Tourism Management 52 (2016) 451e454

Contents lists available at ScienceDirect

Tourism Management
journal homepage: www.elsevier.com/locate/tourman

Worldwide hedonic prices of subjective characteristics of hostels

Glauber Eduardo de Oliveira Santos*, 1
~o Paulo, Brazil
Federal Institute of Education, Science and Technology of Sa

h i g h l i g h t s

Hedonic prices are applied to analyze prices of more than 8000 worldwide hostels.

Hostel prices are explained by their subjective characteristics as rated by guests.

Cleanliness, location and facilities are the most relevant determinants of prices.

A 1% increase in cleanliness rating leads to a 0.317% increase in hostels' prices.

A 1% increase in location rating leads to a 0.2% increase in hostels' prices.

a r t i c l e i n f o a b s t r a c t

Article history: An overnight at an accommodation establishment might be understood as a package of several different
Received 7 July 2015 items. Since they are not traded separately in the market, their values for the consumer are difficult to
Accepted 9 July 2015 measure. On the other hand, this information is be useful for managers to develop better strategies
Available online xxx
regarding the mix of services and facilities of their establishments. Analytical decomposition of total
prices by different characteristics can be done through hedonic price techniques. In this paper, hedonic
Keywords:
prices are used to analyze the impact of different characteristics on prices of hostels. Accommodation
Hedonic price
prices of more than 8000 hostels worldwide are explained by their characteristics as rated by guests.
Hostel
Accommodation
Results show that the most relevant characteristics are cleanliness, location and facilities.
Hospitality © 2015 Elsevier Ltd. All rights reserved.
Pricing

1. Introduction understood as a package of several different items. Since each

The definition of the project for the accommodation establish-
ment is a highly complex process where a long list of services and
physical characteristics must be defined. The planner must choose
whether the property will include characteristics such as a swim-
ming pool, large rooms, sea view, food services and many other.
Each specific decision might be seen as a conditional profit maxi-
mization problem where the revenue expected from the item being
evaluated is compared to its cost, provided all other characteristics
of the business and its environment.
The problem is that the market price of a particular charac-
teristic is usually difficult to assess. An overnight at an accom-
modation establishment generally includes access to most
services and facilities available. Thus, an overnight might be
~o Paulo, R. Pedro
* Federal Institute of Education, Science and Technology of Sa
Vicente, 625, S~
ao Paulo, SP 01109-010, Brazil.
E-mail address: glaubereduardo@gmail.com.
1
Phd in Tourism and Environmental Economics at the University of the Balearic
Islands.
element is not traded separately in the market, its value for the
consumer is difficult to measure. How much would the consumer
be willing to pay for a better location of the accommodation
establishment? What is the accommodation market price of a
better location? When a guest pays a specific price for the over-
night, it is hard to tell how much of it was due the establishment's
location.
Analytically decomposing total prices by different characteris-
tics can be done through hedonic price techniques. By recognizing
that prices vary across offers of differentiated products, this method
is able to estimate the market price of each characteristic of a good
or service. In this paper, hedonic prices are used to analyze the
impact of different characteristics on prices of hostels worldwide.
This category of accommodation service is traditionally character-
ized by dormitory-style accommodations with basic, shared and
supervised facilities (Chon & Maier, 2010).
Previous hedonic price studies about accommodation estab-
lishments have relied on objective characteristics, such as distance
to attractions and the presence of different facilities. However, what
actually makes the difference for explaining consumer behavior is

http://dx.doi.org/10.1016/j.tourman.2015.07.001
0261-5177/© 2015 Elsevier Ltd. All rights reserved.
452 G.E. de Oliveira Santos / Tourism Management 52 (2016) 451e454
not any objective characteristic, but rather the individual's
perception of it. Thus, objective characteristics are no more than
proxies for the actual explanation of the price composition. More-
over, accommodation is a service, rather than a good. Considering
that the ultimate offer is an experience, it is necessary to recognize
that some characteristics of the accommodation service are
intrinsically subjective. The atmosphere as a characteristic of a
hostel is an unquestionable example of this. Considering these ar-
guments, this study relied on subjective characteristics of hostels as
rated by their guests in order to explain market prices. This inno-
vative approach, besides being more direct than the previous one, is
able to provide information on characteristics that were not
examined in the previous literature.
2. Literature review
In the 1920s some economists were already using statistics to
explain prices according to characteristics (Baranzini, Ramirez,
Schaerer, & Thalmann, 2008b). Lancaster (1966) developed a
renowned theory to explain consumers' demand from goods'
characteristics. According to his theory, individuals obtain utility
not directly from goods, but actually from their characteristics.
Rosen (1974) built on the ideas of Lancaster (1966) and developed a
complete theory of hedonic prices, linking consumers' preferences
on goods' characteristics and market prices. According to him,
given some conditions, market price is a proxy for consumers'
preferences and willingness to pay for characteristics of available
goods and services.
Hedonic price methods have long been used to adjust price in-
dexes for quality (Diewert, , Greenlees, & Hulten, 2009; Triplett,
2006). They have also being largely employed to valuate non-
market assets, especially environmental amenities (Baranzini, ,
Ramirez, , Schaerer, & Thalmann, 2008a). In the tourism field, he-
donic prices have been used to assess the value of characteristics in
different markets, including destinations (Rigall-i-Torrent & Fluvia ,
2011; Rigall-i-Torrent & Fluvia, 2007; Rigall-i-Torrent et al., 2011),
tour packages (Aguilo  , Alegre, & Sard, 2003; Thrane, 2005), res-
taurants (Gunawardana & Havrila, 1996; Yim, lee, & Kim, 2014),
national parks (Pearson, Tisdell, & Lisle, 2002), ski lift tickets (Falk,
2008) and scuba diving (Brander, Van Beukering, & Cesar, 2007).
Several studies applied hedonic price methods to the accom-
modation sector. While Hamilton (2007) examined general ac-
commodation establishments, most of the studies focused on
hotels (Chen & Rothschild, 2010; Espinet, Saez, Coenders, & Fluvia ,
2003; Lee & Jang, 2010; Rigall-i-Torrent et al., 2011). Other cate-
gories of accommodation analyzed include bed and breakfasts
(Monty & Skidmore, 2003), rural accommodation (Fleischer &
Tchetchik, 2005) and rural recreational houses for rental
(Rambonilaza, 2006). Juaneda, Raya, and Sastre (2011) compared
prices of hotels and apartments for rental. It is relevant to stress
that no previous hedonic price study about hostels was found in the
Table 1
Summary of previous hedonic price studies of the accommodation sector.
Study Sector
Espinet et al. (2003) Hotels
Monty and Skidmore (2003) Bed and breakfasts
Chen and Rothschild (2010) Hotels
Hamilton (2007) General accommodations
Fleischer and Tchetchik (2005) Rural accommodation
Rambonilaza (2006) Rural recreational houses for rental
Juaneda et al. (2011) Apartments for rental and hotels
Lee and Jang (2010) Hotels
Rigall-i-Torrent et al. (2011) Hotels
main tourism and hospitality journals.
Most hedonic price studies of the accommodation sector
focused on European and North American destinations, Taipei
(Chen & Rothschild, 2010) and Israel (Fleischer & Tchetchik, 2005)
being the only exceptions. Sample sizes vary substantially. Most
studies used samples between one and two hundred observations.
Two remarkably large samples are those used by Espinet et al.
(2003) and Juaneda et al. (2011), surpassing 82 and 16 thousand
observations, respectively. Finally, it is relevant to stress that log-
linear regression is the most common statistical method used to
estimate hedonic function parameters, although some extensions
of this model were employed when necessary.
Several previous studies focused the valuation of accommoda-
tion establishment location. Distance to relevant attractions, such
as the beach, was one of the most studied variables. Price differ-
ences across destinations were also frequently examined. Other
focus of attention was establishments' specific facilities. Table 1
summarizes the main features of these previous studies.
3. Data and method
Data about prices and characteristics of hostels were collected
from the website Hostelworld, which is the world leading hostel
booking channel (Hotel Industry Magazine, 2014). It is also the
most visited URL related to the term “hostel” (Alexa, 2015),
assuming position number 4968 at the global rank of all URLs,
while the second URL related to “hostel” is at rank 9273. The
website was launched in 1999 and lists over 27 thousand properties
in more than 180 countries (Hostelworld, 2015), including hostels,
bed and breakfasts, hotels, camping sites and other categories of
accommodation establishments.
This study analyzed only accommodation establishments clas-
sified as hostels. Information about the complete Hostelword's list
of 9377 hostels located at 2114 cities of 134 different countries were
gathered in January 2015. For hostels offering only one type of ac-
commodation, a single price was recorded. For those offering dor-
mitories and private rooms, the average of both prices was
considered. Regarding price variations related to different dates,
the minimum available price for any date was recorded. Consid-
ering the positive skewness of registered prices, values were
transformed into their natural logarithms. Therefore, the final
dependent variable of the model was the natural logarithm of the
average of minimum available prices in US$ for each hostel. Outliers
representing the 0.25% lowest and highest percentiles for prices of
each type of accommodation were excluded from the dataset due to
the high probability that these values represent transcription errors
or other sorts of unrealistic prices. After outliers elimination, the
sample size dropped down to 9342 hostels.
The main explanatory variables regarded guests' ratings of six
hostels' characteristics: atmosphere, cleanliness, facilities, location,
staff and security. The characteristic value for money was dropped
Place Observations
Costa Brava, Spain 82,000
Southeast Wisconsin, USA 122
Taipei 146
SchleswigeHolstein, Germany 189
Israel 197
Aveyron, France 799
Spain and France 16,222
USA 106
Costa Brava, Spain 4934
 G.E. de Oliveira Santos / Tourism Management 52 (2016) 451e454 453

from the original Hostelworld list of ratings since it intrinsically
refers to the relationship between price and overall rating, thus
bringing endogeneity problems to the model. The overall hostel
rating was not included in the model since it is just a linear com-
bination of lower level ratings. Each rating varies from 0% to 100%,
representing the mean of individual ratings informed by guests
who booked their accommodation through Hostelworld. The
website display no rating information for establishments with less
than 12 months in Hostelworld and for those that received no re-
view over the previous 12-month period. Due to this limitation,
1324 hostels had to be dropped out of the dataset, taking the
sample size down to 8018 establishments.
The natural logarithm of the total number of ratings received by
each hostel was also included as an explanatory variable in order to
capture potential scale effects. Besides, two dummy variables were
used to indicate whether the price of each specific observation
referred to dormitories, private rooms or the average of both. A
graphical synthesis of the model is presented at Fig. 1.
Hostels' prices vary according to their city since inputs' prices
vary geographically. However, since the number of cities examined
in this study is very large, the estimation city effects on hostels'
prices by using dummy variables would be cumbersome. A better
solution was to use a fixed effects regression model where cities
were considered to have an individual effects on prices and to be
correlated with other explanatory variables (Wooldridge, 2002). In
summary, the estimated model was

 X K
ln Pij ¼ bk Xijk þ qD Dij þ qP Pij þ uj þ εij
k¼1
where P is the price, X is the rating of a characteristic, D is a dummy
variable indicating that the price refers to accommodation in dor-
mitories, P is a dummy variable indicating that the price refers to
accommodation in private rooms, u is the city effect, ε is a random
error term, b and q are parameters to be estimated, i indicates the
hostel, j indexes the city and k refers to each specific characteristic.
4. Results and discussion
The estimation of the model was carried out with the help of
Stata 11. The R2 of the model was 0.12, indicating a fair goodness of
fit of the part of the equation composed by explanatory variables.
Apart from that, the city effect was able to explain 73% of the
variation. Therefore, prices' variation is more associated with the
city of the hostel than with its other characteristics under consid-
P
eration. The correlation between uj and Kk¼1 bk Xijk was 0.11. The
Hausman (1978) test indicates that this correlation is highly sig-
nificant (p < 0.001), pointing out that the fixed effect model was
more adequate than the random effect model that assumes no
correlation between these two terms. This result contrasts with the
a priori choice of a random effects model done by Espinet et al.
(2003). Estimates of the model are presented at Table 2.
Among the six hostel characteristics under scrutiny, two were
found to have non-significant effects on prices: atmosphere
(p ¼ 0.185) and security (p ¼ 0.545). Thus, managers should not be
willing to spend a considerable share of their resources on the
improvement of these two characteristics. The effect of staff was
estimated to be negative and significant (p ¼ 0.014). Thus, better
staff ratings are associated with lower hostel prices. The negative
signal could potentially be explained by a correlation of staff ratings
with other non-valued characteristics omitted in this model.
Three other characteristics were found to have positive and
significant effects on hostels' prices: cleanliness, facilities and
location. Among these, cleanliness is the most relevant factor. A 1%
increase in its average rating is associated with a 0.317% increase in
hostels' prices. In the case of location, a similar rating improvement
leads to a price increase of 0.2%. Finally, for hostels' facilities a 1%
improvement in the average rating is associated with a price in-
crease of 0.172%.
The number of reviews is associated with lower prices. As the
number of reviews increases 1%, hostels' prices decrease 0.028%.
This could be associated with economies of scale, where larger
hostels are able to charge lower prices, at the same time that
they have more guests and consequently a greater number of
reviews. Finally, it is useful to note that accommodation in dor-
mitories is about 30% cheaper than in private rooms (0.128-
(-0.168) ¼ 0.296).
Managers should be willing to invest in improving these
characteristics of their hostels as far as the revenue growth arising
from price increase is sufficient to finance a better service able to
boost hostels' rating by a respective percentage. For instance, the
manager should be willing to afford a better location that pro-
vides a 1% better rating always that the cost of this location is
lower than the revenue growth correspondent to a 0.317% price
increase. This analysis is slightly different for elements of fixed
costs, such as location, and for those associated with variable
costs, such as cleanliness. The analysis is obviously context

Table 2
Hedonic prices' estimates.

Variable Coefficient Standard error p-value

Atmosphere 0.078 0.059 0.185

Cleanliness 0.317 0.056 0.000
Facilities 0.172 0.064 0.007
Location 0.200 0.041 0.000
Staff 0.143 0.058 0.014
Security 0.038 0.062 0.545
Number of reviews (ln) 0.028 0.003 0.000
Dormitories 0.168 0.011 0.000
Private rooms 0.128 0.013 0.000
Constant 3.415 0.036 0.000
Fig. 1. Graphical synthesis of the model.
454 G.E. de Oliveira Santos / Tourism Management 52 (2016) 451e454
dependent. In particular, it depends on the real estate market,
geographical characteristics of the destination, competition, sup-
pliers and consumer preferences. It is relevant to stress that the
set of estimates presented in this paper reflects a global mean
that might vary across different countries and destinations.
However, it is also true that the demand for hostels is quite
globalized, indicating that geographical variations of preferences
might be not so relevant.
References
Aguilo, E., Alegre, J., & Sard, M. (2003). Examining the market structure of the
German and UK tour operating industries through an analysis of package hol-
iday prices. Tourism Economics, 9, 255e278.
Alexa. 2015. Website [Online]. Available: http://www.alexa.com/(accessed
14.01.15.).
Baranzini, A., Ramirez, J., Schaerer, C., & Thalmann, P. (Eds.). (2008a). Hedonic
methods in housing markets: Pricing environmental amenities and segregation. ,
New York: Springer.
Baranzini, A., Ramirez, J., Schaerer, C., & Thalmann, P. (2008b). Introduction. In
A. Baranzini, J. Ramirez, C. Schaerer, & P. Thalmann (Eds.), Hedonic methods in
housing markets: Pricing environmental amenities and segregation. New York:
Springer.
Brander, L. M., Van Beukering, P., & Cesar, H. S. J. (2007). The recreational value of
coral reefs: a meta-analysis. Ecological Economics, 63, 209e218.
Chen, C. F., & Rothschild, R. (2010). An application of hedonic pricing analysis to the
case of hotel rooms in Taipei. Tourism Economics, 16, 685e694.
Chon, K., & Maier, T. A. (2010). Welcome to hospitality. Clifton Park (USA): Cengage
Learning.
Diewert, W. E., Greenlees, J., & Hulten, C. R. (Eds.). (2009). Price index concepts and
measurement. Chicago: University of Chicago Press.
Espinet, J. M., Saez, M., Coenders, G., & Fluvia , M. (2003). Effect on prices of the
attributes of holiday hotels: a hedonic prices approach. Tourism Economics, 9,
165e177.
Falk, M. (2008). A hedonic price model for ski lift tickets. Tourism Management, 29,
1172e1184.
Fleischer, A., & Tchetchik, A. (2005). Does rural tourism benefit from agriculture?
Tourism Management, 26, 493e501.
Gunawardana, P. J., & Havrila, I. I. (1996). An analysis of restaurant characteristics
and meal prices in Melbourne, Australia. Tourism Economics, 2, 79e93.
Hamilton, J. M. (2007). Coastal landscape and the hedonic price of accommodation.
Ecological Economics, 62, 594e602.
Hausman, J. A. (1978). Specification tests in econometrics. Econometrica, 46,
1251e1271.
Hostelworld. 2015. About Hostelworld.com [Online]. Available: http://www.
hostelworld.com/aboutus.php 15 Jan. 2015].
Hotel Industry Magazine. (2014). New youth travel report reveals booking lead times
are coming down as mobile usage rises [Online]. Available http://www.hotel-
industry.co.uk/2014/04/new-youth-travel-report-reveals-booking-lead-times-
are-coming-down-as-mobile-usage-rises/ (accessed 14.01.15.).
Juaneda, C., Raya, J. M., & Sastre, F. (2011). Pricing the time and location of a stay at a
hotel or apartment. Tourism Economics, 17, 321e338.
Lancaster, K. J. (1966). A new approach to consumer theory. Journal of Political
Economy, 74, 132e157.
Lee, S. K., & Jang, S. C. (2010). Room rates of U.S. airport hotels: examining the dual
effects of proximities. Journal of Travel Research, 50, 186e197.
Monty, B., & Skidmore, M. (2003). Hedonic pricing and willingness to pay for bed
and breakfast amenities in Southeast Wisconsin. Journal of Travel Research, 42,
195e199.
Pearson, L. J., Tisdell, C., & Lisle, A. T. (2002). The impact of Noosa National Park on
surrounding property values: an application of the hedonic Price method.
Economic Analysis and Policy, 32, 155e171.
Rambonilaza, M. (2006). Labelling and differentiation strategy in the recreational
housing rental market of rural destinations: the French case. Tourism Economics,
12, 347e359.
Rigall-i-Torrent, R., & Fluvi a, M. (2007). Public goods in tourism municipalities:
formal analysis, empirical evidence and implications for sustainable develop-
ment. Tourism Economics, 13, 361e378.
Rigall-i-Torrent, R., & Fluvia , M. (2011). Managing tourism products and destina-
tions embedding public good components: a hedonic approach. Tourism Man-
agement, 32, 244e255.
Rigall-i-Torrent, R., Fluvia , M., Ballester, R., Salo
 , A., Ariza, E., & Espinet, J. M. (2011).
The effects of beach characteristics and location with respect to hotel prices.
Tourism Management, 32, 1150e1158.
Rosen, S. (1974). Hedonic prices and implicit markets: products differentiation in
pure competition. Journal of Political Economy, 82, 34e55.
Thrane, C. (2005). Hedonic price models and sun-and-beach package tours: the
Norwegian case. Journal of Travel Research, 43, 302e308.
Triplett, J. (2006). Handbook on hedonic indexes and quality adjustments in Price
indexes: Special application to information technology products. Paris: OECD.
Wooldridge, J. M. (2002). Econometric analysis of cross section and panel data.
Cambridge: MIT Press.
Yim, E. S., lee, S., & Kim, W. G. (2014). Determinants of a restaurant average meal
price: an application of the hedonic pricing model. International Journal of
Hospitality Management, 39, 11e20.
Glauber Eduardo de Oliveira Santos Professor in the
Tourism and Hospitality Department of the Federal Insti-
tute of Education, Science and Technology of Sa ~o Paulo
(IFSP) and in the postgraduate program of the School of
Arts, Science and Technology of the University of S~ao Paulo
(EACH-USP). PhD in Tourism and Environmental Eco-
nomics at the University of the Balearic Islands. Director of
the International Academy for the Development of
Tourism Research in Brazil.