# RECORD 9ba5bc20c4b8dd4891df1e9ed1d47b53aa5ee9df
# TITLE: The Impact of Housing Market Segmentation between Tourists and Residents on the Hedonic Price for Landscape Quality
# YEAR: 2008.0  VENUE: nan  DOI: 10.1007/bf03399270
# AUTHORS: N. Soguel; Marc J. Martin; Alexandre Tangerini
# PAGES: 24  CANDIDATE TABLE PAGES (1-based): [10, 11, 15, 16, 17]
# ABSTRACT: SummaryMarket segmentation is an important issue when estimating the implicit price for an environmental amenity from a surrogate market like property. This paper tests the hypothesis of a segmentation of the housing market between tourists and residents and computes the implicit price for natural landscape quality in Swiss alpine resorts. The results show a clear segmentation between both groups of consumers, although tests also show that the estimated coefficient for landscape is similar in the tourists’ model and in the residents’. However, since the functional form is non linear, the nominal — rather than relative — value of a change in natural landscape quality is higher in the tourist housing market than in the residents’. Hence, considering the segmentation of the market between tourists and residents is essential in order to provide valid estimates of the nominal implicit price of natural landscape quality.


===== PAGE 1 (score 3, numeric-paren 0, table-words 0, var-words 3) =====

© Swiss Journal of Economics and Statistics 
2008, Vol. 144 (4) 655–678
a 
Institut de hautes études en administration publique-IDHEAP/Swiss Graduate School of 
Public Administration, University of Lausanne, Switzerland.
b 
Service cantonal de recherche et d’information statistiques-SCRIS/ Statistical Office of the 
Canton of Vaud.
 
Corresponding author: nsoguel@idheap.unil.ch. The research benefited from the support of 
the Swiss National Research Foundation allocated as part of the National Research Program 
48 “Landscapes and habitats of the Alps”. Helpful comments by Anil Markandya, Tim Taylor 
and the editors of this special issue are gratefully acknowledged.
The Impact of Housing Market Segmentation  
between Tourists and Residents on the Hedonic Price  
for Landscape Quality
Nils Soguela, Marc-Jean Martinb and Alexandre Tangerinia
JEL-Classification: D49, Q26, Q51
Keywords: Hedonic prices method, Market segmentation, Landscape quality, Landscape value, 
MACBETH
1. Introduction
Market segmentation has long been a controversial empirical issue in the use of 
the hedonic price method. It happens when a given market, such as housing, is 
organised into a series of submarkets each of them showing a unique functional 
relationship between prices and (property) attributes. If segmentation does exist, 
estimating a single hedonic price function for the whole market provides faulty 
estimates of the implicit prices. A bias in implicit prices is especially problemati-
cal when the hedonic price technique utilises the considered market to estimate 
economic values for environmental services, like landscape. The very idea of the 
method is to overtake the problem of the absence of clearly defined markets for 
environmental resources and to observe some surrogate market instead. The most 
common markets used as surrogates are those for property or labour. Therefore 
any bias in implicit price for an environmental amenity is all the more damag-
ing as few alternative techniques may exist.
Segmentation does not render the hedonic pricing technique invalid. Rather 
it makes its use more difficult, as separate functions must be computed for each 


===== PAGE 10 (score 37, numeric-paren 24, table-words 2, var-words 7) =====

664 
Soguel / Martinl / Tangerini
8 
Another strategy to cope with the differences between resorts consists of introducing in the 
model a dummy variable for each resort. Results for this strategy are not reported for two 
reasons. First, it leads to identical conclusions: the impact of the landscape quality is simi-
lar and so is the finding about the segmentation between residents and tourists as well as the 
spatial segmentation. Second, this strategy comes at the cost of having to abandon all other 
local neigbourhood characteristics Ni since these variables do not vary within the resorts and 
are thus colinear with the resort-specific dummies. It also offers less explanatory power than 
the models where the number of rooms interacts with a dummy for each resort (see Models 2 
and 4).
resort (e.g. lnRooms × Verbier) resulting in strictly positive values when flats 
are located in Verbier (Verbier = 1) and null values when flats are located else-
where (Verbier = 0)8.
Hence, the results for four models are presented in Table 2: two for tourists and 
two for residents. Breusch-Pagan, Cook-Weisberg and Szroeter tests signal hetero-
scedasticity problems especially when modelling without spatially segmenting the 
number of rooms and modelling the market for residents. Therefore we use the 
Huber-White-sandwich estimator of the variance of the regression estimators.
Table 2: Hedonic Price Functions for Apartments Rented either by Tourists or by 
Residents in Six Swiss Alpine Resorts (OLS, log-log Specification)a
Independent variables
Models
Tourists
Residents
1
2
3
4
Ln Rooms (number of rooms in the apartment 
excluding kitchen and bathrooms)
0.5748**
(21.73)
0.5190**
(14.56)
Ln Rooms × Verbier (number of rooms in the 
apartment if located in Verbier)
0.9036**
(15.17)
0.6100**
(7.87)
Ln Rooms × Anzere 
0.6507**
(15.17)
0.4445**
(12.62)
Ln Rooms × Ovronnaz 
0.5234**
(9.50)
0.5264**
(9.83)
Ln Rooms × Champery
0.5178**
(12.88)
0.4082**
(8.22)
Ln Rooms × Grimentz
0.4813**
(12.54)
0.5168**
(14.46)


===== PAGE 11 (score 68, numeric-paren 46, table-words 2, var-words 11) =====

The Impact of Housing Market Segmentation between Tourists and Residents 
665
Independent variables
Models
Tourists
Residents
1
2
3
4
Ln Rooms × Haute-Nendaz 
0.4209**
(9.91)
0.4182**
(4.80)
Ln Landscape-Quality (visibility-adjusted 
MACBETH score)
0.0449**
(6.57)
0.0225**
(3.84)
0.0351**
(4.28)
0.0287**
(2.67)
Ln Resort-Infrastructureb
0.1297**
(6.19)
−0.0260
(−0.54)
0.1962
(0.82)
0.0156
(0.58)
Ln Distances-to-Infrastructure (sum of the 
distances to the resort’s main infrastructures: 
ropeways, grocery stores, and resort centres)
0.0428**
(2.68)
0.0001
(0.01)
0.1393**
(2.63)
0.1509**
(2.91)
Ln Average-Ski-Runs (ratio between the 
length of the runs in kilometres and the 
number of ski lifts) c
0.3234**
(10.45)
0.1678*
(2.40)
−0.0380
(−0.53)
-0.1663*
(-2.22)
Ln Flat-Equipment (1+sum of dummies if the 
flat features a TV-cable, fire-place, balcony, 
dishwasher, security system)
0.1545**
(3.32)
0.1914**
(4.44)
0.1027**
(2.75)
0.1091*
(2.15)
Constant
7.3369**
(92.10)
7.2409**
(52.94)
6.1568**
(38.02)
6.3375**
(42.43)
N
Adjusted-R2
F
Largest VIF
Mean VIF
σ2 (OLS)
401
0.7472
202.56
1.38
1.25
0.22
401
0.8151
187.30
6.93
3.33
0.19
105
0.8517
91.15
2.43
1.98
0.15
105
0.8677
80.91
4.36
2.62
0.15
a.  The dependent variable is the natural logarithm of the rent during the period of February to 
March 2003 (including bills). For residents, the average nominal monthly rent equals 812 CHF, 
with a standard deviation of 308. For tourists, the average nominal rent on a monthly basis 
equals 4635 CHF, with a standard deviation of 2301. In March 2003, 1 CHF = 0.681 €.
 
T-values are shown between parentheses under the estimated parameters according to Huber-
White-sandwich estimator of the variance of the regression estimators. Coefficients with 
double asterisk are significant to 99%. Those with a single asterisk are significant to 95% 
(bilateral test). 
b. The infrastructure index is made up of several sub-indexes. Each of them measures the impor-
tance of a type of infrastructure relative to the total offered by the six resorts. For example, Ver-
bier hosts 13 (34%) of the 38 bars of the six resorts (100%) and 18 (46%) of the sports stores. 
After averaging the percentages over the seven different types of infrastructure, Verbier offers 
the largest total infrastructure (37%), and Grimetz and Ovronnaz the smallest (7% each).
c. This average length offers an estimate of the “rate-of-return” one can expect from each lift, in 
terms of the duration of the descent.


===== PAGE 12 (score 9, numeric-paren 0, table-words 1, var-words 7) =====

666 
Soguel / Martinl / Tangerini
9 
Several additional variables were tested and rejected in light of the criteria used to select the 
variables to be retained in the models. Just to mention a few, they include the quality of the 
built environment in the resort (MACBETH score), the level of noise pollution, the age of 
the house, the number of years since the last repair either in the house or in the apartment, 
the number of apartments in the house, and the floor on which the apartment is located.
There is no indication of multicollinearity, according to the variance inflation 
factors (the largest and mean VIF peaks at a value of 6.93 and 3.33 which is far 
from the generally accepted threshold of 10).
The independent variables retained in the hedonic function explain a share of 
the rent variability ranging between 0.75 (Model 1) and 0.87 (Model 4). How-
ever the set of independent variables that influence significantly and with the 
expected sign both tourists’ and residents’ rent is rather small.9
As in many hedonic price studies, the number of rooms (Ln Rooms) influ-
ences positively and significantly both tourists’ and residents’ rent. It is true in 
both types of models: with or without interaction between the number of rooms 
and a dummy variable for each resort (Models 4 and 2, respectively 3 and 1). 
However, according to the size of the coefficient, the effect is larger on the tour-
ist market.
It is interesting to see that interacting the number of rooms with a dummy for 
each resort captures the effect of local neighbourhood characteristics like the visi-
bility-adjusted quality of the landscape (Ln Landscape-Quality), the resorts infra-
structure (Ln Resort-Infrastructure), the distance to infrastructure (Ln Distances-
to-Infrastructure) and the average length of the ski runs (Ln Average-Ski-Runs). 
In fact, these latter characteristics are able to significantly explain the variabil-
ity of the rents especially in Model 1 (model for tourists). However the value of 
their coefficients obviously drops when one introduces interaction variables for 
the number of rooms into the model, as does their degree of significance. This 
demonstrates that the local neighbourhood characteristics at our disposal are 
not powerful enough to explain fully the variability of the rents between the six 
considered resorts. Therefore, Model 1 still suffers to some extent from a spatial 
segmentation bias.
Although the statistics for the visibility-adjusted quality of the landscape are 
affected when moving from Model 1 to Model 2 or from Model 3 to Model 4, 
they nevertheless convey that the quality of landscape impacts significantly and 
positively on the rents.
The distance to a resort’s main infrastructure influences the rents with a sub-
stantial probability, except in Model 2. Interestingly, it impacts positively, mean-
ing that rents are relatively higher for apartments located away from the resorts’ 


===== PAGE 15 (score 21, numeric-paren 11, table-words 2, var-words 4) =====

The Impact of Housing Market Segmentation between Tourists and Residents 
669
An important consequence of the existence of submarkets is that estimates of 
implicit prices derived from the parameters of a pooled data hedonic function 
are likely to be biased. To compare the parameters across submarkets we employ 
the Tiao-Goldberger test suggested by Michaels and Smith (1990). In our case 
the test is particularly useful since it considers the null hypothesis that the coef-
ficients of a particular regressor take the same value in each of the models. The 
test works by comparing each parameter estimate to the weighted sum of param-
eter estimates across Models 1 and 3 as well as across Models 2 and 4. Table 3 
reports the parameter estimates and the Tiao-Goldberger test statistics for the 
Table 3: Results of Tests for Differences between Tourists’  
and Residents’ Hedonic Functionsa
Comparison of models …
… without geographical 
interaction terms 
(Models 1 and 3)
… with geographical 
interaction terms 
(Models 2 and 4)
Chow test F-statistics
547.69**
401.49**
Wald test Chi-squared statistics
8047.6314**
12789.516**
Tiao-Goldberger test F-statistics
Ln Rooms
0.66
Ln Rooms × Verbier
11.89**
Ln Rooms × Anzere 
10.29**
Ln Rooms × Ovronnaz 
0.02
Ln Rooms × Champery
0.62
Ln Rooms × Grimentz
0.00
Ln Rooms × Haute-Nendaz 
0.02
Ln Landscape-Quality
1.29
0.38
Ln Resort-Infrastructure b
1.08
0.06
Ln Distances-to-Infrastructure
1.53
4.39*
Ln Average-Ski-Runs
18.32**
4.26*
Ln Flat-Equipment
0.48
0.12
Constant
309.97**
13.94**
N
506
506
a 
Coefficients with double asterisks are different at the 99% level between the compared models, 
those with a single asterisk at the 95% level.


===== PAGE 16 (score 14, numeric-paren 4, table-words 2, var-words 5) =====

670 
Soguel / Martinl / Tangerini
coefficients that were estimated across the two submarkets. Of most interest in 
this research is the possible difference in the implicit price of the landscape qual-
ity. The test statistics for both comparisons indicate that the null hypothesis of 
equality of parameter estimates cannot be rejected, even at the 1% level of con-
fidence. This finding is important and shows that the quality of the landscape 
does not affect the rental markets of tourists and residents in different ways. In 
turn the null hypothesis can be rejected for some other parameters (test scores 
italicised). Models with geographical interaction terms are different as far as five 
coefficients are concerned, including the constant.
This finding leads us to adapt the estimation strategy and to compute the two 
more general models presented in Table 4. Each model is a joint model for both 
tourists and residents. When the Tiao-Goldberger test indicates that the param-
eters are different between the two markets the concerned explanatory variable 
is interacted either with a dummy indicating that the tenant is a resident (R) or 
with a dummy indicating the tenant is a tourist (T). Model 5 presents results 
without geographic interaction terms for the number of rooms and Model 6 
presents results with geographic interaction terms.
This further modelling step has three consequences. First, it sheds further 
light on the differences in parameters and their impact on rents. For example, the 
difference between constants in tourists’ and residents’ models is approximately 
equal to 1 (7.33 − 6.29 in Model 5 and 7.31 − 6.13 in Model 6). The significance 
of the parameters echoes the one found in Models 1 to 4. Second, it provides us 
with more accurate estimates for parameters that are common to both tourists 
and residents. The 95%-confidence intervals for the landscape quality variable 
show clearly the increased accuracy. Intervals are broader in Model 2 (0.0110 to 
0.0340) and Model 4 (0.0074 to 0.0501) than in Model 6 (0.0157 to 0.0341). 
Third, since Models 5 and 6 are general and no longer specific to either tourists 
or residents, the estimated hedonic functions explain a slightly lower share of the 
rent variance than the previous – specific – models. Considering models with 
geographic interaction terms for the number of rooms, the adjusted-R2 is slightly 
lower in Model 6 (0.8087) than in Model 2 (0.8151) for tourists. It is also slightly 
lower in Model 6 (0.8587) than in Model 4 (0.8677) for residents.


===== PAGE 17 (score 40, numeric-paren 22, table-words 5, var-words 3) =====

The Impact of Housing Market Segmentation between Tourists and Residents 
671
Table 4: Joint Hedonic Price Functions for Apartments Rented by Tourists  
and by Residents in Six Swiss Alpine Resorts (OLS, log-log Specification)
Independent variables 
Model 5, without  
geographical  
interaction terms a
Model 6, with  
geographical  
interaction terms b
Parameters
T-values
Parameters
T-values
Ln Rooms
0.5770**
25.76
Ln Rooms × Verbier × R
0.5995**
10.66
Ln Rooms × Verbier × T
0.8857**
20.46
Ln Rooms × Anzere × R
0.4432**
16.66
Ln Rooms × Anzere × T
0.6499**
16.03
Ln Rooms × Ovronnaz 
0.5267**
12.55
Ln Rooms × Champery
0.5102**
14.11
Ln Rooms × Grimentz
0.4905**
15.77
Ln Rooms × Haute-Nendaz 
0.4122**
11.5
Ln Landscape-Quality
0.0417**
7.58
0.0249**
5.31
Ln Resort-Infrastructure b
0.1138**
6.39
−0.0060
−0.19
Ln Distances-to-Infrastructure
0.0470**
3.11
Ln Distances × R
0.1320**
2.69
Ln Distances × T
0.0020
0.13
Ln Average-Ski-Runs × R
−0.0592
−1.01
−0.0777
−1.19
Ln Average-Ski-Runs × T
0.3251**
10.89
0.1695**
2.72
Ln Flat-Equipment
0.1324**
3.93
0.1672**
5.26
Constant × R
6.2926**
56.28
6.1325**
46.96
Constant × T
7.3389**
110.51
7.3079**
66.12
N
Adjusted pseudo-R2 if R = 1
Adjusted pseudo-R2 if T = 1
506
0.8270
0.7422
506
0.8587
0.8087
a 
Same dependent variable as in Models 1 and 3. 
b 
Same dependent variable as in Models 2 and 4. 
R Dummy variable taking the value 1 if the tenant is a resident, 0 if a tourist.
T Dummy variable taking the value 1 if the tenant is a tourist, 0 if a resident.
Coefficients with double asterisk are significant to 99%.  Those with a single asterisk are signifi-
cant to 95% (bilateral test).  T-values according to Huber-White-sandwich estimator of the vari-
ance of the regression estimators.


===== PAGE 18 (score 10, numeric-paren 0, table-words 3, var-words 4) =====

672 
Soguel / Martinl / Tangerini
11 lnY =
1
2
1
1
2
2
1
2
ln
ln
ln
...
ln
, hence 
...
.
nb
b
b
n
n
n
a
b
X
b
X
b
X
Y
aX X
X
+
+
+
+
=
7. Simulating the Impact of Changes in Landscape Quality
Table 5 presents the simulated rent according to the visibility-adjusted quality 
of the landscape in each resort. All other characteristics of the apartment (Si, Ni) 
are held constant at their average value. Simulations are carried out for rents paid 
by tourists (upper part of the table) or by residents (lower part of the table) using 
either models with or without geographic interaction terms for the number of 
rooms (Model 6 and Model 5). When the simulation is based on Model 6 (i.e. 
with interaction terms), two extremes are considered: an apartment specifically 
located in Verbier vs. one in Champery (i.e. interaction variables for resorts other 
than the considered one being set equal to 0). Therefore the results reported in 
the column “Verbier” are for an average flat located in Verbier and rented by a 
tourist (upper part of the column) or a resident (lower part) that benefits from a 
landscape quality varying between 10 and 110 according to the visibility-adjusted 
MACBETH score. Results reported in the column “Haute-Nendaz” are for a flat 
located in this resort, rented either by a tourist or a resident and benefiting from 
varying levels of landscape quality.
Considering the two extreme visibility-adjusted MACBETH scores for the 
landscape quality, the rent for a score of 110 (as in Champery) is considerably 
higher than for a score of 10 (as in Haute-Nendaz). Comparison between the 
simulations using both models also shows differences. Simulations with Model 5 
indicate a percentage of variation due to landscape quality almost twice as high 
as the one with Model 6 (21.6% and 12.4% respectively. Since both models are 
joint models for tourists and residents, the percentage of variation in the rent is 
the same across the two groups. This underscores the assertion that the relative 
marginal implicit price of landscape quality is identical.
However, when it comes to the nominal marginal implicit price (measured 
in CHF), the double-log specification of the hedonic equations must be taken 
into consideration. Due to the multiplicative property of such a specification11, a 
“boosting” and multiplicative effect appears as the constant differs across the two 
groups. Since the difference in constant is 1.05 in Model 5 and 1.18 in Model 6, 
a multiplicative effect of 2.85 (Model 5) and 3.24 (Model 6) occurs on the nom-
inal tourists’ market price when the visibility-adjusted quality of the landscape 
varies (or any other characteristic). Were the landscape quality variable to double, 
then the effect would be an increase in rent of 4.17% according to Model 5 and 
of 2.49% according to Model 6. However the nominal effect measured in CHF