Is there an economic rent for island hotels

1
Is there an economic rent for island hotels?

• Juan Prieto-Rodríguez
• Manuel González-Díaz

2
Aim and motivation

• AIM
• Analyse if there is an insular rent for the
  Spanish hotels
• One of the most important reasons to choose a
  particular hotel is its location
• We study if the island hotels establish higher
  prices than mainland located hotels, once the
  quality of the hotel is controlled
• MOTIVATION
• Little research has been done in this issue,
  specially studying the whole price distribution

3
Exposition order

• Hotel sector
• Hypotheses
• Sample and methodology
• Results
• Conclusions

4
Hotel sector

• The production of the accommodation service
  requires several assets
• A building, goods and chattels, know-how, booking
  center and a brand name.
• Hotel firms may have
• Single establishment ? independent hotel
• Multiple establishments ? Hotel chain
• Hotel chains control the establishments through
  different organizational forms
• Wholly-ownership (vertical integration)
• Hotel chain is the owner of all assets
• Leasing contract
• Hotel chain hires some assets to local partners
  (building, goods chattels, etc.)
• Management service contract
• Hotel chain only owns the brand name and booking
  center but manages other assets
• Franchise contract
• Hotel chain cedes his brand name to a local
  partner

5
Exposition order

• Hotel sector
• Hypotheses
• Sample and methodology
• Results
• Conclusions

6
Hypotheses

• H1 insular rent hypothesis
• Hotels located at Balearic and Canary Islands are
  able to fix higher price than hotels with
  equivalent standards of quality located mainland
• H2 Klein and Lefflers (1981) hypothesis
• Companies develop reputational capital (brand
  names) to solve informational asymmetry between
  producer and consumer
• In exchange for this guarantee, the consumer is
  willing to pay higher prices than he would pay
  for a similar quality hotel without this
  guarantee

7
Exposition order

• Hotel sector
• Hypotheses
• Sample and methodology
• Results
• Conclusions

8
Sample and methodology

• Data set
• The information was obtained initially from the
  Hostelmarket Census of Hotels and Apartments for
  January 2004 and the Hostelmarket Annual Reports
  for 2003 and 2004
• Additional data were taken from the chains
  reports, the SABI data base (System for the
  Analysis of Balance Sheet on the Iberian
  Peninsula) and Turespaña
• Finally, the Official Hotel Guides for 2004 and
  2005 (professional edition) give detailed
  information on the specific characteristics,
  services and prices of the hotels
• Sample size 744 hotel establishments located in
  Spain
• 464 establishments belong to the six main hotel
  chains operating hotels in Spain
• The remaining 280 establishments were chosen by
  random sampling of independent hotels

9
Sample and methodology
10
Sample and methodology

• The quantile regression estimator (Koenker and
  Basset, 1978) consists in estimating regressions
  at several quantiles of the dependent variable
  distribution
• If ln(pi) is the logarithm of the room prices of
  a hotel and it is a random variable and x is a
  vector of explanatory variables, we can define
  the ?th conditional quantile of ln p given X as
• This equation indicates the log price of that
  hotel with characteristics x that leave behind a
  fraction ? of hotels with the same characteristics

11
Sample and methodology

• We can express the linear quantile regression
  model in terms of the log price of hotel rooms
  as
• The distribution function of the error term is
  left unspecified with the only assumption being
  that
• This method allows different effects of the
  explanatory variables depending on the position
  held by hotels in the price distribution
• The estimated coefficients represent the marginal
  change in the ?th quantile conditional price
  induced by a marginal change in the explanatory
  variables

12
Exposition order

• Hotel sector
• Hypotheses
• Sample and methodology
• Results
• Conclusions

13
Results
Figure 2 Estimated coefficient of Island and
Canary Island variables in Model 2
14
Results

• Hotels located at any Spanish island charge
  higher prices than similar hotels located in the
  mainland
• Island hotels charge between a 10 and 30 per cent
  more than mainland hotels with similar
  characteristics (number of stars, facilities and
  located at the coast,)
• This premium increases with prices so medium and
  high quality island hotels are able to get a
  higher location premium
• Canary Islands estimated coefficients are
  negative for almost the whole price distribution
  but they are only significant around the median
• Medium quality Canarian hotels could charge lower
  prices than similar Balearic hotels but they have
  a price premium compared to average mainland
  hotels

15
Results
Figure 3 Island and Canary Island estimated
elasticity of hotel price in Model in Model 3
(chain variables included)
16
Results

• In order to check if results are robust, Model 3
  includes hotel chain dummies and contractual
  variables to take into account the control that
  the hotel chain has over the hotel services
• Once we control for these factor, the price
  premium disappears
• The estimated coefficients for the Island
  variable are lower and only significantly
  different from zero at the 55th and 70th centiles
• Canary Islands coefficients are not significant
  any more
• The economic rent of island hotels of Model 2 is
  related to the large presence of chain hotels in
  the Spanish hotel sector at the islands
• Hotel chains extract a rent from their ability to
  reduce consumers uncertainty about the quality
  of the hotel
• Island hotels charge a higher price because they
  are controlled by hotel chains and not by a
  location rent

17
Results
Figure 4 Estimated coefficient for hotel chain
variables in Model 3
18
Results

• Estimated coefficients for the more important
  Spanish hotel chains (Sol-Melia, HUSA, NH, Riu,
  AC and Hoteltur) are mainly positive
• This result can be interpreted as empirical
  evidence in support of Klein and Lefflers (1981)
  hypothesis
• However, not all the chains have the same ability
  to charge higher prices
• maybe they dont have the same capacity to use
  their brand as a good quality signal
• Sol-Melia, NH and Riu Hotels have positive and
  significantly different from zero estimated
  coefficients at large areas of the price
  distribution

19
Exposition order

• Hotel sector
• Hypotheses
• Sample and methodology
• Results
• Conclusions

20
Conclusions

• Once all visible signals of quality are taking
  into account, it seems that there is a positive
  rent for island hotels
• However, this rent is related to a hotel brand
  name effect and the fact that the most important
  Spanish hotel chains are heavily implanted in
  Canary and Balearic islands
• Hotel brand name reduces consumers information
  asymmetry problems by signalling higher quality
  standards or reducing the expected variance in
  the quality of a hotel given its ex-ante known
  characteristics
• This informational asymmetry seems to be more
  important with regards to island hotels since the
  implicit prices are higher due to the higher
  transporting costs and to the lesser number of
  alternatives in case of a great failure

21
Conclusions

• In our empirical analysis, quantile regression
  estimations have some advantages over OLS
  estimations since the effects of impendent
  variables on hotel prices are not constant along
  the price distribution
• Possible political economy implication it could
  be worthy to help the establishment of strong
  hotel chains with international projection in
  order to persuade new foreign customers more
  easily by reducing their doubts about the quality
  of the hotel service