considered in the CDM. Li, et al. (2015) used a
binary logit model to understand the location choices of high-end vs. low-end hotels in Hong
Kong. Yang, et al. (2012) applied an ordered logit model to understand hotel location choices in
Beijing based on proximity to the city center. Kalnins and Chung (2004, 2006) used a
conditional logit model to investigate location factors of lodging firms in Texas, and the model is
able to incorporate a large number of location options covered in the dataset over several years.
However, a major drawback associated with previous DCM applications is the property of the
independence of irrelevant alternatives (IIA), which postulates a uniform substitutability and
complementarity pattern across possible choices (Cameron and Trivedi, 2005). The IIA property
can be easily violated in spatial choice modeling when choice alternatives are characterized by
spatial continuity, aggregation and heterogeneity (Hunt, Boots, and Kanaroglou, 2004), which
leads to inconstant and problematic coefficient estimates (Sener, Pendyala, and Bhat, 2011).
Also, previous DCMs are unable to capture the nuanced heterogeneity across choice decision-
makers, suggesting that hotel location mechanism is completely identical across all decision-
makes. The absence of random taste heterogeneity may mask the variation of location preference
across decision-maker and result in misleading results (Guimarães, et al., 2004). Some more
advanced DCMs (i.e., the mixed logit model and the Bayesian conditional logit model) are able
to overcome IIA restrictions by allowing for a more flexible pattern of interdependence among
choice alternatives and a varying coefficient across decision-makers.
Data and models
Beijing, the capital city of China, was selected in this study for empirical analysis. As of 2012,
Beijing had a total of 612 star-rated hotels with an ADR of 523.2 CNY (Chinese Yuan) and 768
non-rated budget hotels with an ADR of 245.05 CNY. As a mono-centric city, Beijing has a high
density of population and high-end service business in the inner urban areas (Yang, Cai, Ottens,
and Sliuzas, 2013). Several ring roads constructed the metropolitan transportation skeleton, and
they are located successively from city center. We restricted our sample to the urbanized area of
Beijing, which is conventionally defined as four inner-urbanized districts (Dongcheng, Xicheng,
Xuanwu and Chongwen) and four outer-urbanized districts (Chaoyang, Fengtai, Haidian and
Shijingshan). The basic geographic unit in Beijing is the jiedao, and each district is comprised of
multiple jiedaos. The urbanized area is mostly within the Ring V road, and includes some
jiedaos north of it (Wu, Zhang, and Dong, 2013). In 2012, there were more than 92 jiedaos
within our research area, covering a total area of 1,180 km2. To create a reasonable choice set for
the empirical model and to avoid problems related to minor jiedao boundary changes during
research period, we decided to combine several nearby and similar jiedaos into research zones,
following the method proposed by Zheng, Peiser, and Zhang (2009), and Wu, et al. (2013). We
created a final set of 46 research zones, depicted in Figure 1. We eliminated four zones without
hotel entrance records, which is a common practice in DCM modelling of industrial location
Arauzo-Carod, et al. (2010).
We introduce two GIS based geo-spatial methods to investigate the spatial distribution of Beijing
hotels and how the distribution pattern changes over time. First, we estimate the kernel density at
each location (grid) in the research area as the average density value of hotel counts with a kernel
quartic function assigning more weights to closer hotels in ArcGIS software (Fotheringham,
Brunsdon, and Charlton, 2000; Wang, Chen, Xiu, and Zhang, 2014). In the area of hospitality
management, kernel density visualization has been used to investigate the location pattern of
convention facilities (Cong, Wu, Morrison, and Xi, 2014) and second homes (Tjørve, Flognfeldt,
and Tjørve, 2013). Second, the tool of standard deviational ellipse (SDE), covering mean center,
an average location of hotels, and the standard distance, a radius equal to one standard deviation
of the distances of all hotels from the mean center, is used to examine the distribution pattern
using ArcGIS software (Fotheringham, et al., 2000; Wang, et al., 2014). In the hospitality and
tourism literature, Wall, Dudycha, and Hutchinson (1985) used the SDE tool to examine the
spatial pattern of accommodation establishments in Toronto, and more recently, Onome Daniel
(2007) employed this tool to evaluate the distribution of tourism resources in different
The expected profit that hotel entrant i derives from choosing a location t is given as follows:
it it it it i t it
where the total profit
consists of the deterministic component,
, and the stochastic
, which is assumed to be an independently and identically distributed extreme
value. Two groups of explanatory variables are incorporated to explain the deterministic
is a set of variables measuring the attributes of location t for hotel i, and varies
across alternatives; whereas
is a set of hotel-specific variables comprised of individual
characteristics of hotel i. We assume the density for
distribution parameters (Train, 2009). Conditional on the estimated coefficients of
can calculate the probability of hotel i choosing location t as:
Pr( ) Pr( )
i it ij
it i t
ij i j
Y t U U j t
denotes the location choice of hotel entrant i, and j indexes all n location alternatives.
By allowing for unrestricted substitution patterns, the mixed logit model relaxes the assumption
of IIA embedded in the conventional DCM without random coefficients. The log likelihood (LL)
of the model is given by:
LL Y t f d
Equation (3) does not have a closed form solution, and thus can hardly be solved analytically.
Therefore, to estimate the proposed mixed logit model, we use the maximum simulated
likelihood with a simulation based on Halton sequences (Train, 2003). The simulated log
likelihood (SLL) is given by:
SLL Y t
is the rth draw from
, and r = 1, …, R. The great flexibility of the mixed logit
model comes at the cost of the computational burden associated with simulation techniques.
With advancements in information and computer technology, these techniques have become
available and have been introduced by Correia, Santos, and Barros (2007) and Masiero,
Yoonjoung Heo, and Pan (2015). This innovative method enables us to incorporate locational
heterogeneity and relax the restrictions on inter-choice substitution patterns better than
conventional discrete choice models (Kalnins and Chung, 2004, 2006). To the best of
knowledge, our study represents the very first research effort in which the mixed logit model is
applied to understand hotels’ location decisions.
The dependent variable in the mixed logit model is a set of binary variables, indicating whether
the hotel entrant chose to locate in a specific zone within the study area. This variable is assigned
a value of 1 if a hotel entrant chose the zone, and 0 otherwise. Therefore, for each single hotel
entrant, there is a set of dummy variables where only a single variable is assigned a value of 1.
We specify the following zone-specific variables as independent variables in the proposed
subway: the presence of subway stations in each zone; subway = 1 for the zone with
subway stations, and subway = 0 otherwise. During the study period, Beijing’s subway
system was expanded farther into outlying areas, facilitating access to the city center. As
suggested by Li, et al. (2015), accessibility to subway stations plays an important role in
shaping hotel location patterns.
Olympics: Beijing hosted the 2008 Summer Olympic Games, and Beijing Olympic Park
was constructed on the border between the Haidian and Chaoyang districts. As the
Olympic core district, major stadiums (including Beijing National Stadium and National
Aquatics Center) and National Convention Center were built in the park. To capture the
influence of Beijing Olympic, Olympics = 1 for the zone where Beijing Olympic Park is
located, and Olympics = 0 otherwise.
college_stud: the density of college student population in each zone, measured by the
number of registered college students (in 10,000) including both undergraduates and
graduates per square kilometer. We collected the data from the official website of each
university and college located in the study area. According to Broadway (1993), hotels
tend to co-locate with university campuses.
urbanization: log of the population density in each zone, measured by the number of
registered residents (in 10,000) per square kilometer. We obtained these data from the
National Population Census (2000 and 2010). This variable has been widely used as a
proxy for the advantages of urbanization economies (Frenken, et al., 2007; Melo, et al.,
2009). In a centralized planning economy like China, the city government plays a
dominate role in allocating economic and social infrastructure based on population
concentration. Also, a high population density leads to a high concentration of businesses
that are directly or indirectly related to the lodging industry, such as retailing, transport,
and restaurants (Lall, et al., 2004).
localization: the number of star-rated and non-rated budget hotels in each zone. It is used
to capture localization economies. We obtained a list of star-rated hotels in Beijing and
their characteristics from the Beijing Tourism Bureau and a list of budget hotels in
Beijing and their characteristics from a leading budget hotel consulting website,
www.hostelcn.com, as well as the official websites of major Chinese budget hotel chains.
This variable captures within-sector clustering aspects of localization economies.
same_brand: the number of same-brand budget hotels in each zone, used to capture the
chain-affiliated hotel location behavior. Since more than 70% of star-rated hotels are not
affiliated with any hotel chains in our data set, we incorporate this variable only in the
location choice model of non-rated budget hotels.
We also incorporate localization_sq, the square term of localization, in some models to capture
the quadratic effect of localization economies as a type of agglomeration economies (Baum and
Haveman, 1997). Note that in the empirical model, since the zonal characteristics were observed
at the time of the location decision rather than at the time of the hotel opening, we use 1-year
lagged values for all independent variables of zonal characteristics except sub. In general,
members of the general public are informed about construction plans for new subway systems in
advance. The use of a lagged variable in the location choice model also alleviates the potential
reverse causality problem (Freedman and Kosová, 2012). Apart from variables explaining the
zonal characteristics, we also consider the hotel-specific variable, star, which measures the star
rating of a star-rated hotel. To keep the empirical model parsimonious, we enter the hotel-
specific variable as interaction terms with other site characteristic variables (
in Equation 1).
Figure 1 demonstrates the spatial pattern of star-rated and budget hotels in Beijing as of 2012. A
general pattern indicates that most hotels chose to locate around the city center, which is
consistent with the prediction from the monocentric hotel location model (Yokeno, 1968). The
map does not present any obvious location pattern differences between star-rated and non-rated
(Please insert Figure 1 about here)
Table 1 presents the summary of hotel properties covered in our data set. A total of 110 star-rated
hotels and 535 budget hotels were established in Beijing from 2005 to 2012. Among these,
77.27% of the star-rated hotels and 46.35% of the budget hotels were open before 2008, the year
when Beijing hosted the Summer Olympic Games. We find that, over the years, fewer star-rated
hotels entered the market, and budget hotels comprised the vast majority of new hotel entrants.
Apart from a single one-star hotel, star-rated hotels were relatively evenly distributed. Table 2
further presents the descriptive statistics of zone-specific variables across the 42 zones in the
urbanized area of Beijing from 2005–2012. Table 3 presents the collinearity diagnostics of
independent variables in two samples. VIF measures are all variables are way lower than the
cutoff value 10, and their tolerance values are higher than 0.2, suggesting the absence of multi-
collinearity issue (Dormann, et al., 2013). Also, the condition numbers in two samples are
significantly lower than the threshold value, which is 30 (Dormann, et al., 2013). The correlation
coefficient tables of independent variables, which are available upon request, suggest that most
coefficients are below 0.50, suggesting that correlations are low to moderate among explanatory
variables. To guarantee the robustness and reliability of results, we sequentially include
additional independent variables to a baseline model (Greene, 2007).
(Please insert Table 1 about here)
(Please insert Table 2 about here)
(Please insert Table 3 about here)
Figure 2 presents the kernel density maps of new hotel entrants in Beijing during the research
period. In Figure 2 (a), two clusters are identified for newly established star-rated hotels, and
they are the city center cluster covering Fangfujing-Jianguomen area and central business district
(CBD), and the Olympic park in the north to city center. In Figure 2 (b), we identify two clusters
for new budget hotels. The larger one in the city center is almost the same as the cluster found
for star-rated hotels, and the other is located in the northwest to the city center (Zhongguancun
area), which is a high-tech zone called the Chinese Silicon Valley.
Figure 3 provides the maps of SDEs for all hotels as of 2005 and 2012. In Figure 3 (a), the SDEs
of star-rated hotels in 2005 and 20012 are similar, suggesting that the location pattern of star-
rated hotels varied little during the research period. The two SDEs are mainly elongated along
the west-east direction with the center in the city center. In Figure 3 (b), the SDE of budget
hotels in 2005 is very similar to its counterpart of star-rated hotels in terms of orientation, shape
and location. However, the SDE in 2012 changed significantly; its scope shrank without a shift
of center, indicating that budget hotels tended to locate in a more concentrated way around the
city center over the research period.
Results for star-rated hotels
Table 4 presents the mixed logit model estimation results for star-rated hotels. Model 1 includes
all independent variables previously proposed for all star-rated hotels in the sample, and is
regarded as the baseline model. Among five variables, Olympics and localization are estimated
to be significant. A positive coefficient of Olympics suggests that star-rated hotels are more
likely to choose to locate in the zone where Beijing Olympic Park is located, whereas a positive
coefficient of localization indicates that the seek for the advantage of localization economies, a
type of agglomeration economies, drives the location choice decision of star-rated hotels. For
other independent variables, subway, college_stud, and urbanization, are estimated to be
Models 2 and 3 were estimated based on the sub-sample between 2005 and 2008 and that
between 2009 and 2012, respectively. In Model 2, a significant and positive coefficient of
subway suggests that star-rated hotels preferred locations with easy access to public
transportation in 2005-2008. However, according to the results in Model 3, this coefficient
becomes negative and significant in 2009-2012. One potential reason is that the Beijing subway
system expanded rapidly during the research period, and most zones had easy access to subway
lines in and before 2008. Additional expansion of subway network did not help attract more
budget hotels after 2008. Another noticeable difference between the estimates of these two
models is the different estimated coefficient of Olympics, and it is estimated to be positive and
significant in and before the 2008 Beijing Olympics (Model 2) but negative and significant after
that (Model 3). This result indicates that the location around Beijing Olympic Park was greatly
attractive for new star-rated hotel entrants before the Olympic Games, but after the Games, new
entrants avoided that location. This result can be explained by the facts that after 2008, even
though some events were hold using facilities constructed for Olympic Games, they were hardly
generate sufficient lodging demand; therefore, the overall supply in the area close to Beijing
Olympic Park tended to substantially lower the performance of hotels and impeded new hotels
from entering that area. Another variable, localization, becomes statistically insignificant albeit
positive after 2008 (Model 3). This result can be partly explained by our previous findings of the
crowding-out effect around Beijing Olympic Park, which tends to offset the agglomeration
benefits in other locations, leading to an insignificant effect of agglomeration economies overall.
Lastly, in Model 3, we find that urbanization is estimated to be moderately significant,
suggesting that the advantage of urbanization economies became a location determinant when
star-rated hotel choose location sites after 2008.
In Model 4, we introduced interaction terms between star and two agglomeration-related
variables. However, both interaction terms, urbanization*star and localization*star, are
estimated to be statistically insignificant. Moreover, as suggested by Marco-Lajara, et al. (2015),
the agglomeration effect from nearby peers can be non-linear and quadratic. Therefore, we added
the quadratic term of localization, localization_sq, into Model 5. This quadratic term is estimated
to be negative and moderately significant at the 0.10 significance level, suggesting that the effect
of localization economies on hotel location probability is inverted-U-shaped. In Model 5,
Olympics is estimated to be insignificant, and this result might be caused by multi-collinearity
problems. Therefore, we re-run the model by excluding Olympics in Model 6. The signs,
magnitudes, and significance levels of remaining variables vary little, suggesting the absence of
(Please insert Table 4 about here)
In Model 7, we further included the interaction terms between star and the quadratic term.
Among these variables, localization*star is estimated to be moderately significant whereas
localization_sq*star is highly significant. The results show that the concentration of nearby
hotels has a quadratic effect on the location choices of hotel entrants, and this quadratic effect
varies across hotels with different star ratings. Figure 4 plots the relationship between
localization and hotel location probability after taking quadratic and interaction term with star
into account. For one-star hotel entrants, a consistent positive relationship is found between
location probability and incumbent hotel concentration of the location. For two- to five-star
hotels, this relationship is inverted-U-shaped. When the location is concentrated with a moderate
number of neighboring hotels, an increase in this number is associated with an increase in the
location selection probability for hotel entrants. However, after a certain point (the turning
point), the existence of more hotels nearby leads to a lower probability of that location being
selected. It is interesting to note that hotel entrants with a higher star rating have a lower turning
point, suggesting that high-end hotels are more vulnerable in presence of more neighboring
hotels, and these neighbors tend to erode the operating performance of high-end ones (Kalnins
and Chung, 2004). In Model 8, we re-run the model by excluding Olympics, and the estimation
results vary little, suggesting the robustness of our results.
(Please insert Figure 4 about here)
The random effects (standard deviations) of all coefficients are estimated to be insignificant in
Table 4, suggesting that there is little variation in the effects of independent variables across
observations. Thus, the results demonstrate no unspecified heterogeneity in preferences related to
hotel location choices for different zones and across different investors.
Results for non-rated budget hotels
In Table 5, we present the estimation results of the mixed logit model for non-rated budget
hotels. Model 9 incorporates five independent variables without any interaction and quadratic
terms as the baseline model for budget hotels. Similar to the results from Model 1, Olympics is
estimated to be positive and significant for budget hotels. However, its coefficient is only half of
its counterpart in Model 1. One possible reason for this result is that budget hotels, which have
less affordability for high land rent compared to star-rated ones (Egan and Nield, 2000), are less
likely to enter the area around the Olympic areas because of the high entry land cost. Also, a
large number of budget hotels did not get the permit from the government to serve international
guests, and therefore, location site very proximate to Olympic facilities seem less attractive to
new budget hotel entrants. Furthermore, both subway and college_stud are found to be
statistically insignificant. Regarding agglomeration-related variables, unlike in the star-rated
hotel location model, urbanization, which captures the benefits from urbanization economies, is
estimated to be positive and significant in the budget hotel location model. This result indicates
that the advantages of urbanization economies are highly attractive to new budget hotel entrants.
The affluent supply of public goods that urbanization economies bring in can be particularly
important for guests of budget hotels with a medium or limited budget. Moreover, localization is
estimated to be positive and significant, suggesting that budget hotels also prefer co-locating
with other hotels like their star-rated peers. The magnitude of this coefficient is very close to its
counterpart in Model 1.
We also split the non-rated budget hotel sample into two sub-samples, hotels established in
2005-2008, and those opened in 2009-2012. Models 10 and 11 presents the estimated hotel
location models for these two sub-samples, respectively. The results suggest that the
attractiveness of locations close to Beijing Olympic Park (Olympics) disappeared after 2008; but
unlike the results from star-rated hotels (Model 3), budget hotels did not avoid these locations in
the post-Olympics era. Regarding agglomeration-related variables, we find that the magnitudes
of coefficients of both urbanization and localization get smaller in the post-Olympics era for
budget hotels (Model 11). This result can be attributed to the decreasing benefits from
agglomeration economies over the scale of agglomeration (Borowiecki, 2015), which will be
further investigated later in this section.
(Please insert Table 5 about here)
Model 12 includes two additional variables to examine the location choices of chain-affiliated
budget hotels. We introduce same_brand and its squared term same_brand_sq to capture the
potential non-linear effect of the number of same-brand hotels on the location choices of budget
hotels. However, neither of them is estimated to be statistically significant, and a joint test on
these two also shows that they are jointly insignificant. In Model 13, we added an additional
quadratic terms localization_sq to Model 9. Both localization_sq and localization are estimated
to be statistically significant, highlighting a quadratic relationship between the location
probability of a budget hotel and the hotel concentration of that location. Based on the estimated
coefficient, the effect of localization economies (localization) becomes negative for zones with
more than 38 (0.0829/(-2*(-0.00109))) hotels within the zone. Figure 5 depicts this relationship.
Similar to the results presented in Figure 4, this relationship is inverted-U-shaped. At low values
of hotel concentration, locations with large concentrations of hotels are more attractive to budget
hotel entrants, whereas at high values, these locations are avoided by budget hotel entrants. With
regard to the estimated random effects (standard deviations), most are statistically insignificant.
Lastly, since the change of significance level of Olympics in Model 13 might be cause by multi-
collinearity, we run Model 14 by excluding this variable. The results highlight the robustness of
estimates of remaining variables.
(Please insert Figure 5 about here)
Due to the relatively high correlation across two agglomeration-related independent variables
shown in Table 3, we estimated a series of location models by including only one out of two at a
time based on the specification of baseline models. Major conclusions on significances and signs
of estimated coefficients varied little. We further checked the robustness of our results by
utilizing a set of different econometric models. In particular, we used the conditional logit model
(CLM) and nested logit model (NLM) to re-run the model specifications in Tables 4 and 5. The
CLM assumes no random effect compared to the mixed logit model specification, whereas the
NLM alleviates the IIA assumption by grouping certain alternatives (zones) into a nest and
assuming that within-nest alternatives are substitutable and between-nest alternatives are
independent (Train, 2009). In the NLM, we specified the nests of location alternatives based on
zones’ relative locations to the four major ring roads in Beijing: within Ring II, between Rings II
and III, between Rings III and IV, between Rings IV and V, and outside Ring V. The results
from the CLM and NLM are consistent with the findings from the mixed logit model presented
in Tables 4 and 5 in terms of coefficient magnitude, sign, and significance. Therefore, the
robustness of estimation results is confirmed. The complete estimation results from the
robustness check are available upon request.
In this paper, we investigated the intra-metropolitan location patterns of star-rated and non-rated
budget hotels in Beijing. This study introduced the concept of urbanization economies, as a type
of agglomeration economies stemming from inter-sectional business diversity, in evaluating
hotel location choice, and empirically confirmed the importance of urbanization economies in
attracting budget hotel entrants but not star-rated hotels. The location sites associated with
advantages of urbanization economies are assume to be endowed with more developed public
infrastructure and higher demand for business travelers. Regarding localization economies from
agglomeration of hotels, our results found that, for both star-rated and budget hotels, an inverted-
U-shaped relationship exists between the hotel concentration and the probability of selecting that
hotel location. Our results corroborate the previous results from Baum and Haveman (1997), and
competition outweighs agglomeration in a presence of large number of peers in the
neighborhood. However, the result partly contradicts the findings from Marco-Lajara, et al.
(2015) of a U-shaped relationship between agglomeration level and hotel performance. One
possible reason is that a constant negative relationship between agglomeration and hotel income
may not hold due to demand spillovers across hotel properties.
We also unveiled the impact of 2008 Beijing Olympics on the location pattern of hotels under the
rigorous econometric modelling framework. Star-rated hotels serving a relatively higher-end
market and with a higher land rent affordability were found to be more likely to locate nearby the
Olympic core district, which attracted a significant proportion of new star-rated and budget
hotels in and before 2008. However, the results suggested the overall supply of rooms deterred
new star-rated hotels from entering this area after the Olympic Games. This result is consistent
with Ferreira and Boshoff (2014)’s findings on the detrimental effect of hotel over-supply on
hotel performance after 2010 FIFA Soccer World Cup in South Africa.
Our results have several important implications for hotel investors and urban planners. First,
when new hotel entrants choose locations in metropolitan settings, they consider not only gains
associated with co-location, but also possible costs associated with competition across nearby
incumbents. When choosing to co-locate with other hotels, investors should evaluate the profiles
of incumbents. Low-end hotels should be avoided because according to our results, they tend to
generate negative externalities for star-rated hotels. Second, since an inverted-U-shaped
relationship exists between star-rated hotel density and hotel location probability, investors
should deliberately analyze these factors to determine the location sites with the optimal level of
hotel concentrations (Figures 4 and 5) that can provide the highest level of agglomeration
benefits. Third, the dynamics of urban economies may be favorable for budget hotels; when a
specific industrial zone or business district is proposed, investors should investigate the
opportunity of locating budget hotels within the zone/district to take advantage of urbanization
economies in these areas. Lastly, the market entry and location decisions into mega-event zones
should be very strategic and both short-run and long-run evaluations should be carried out.
According to our results, the core area hosting 2008 Beijing Olympics attracted a large number
of new hotel entrants before 2008, but this area became abandoned by hotel investors, especially
for star-rated hotels, due to an oversupply of hotel rooms and a lack of lodging demand in the
Several limitations may hinder the applicability of our findings. First, we did not take ownership
factors into consideration due to data unavailability. Hotel ownership type tends to shape a
hotel’s ability to pay the rent for a location and its capability to generate positive externalities to
proximate hotels. Second, due to the inherent limitation of DCMs, zones without hotel entry
records have to be eliminated, and this may lead to a loss of useful information. Third, we did
not consider any performance measures in the empirical model. In a typical hotel location choice
process, an entrant may have access to some historical performance data of incumbents, which
may help them identify optimal location sites. Therefore, in future studies we recommend
incorporating ownership and performance measures into a more integrated hotel location model,
investigating the crowding out effect associated with agglomeration, and using data from other
countries in order to compare the results against our findings.
Adam, I., Amuquandoh, F. E., 2014. Hotel characteristics and location decisions in Kumasi
Metropolis, Ghana. Tourism Geographies 16 (4), 653-668.
Adam, I., Mensah, E. A., 2014. Perceived spatial agglomeration effects and hotel location
choice. Anatolia 25 (1), 49-60.
Alcácer, J., Chung, W., 2014. Location strategies for agglomeration economies. Strategic
Management Journal 35 (12), 1749-1761.
Aliagaoglu, A., Ugur, A., 2008. Hotels as a model of regional life: The Erzurum sample.
European Planning Studies 16 (10), 1405-1422.
Arauzo-Carod, J.-M., Liviano-Solis, D., Manjón-Antolín, M., 2010. Empirical studies in
indutrial location: An assessment of their methods and results. Journal of Regional
Science 50 (3), 685-711.
Ashworth, G. J., Tunbridge, J. E., 1990. The Tourist-historic City. London: Belhaven Press.
Balaguer, J., Pernías, J. C., 2013. Relationship between spatial agglomeration and hotel prices.
Evidence from business and tourism consumers. Tourism Management 36 (391-400.
Barrios, S., Görg, H., Strobl, E., 2006. Multinationals' location choice, agglomeration economies,
and public incentives. International Regional Science Review 29 (1), 81-107.
Barros, C. P., 2005. Measuring efficiency in the hotel sector. Annals of Tourism Research 32 (2),
Baum, J. A. C., Haveman, H. A., 1997. Love thy neighbor? Differentiation and agglomeration in
the Manhattan hotel industry, 1898-1990. Administrative Science Quarterly 42 (2), 304-
Baum, J. A. C., Mezias, S. J., 1992. Localized competition and organizational failure in the
Manhattan hotel industry, 1898-1990. Administrative Science Quarterly 37 (4), 580-604.
Becerra, M., Santaló, J., Silva, R., 2013. Being better vs. being different: Differentiation,
competition, and pricing strategies in the Spanish hotel industry. Tourism Management
Borowiecki, K. J., 2015. Agglomeration economies in classical music. Papers in Regional
Science 94 (3), 443-468.
Broadway, M. J., 1993. Montreal's changing tourist landscape. Canadian Journal of Urban
Research 2 (1), 30-48.
Cameron, A. C., Trivedi, P. K., 2005. Microeconometrics: Methods and Applications.
Cambridge: Cambridge University Press.
Canina, L., Enz, C. A., Harrison, J. S., 2005. Agglomeration effects and strategic orientations:
Evidence from the U.S. lodging industry. Academy of Management Journal 48 (4), 565-
Chan, W. W., Ni, S., 2011. Growth of budget hotels in China: Antecedents and future. Asia
Pacific Journal of Tourism Research 16 (3), 249-262.
Chung, W., Kalnins, A., 2001. Agglomeration effects and performance: A test of the Texas
lodging industry. Strategic Management Journal 22 (10), 969-988.
Cong, L., Wu, B., Morrison, A. M., Xi, K., 2014. The spatial distribution and clustering of
convention facilities in Beijing, China. Asia Pacific Journal of Tourism Research 19 (9),
Correia, A., Santos, C. M., Barros, C. P., 2007. Tourism in Latin America A Choice Analysis.
Annals of Tourism Research 34 (3), 610-629.
Dormann, C. F., Elith, J., Bacher, S., Buchmann, C., Carl, G., Carré, G., Marquéz, J. R. G.,
Gruber, B., Lafourcade, B., Leitão, P. J., 2013. Collinearity: A review of methods to deal
with it and a simulation study evaluating their performance. Ecography 36 (1), 27-46.
Duranton, G., Puga, D., 2000. Diversity and specialisation in cities: Why, where and when does
it matter? Urban Studies 37 (3), 533-555.
Egan, D. J., Nield, K., 2000. Towards a theory of intraurban hotel location. Urban Studies 37 (3),
Enz, C. A., Canina, L., Liu, Z., 2008. Competitive dynamics and pricing behavior in US hotels:
The role of co-location. Scandinavian Journal of Hospitality and Tourism 8 (3), 230-250.
Falk, M., Hagsten, E., 2015. Modelling growth and revenue for Swedish hotel establishments.
International Journal of Hospitality Management 45 (59-68.
Ferreira, S. L., Boshoff, A., 2014. Post-2010 FIFA Soccer World Cup: oversupply and location
of luxury hotel rooms in Cape Town. Current Issues in Tourism 17 (2), 180-198.
Fotheringham, A. S., Brunsdon, C., Charlton, M., 2000. Quantitative Geography: Perspectives on
Spatial Data Analysis. London: Sage.
Freedman, M. L., Kosová, R., 2012. Agglomeration, product heterogeneity and firm entry.
Journal of Economic Geography 12 (3), 601-626.
Frenken, K., Van Oort, F., Verburg, T., 2007. Related variety, unrelated variety and regional
economic growth. Regional Studies 41 (5), 685-697.
Fujita, M., Krugman, P. R., Venables, A., 2001. The Spatial Economy: Cities, Regions, and
International Trade. Cambridge: MIT press.
Granovetter, M., 1992. Problems of explanation in economic sociology. In N. Nohria & R.
Eccles (Eds.), Networks and organizations: Structure, form, and action (pp. 25-56).
Cambridge, MA: Boston: Harvard Business School Press.
Greene, W. H., 2007. Econometric Analysis (6th ed.). Upper Saddle River, NJ: Prentice Hall.
Gross, M. J., Gao, H., Huang, S. S., 2013. China hotel research: A systematic review of the
English language academic literature. Tourism Management Perspectives 6 (68-78.
Guimarães, P., Figueiredo, O., Woodward, D., 2004. Industrial location modeling: Extending the
random utility framework. Journal of Regional Science 44 (1), 1-20.
Holl, A., 2004. Transport infrastructure, agglomeration economies, and firm birth: Empirical
evidence from Portugal. Journal of Regional Science 44 (4), 693-712.
Hua, W., Chan, A., Mao, Z., 2009. Critical success factors and customer expectation in budget
hotel segment—A case study of China. Journal of Quality Assurance in Hospitality &
Tourism 10 (1), 59-74.
Huang, S., Liu, Z., Hsu, C. H. C., 2014. Customer experiences with economy hotels in China:
Evidence from mystery guests. Journal of Hospitality Marketing & Management 23 (3),
Hunt, L. M., Boots, B., Kanaroglou, P. S., 2004. Spatial choice modelling: New opportunities to
incorporate space into substitution patterns. Progress in Human Geography 28 (6), 746-
Ingram, P., Baum, J. A. C., 1997. Chain affiliation and the failure of Manhattan hotels, 1898-
1980. Administrative Science Quarterly 42 (1), 68-102.
Ingram, P., Inman, C., 1996. Institutions, intergroup competition, and the evolution of hotel
populations around Niagara falls. Administrative Science Quarterly 41 (4), 629-658.
Juan, P.-J., Lin, S.-Y., 2011. Resort hotel location. Tourism Economics 17 (4), 925-931.
Kalnins, A., 2004. An empirical analysis of territorial encroachment within franchised and
company-owned branded chains. Marketing Science 23 (4), 476-489.
Kalnins, A., Chung, W., 2004. Resource-seeking agglomeration: A study of market entry in the
lodging industry. Strategic Management Journal 25 (7), 689-699.
Kalnins, A., Chung, W., 2006. Social capital, geography, and survival: Gujarati immigrant
entrepreneurs in the U.S. lodging industry. Management Science 52 (2), 233-247.
Lall, S. V., Shalizi, Z., Deichmann, U., 2004. Agglomeration economies and productivity in
Indian industry. Journal of Development Economics 73 (2), 643-673.
Lee, S. K., Jang, S., 2011. Room rates of U.S. airport hotels: Examining the dual effects of
proximities. Journal of Travel Research 50 (2), 186-197.
Lee, S. K., Jang, S., 2013. Early mover or late mover advantage for hotels? Journal of Hospitality
& Tourism Research.
Lee, S. K., Jang, S., 2015. Conditional agglomeration externalities in lodging markets. Journal of
Hospitality & Tourism Research 39 (4), 540-559.
Li, M., Fang, L., Huang, X., Goh, C., 2015. A spatial–temporal analysis of hotels in urban
tourism destination. International Journal of Hospitality Management 45 (0), 34-43.
Liu, Z.-Q., Liu, J. C., 1993. Assessment of the hotel rating system in China. Tourism
Management 14 (6), 440-452.
Marco-Lajara, B., Claver-Cortés, E., Úbeda-García, M., 2014. Business agglomeration in tourist
districts and hotel performance. International Journal of Contemporary Hospitality
Management 26 (8), 1312-1340.
Marco-Lajara, B., Claver-Cortés, E., Úbeda-García, M., Zaragoza-Sáez, P. D. C., 2014. Hotel
performance and agglomeration of tourist districts. Regional Studies 1-20.
Marco-Lajara, B., Claver-Cortés, E., Úbeda-García, M., Zaragoza-Sáez, P. d. C., 2015. A
dynamic analysis of the agglomeration and performance relationship. Journal of Business
Martin, S., Rod, M., 1990. Competitive strategies in the budget hotel sector. International Journal
of Contemporary Hospitality Management 2 (3), 3-9.
Masiero, L., Yoonjoung Heo, C., Pan, B., 2015. Determining guests’ willingness to pay for hotel
room attributes with a discrete choice model. International Journal of Hospitality
Management 49 (117-124.
Melo, P. C., Graham, D. J., Noland, R. B., 2009. A meta-analysis of estimates of urban
agglomeration economies. Regional Science and Urban Economics 39 (3), 332-342.
Moomaw, R. L., 1988. Agglomeration economies: Localization or urbanization? Urban Studies
25 (2), 150-161.
Morikawa, M., 2011. Economies of density and productivity in service industries: An analysis of
personal service industries based on establishment-level data. Review of Economics and
Statistics 93 (1), 179-192.
Onome Daniel, A., 2007. Evaluating tourism resource areas in Nigeria for development. Tourism
Review 62 (2), 37-45.
Ren, L., Qiu, H., Wang, P., Lin, P. M., 2016. Exploring customer experience with budget hotels:
Dimensionality and satisfaction. International Journal of Hospitality Management 52 (13-
Rigall-I-Torrent, R., Fluvià, M., 2011. Managing tourism products and destinations embedding
public good components: A hedonic approach. Tourism Management 32 (2), 244-255.
Rigall-I-Torrent, R., Fluvià, M., Ballester, R., Saló, A., Ariza, E., Espinet, J.-M., 2011. The
effects of beach characteristics and location with respect to hotel prices. Tourism
Management 32 (5), 1150-1158.
Roehl, W. S., Van Doren, C. S., 1990. Locational characteristics of American resort hotels.
Journal of Cultural Geography 11 (1), 71-83.
Rogerson, J. M., 2012. The changing location of hotels in South Africa’s coastal cities, 1990–
2010. Urban Forum 23 (1), 73-91.
Sener, I. N., Pendyala, R. M., Bhat, C. R., 2011. Accommodating spatial correlation across
choice alternatives in discrete choice models: An application to modeling residential
location choice behavior. Journal of Transport Geography 19 (2), 294-303.
Shoval, N., 2006. The geography of hotels in cities: An empirical validation of a forgotten
model. Tourism Geographies 8 (1), 56-75.
Shoval, N., Cohen-Hattab, K., 2001. Urban hotel development patterns in the face of political
shifts. Annals of Tourism Research 28 (4), 908-925.
Tjørve, E., Flognfeldt, T., Tjørve, K. M. C., 2013. The effects of distance and belonging on
second-home markets. Tourism Geographies 15 (2), 268-291.
Train, K., 2003. Discrete Choice Methods with Simulation. Cambridge: Cambridge University
Train, K. E., 2009. Discrete Choice Methods with Simulation. Cambridge: Cambridge University
Tsai, H., Wu, J., Zhou, Z., 2011. Managing efficiency in international tourist hotels in Taipei
using a DEA model with non-discretionary inputs. Asia Pacific Journal of Tourism
Research 16 (4), 417-432.
Tsang, E. W. K., Yip, P. S. L., 2009. Competition, agglomeration, and performance of Beijing
hotels. The Service Industries Journal 29 (2), 155-171.
Wall, G., Dudycha, D., Hutchinson, J., 1985. Point pattern analyses of accommodation in
Toronto. Annals of Tourism Research 12 (4), 603-618.
Walsh, K., Enz, C. A., Canina, L., 2004. The impact of gasoline price fluctuations on lodging
demand for US brand hotels. International Journal of Hospitality Management 23 (5),
Wang, F., Chen, C., Xiu, C., Zhang, P., 2014. Location analysis of retail stores in Changchun,
China: A street centrality perspective. Cities 41, Part A (54-63.
Wang, S., Sun, X., 2014. Economy Hotels in China: A Glocalized Innovative Hospitality Sector.
New York, NY: Routledge.
White, P. J., Mulligan, G. F., 2002. Hedonic estimates of lodging rates in the Four Corners
Region. The Professional Geographer 54 (4), 533-543.
Wu, W., Zhang, W., Dong, G., 2013. Determinant of residential location choice in a transitional
housing market: Evidence based on micro survey from Beijing. Habitat International 39
Yang, Y., Luo, H., Law, R., 2014. Theoretical, empirical, and operational models in hotel
location research. International Journal of Hospitality Management 36 (209–220.
Yang, Y., Wong, K. K. F., Wang, T., 2012. How do hotels choose their location? Evidence from
hotels in Beijing. International Journal of Hospitality Management 31 (3), 675-685.
Yang, Z., Cai, J., Ottens, H. F. L., Sliuzas, R., 2013. Beijing. Cities 31 (491-506.
Yokeno, N., 1968. La localisation de l' industrie touristique: application de l’analyse de Thunen-
Weber. In Cahiers du Tourisme. Aix-en-Provence: C.H.E.T.
Zhang, H. Q., Guillet, B. D., Gao, W., 2012. What determines multinational hotel groups’
locational investment choice in China? International Journal of Hospitality Management
31 (2), 350-359.
Zhang, H. Q., Luo, J. M., Xiao, Q., Denizci Guillet, B., 2013. The impact of urbanization on
hotel development: Evidence from Guangdong Province in China. International Journal
of Hospitality Management 34 (92-98.
Zheng, S., Peiser, R. B., Zhang, W., 2009. The rise of external economies in Beijing: Evidence
from intra-urban wage variation. Regional Science and Urban Economics 39 (4), 449-
Table 1. Descriptive statistics of hotel characteristics
Non-rated budget hotels
year of opening
Table 2. Descriptive statistics of zonal characteristics
Table 3. Collinearity diagnostics of independent variables
Non-rated budget hotels
Table 4. Mixed logit estimation results for star-rated hotels
Random coefficients (SD)
Number of hotels
Number of zones
Notes: *** indicates significance at the 0.01 level, ** indicates significance at the 0.05 level, * indicates significance at the 0.1 level. Robust
standard errors are presented in parentheses. † indicates the sample on and before 2008, and †† indicates the sample after 2008.
Table 5. Mixed logit estimation results for non-rated budget hotels
Random coefficients (SD)
Number of zones
Notes: *** indicates significance at the 0.01 level, ** indicates significance at the 0.05 level, * indicates significance at the 0.1 level. Robust
standard errors are presented in parentheses. † indicates the sample on and before 2008, and †† indicates the sample after 2008.
Figure 1. Spatial location of star-rated and budget hotels in Bejing (2012)
0 3.5 7 10.5 141.75 Miles
(a). Star-rated hotels
(b). Budget hotels
Figure 2. Kernal density maps of new hotel entrants in Beijing, 2005-2012
(a). Star-rated hotels
(b). Budget hotels
Figure 3. Standard deviational ellipse of hotel location in 2005 and 2012
Figure 4. The effect of variables on hotels with different star-ratings
Figure 5. Relationship between star-rated hotel density and location probability on hotels with different
Figure 6. Relationship between star-rated hotel density and location probability on non-rated budget