Examining the determinants of room rates for hotels in capital cities: The Oslo experience
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
www.palgrave-journals.com/rpm Examining the determinants of room rates for hotels in capital cities: The Oslo experience Christer Thrane Received (in revised form): 7th November, 2006 Department of Social Sciences, Lillehammer University College, 2626 Lillehammer, Norway. Tel: þ 47 61 28 82 47; Fax: þ 47 61 28 81 70; E-mail: Christer.Thrane@hil.no Christer Thrane is Professor of Tourism in the what causes room rates to diverge. In this Department of Social Sciences at Lillehammer respect, the multidimensional concept of University College, Norway. His research quality is expected to be associated with hotel interests include, among other things, quanti- prices in a more or less linear fashion: a higher tative research in the areas of tourism, quality equals a higher price. In other words, hospitality and recreation. The author would price differences between hotels signal quality like to thank Jo Kleiven and the anonymous differences between hotels. In a related manner reviewers for valuable comments to an earlier the presence or absence of various hotel draft of this paper and Tone Kvamme and attributes (eg, a spa, a restaurant, a central Guro Larsson for excellent research assis- location etc.) will be among the factors that tance. The usual disclaimer applies. most people will expect to influence hotel room prices. In addition, it stands to reason that ABSTRACT lodging in hotels possessing many desirable KEYWORDS: hotels, Oslo, room rates, price attributes will be more expensive than lodging hedonics, SUR in hotels in which few or no such attributes are present. This paper is concerned with Price hedonic theory states that the price for a product the prices consumers pay for staying in may be thought of as an additive function of the various the various hotels in the proximity of the utility-bearing attributes embedded in the product. Norwegian capital of Oslo. In particular, it Within this framework, the present study demonstrates examines how a number of these hotels’ how the room rates for hotels in and around the attributes explain variation in the room Norwegian capital of Norway can be linked to certain rates for single and double rooms. The hotel attributes. Seemingly unrelated regression (SUR) theoretical foundation for this research ques- models incorporating nine hotel attributes explain tion is found in the price hedonic literature. In about 70% of the variation in room rates. Of particular general terms, the results of price hedonic importance in this respect, are the attributes mini-bar, studies may have important implications for hairdryer, free parking and distance to downtown. consumers who want to compare the ‘marginal Journal of Revenue and Pricing Management (2007) 5, utility’ of different product attributes. As such, 315–323. doi:10.1057/palgrave.rpm.5160055 the results shed light on which (hotel) attributes they have to pay extra for and which (hotel) INTRODUCTION attributes can be bought without a price It is common knowledge that the prices people surcharge. The flip side of this, however, is have to pay for accommodation in hotels vary that the findings also can be used by the enormously. Furthermore, most people prob- hospitality industry to enhance their strategic ably possess a more or less accurate intuition of pricing. & 2007 Palgrave Macmillan Ltd, 1476-6930 $30.00 Vol. 5, 4 315–323 Journal of Revenue and Pricing Management 315
Examining the determinants of room rates for hotels THEORETICAL FRAMEWORK AND Skidmore (2003) scrutinised how a number of PRIOR RESEARCH hotel attributes could be linked to room rates Price hedonic theory states that the price for a local US market. Finally, Coenders et al. of a product can be regarded as a function (2003) examined how hotel room prices in the of its immanent attributes or characteristics sun-and-beach segment in the Mediterranean (Lancaster, 1966; Rosen, 1974). Typical appli- region were influenced by a number of hotel cations of the theory involve the price of attributes. residential properties (Fujita, 1989), whereas In a related spirit, a number of papers have more esoteric applications concern that of sought to determine how prices for (all- Bordeaux wine (Combris et al., 1997). In the inclusive) package holidays are affected by case of residential properties, the sale price of a hotel characteristics and related features (eg, house can be regarded as an additive function Sinclair et al., 1990; Aguilo et al., 2001, 2003; of its size, number of rooms, amenities, Papatheodorou, 2002; Espinet et al., 2003; neighbourhood quality and proximity to key Thrane, 2005). Also, Roubi and Litteljohn institutions. By the same logic, a wine’s district (2004) developed a regression model within a of production, vintage and subjective quality price hedonic framework to explain the (assessed by taste panels) will to a large degree variation in sales prices for hotel properties determine its sale price. In other words, sold in the UK between 1996 and 2002. hedonic price analysis assumes that the price of a product or a service is a function of a THE PRESENT STUDY bundle of attributes. Based on the general ideas embedded in price A few studies with a focus on the determi- hedonic theory and its prior hospitality and nants of hotel room prices or room rates have tourism applications, there is ample evidence to adopted the framework described above, albeit suggest that the presence or absence of certain to varying degrees. Carvell and Herrin (1990) hotel attributes affect the room rates faced by showed that price variations among hotels in consumers when they make their decisions San Francisco were attributable to features such regarding hotel accommodation. Formally, as location and other hotel characteristics. In a the determinants of the room rate (R) a similar vein, Bull (1994) scrutinised how consumer must pay for a hotel stay is a function certain locational and site-specific attributes of various objective hotel attributes (O) (eg, affected motel room prices in Ballina, Australia. type of board; distance to downtown; presence The small number of hotels/motels included in of swimming pool, bar and restaurant etc.) these studies (20 and 15, respectively), how- and, possibly, more subjective attributes (S) ever, precluded any certain generalisations. Wu (eg, service quality; hotel star rating; (1998) examined whether chain-associated atmosphere etc.). Accordingly, the room rate motels in Arkansas and Kansas were more for the ith hotel stay (Ri) can be described as expensive to stay in than their non-chained Ri ¼ P(Oi, Si) where both Oi and Si are vectors counterparts. More recently, Israeli (2002) of attributes. Ordinary Least Squares (OLS) studied how hotel room prices in Israel were regression or the related log-linear form have in influenced by hotels’ star rating, number of prior hospitality or tourism applications mostly rooms and corporate affiliation (ie, chain been used to estimate this type of hedonic price association). In a related manner, White and model. Mulligan (2002) demonstrated how attributes As there exists no official star-rating classi- such as presence of a pool and a spa, as well as fication for the hotels that make up the data in climatic and locational features, affected the this study, the subsequent analysis focuses solely hotel room rates in four southwestern US on objective attributes. At first glance this states. Using the same approach, Monty and might appear as a limitation because most 316 Journal of Revenue and Pricing Management Vol. 5, 4 315–323 & 2007 Palgrave Macmillan Ltd, 1476-6930
Thrane previous scholars have included both star rating reasons for why the data could not be and objective attributes in their price hedonic envisioned as a random sample from a universe regressions. However, as argued by Papatheo- of hypothetical cases (Henkel, 1976). dorou (2002) and Thrane (2005), using star The two dependent price variables in rating as an independent variable alongside the study were the weekday rates in March objective attributes amounts to a specification 2005 for a one-night stay in a single or a double error because star rating is an endogenous room. It is important to note that these independent variable. The simple reason for room rates do not necessarily reflect the price this is that the star rating variable becomes a consumers actually pay because stays booked a function of objective attributes to a substantial long time in advance usually are bought degree, in much the same sense as price. at discount prices. As the hotels or hotel Another problem that arises with this proce- chains set their prices in advance, however, dure is that it likely will cause multicollinearity they signal hotel quality. For this reason it is in many cases (ie, high correlations between the unproblematic to use these ‘price-proposals’ as independent variables). Thus, the lack of an dependent variables within a price hedonic official star rating classification for the hotels in framework (cf. Israeli 2002; Papatheodorou the data does not create problems in the present 2002). Table 1 presents descriptive statistics for context. these dependent variables, both in level and In summary, prior research has clearly logged form. Table 1 also presents descriptive demonstrated that the prices faced by con- statistics for the attributes or independent sumers in their choice of accommodation variables considered in the study. One parti- depend on the attributes embedded in the cular problem with price hedonic theory is lodging facilities. In line with this research the that it offers few theoretical guidelines for purpose of the present study is to examine the selecting independent variables (Andersson relationships between a number of hotel 2000). The selection of independent variables attributes and room rates in and around the in the present context was based on the Norwegian capital of Oslo. research cited above, and in particular on the list of relevant attributes compiled by Data Espinet et al. (2003). The information about The data for this study were extracted from the the attributes were mostly gathered from Internet-based search engine for hotels in the various hotels’ websites, but in some Norway in March 2005 (www.hotell.no). instances obtained through phone calls. Originally 88 hotels came up on this list. Seven It should be mentioned that the attributes of these lodging premises could, however, not considered in this study also included the be classified as hotels and were therefore features that are regularly quality-tested discarded from the sample. Furthermore, in Norway’s leading financial newspaper, because an important aspect of hedonic price Dagens Næringsliv (DN). Two attributes — modelling is to make sure that the data TV in hotel room and Internet access — where are homogenous enough to make relevant present in 72 out of the 74 hotels. As these comparisons, seven more hotels that could be two attributes for all practical purposes classified as either very influential cases or as came close to being constants, they were not outliers were deleted from the sample. Thus, included among the independent variables/ the present analysis is restricted to 74 hotels. attributes in the analysis. It should be noted that Arguably these data come closer to a popula- the mean values of the binary independent tion than to a sample. Nevertheless, the use of variables in Table 1 indicate the proportion in significance tests was deemed appropriate in the the data where the attribute of interest is statistical analysis since there were few obvious present. & 2007 Palgrave Macmillan Ltd, 1476-6930 $30.00 Vol. 5, 4 315–323 Journal of Revenue and Pricing Management 317
Examining the determinants of room rates for hotels Table 1: Descriptive statistics for study variables (N=74) Variable Description of variable/attribute Mean Standard deviation RSING Room rate per night for single room (NOK) 1001 295.8 RDOUB Room rate per night for double room (NOK) 1247 360.9 LogRSING RSING, logged 6.86 0.331 LogRDOUB RDOUB, logged 7.08 0.306 CHAINa Hotel is associated with a chain (yes=1) 0.810 0.394 MINIBARa Mini-bar is present in hotel room (yes=1) 0.757 0.431 POOLa Pool is present in hotel (yes=1) 0.176 0.383 PARKFREEa Free parking space is present in hotel (yes=1) 0.500 0.503 RESTAUa Restaurant is present in hotel (yes=1) 0.649 0.480 HAIRDRYa Hairdryer is present in hotel room (yes=1) 0.838 0.371 ROOMSERVa Room service is present in hotel (yes=1) 0.338 0.476 DIST Distance in kilometers to Oslo Central Station 11.89 14.64 BEDS Number of beds in hotel 257.7 191.5 a Indicates a binary variable (ie, an attribute). The mean value refers to the proportion in the data where the attribute of interest is present. Estimation superior compared with two sequential OLS Most previous price hedonic research has regressions in terms of obtaining statistical followed Rosen’s (1974) original advice and efficiency (Zellner, 1962). Thus, the estimation used the log-linear specification instead of the of the two regression equations was carried out linear one. Following convention, therefore, by means of Zellner’s seemingly unrelated the natural logarithms of the two room rates regression (SUR) approach. Another advantage were used as the dependent variables in this of the SUR technique is that it permits testing study. An additional advantage of the log-linear for significant differences in regression coeffi- form compared to the linear, is the former cients between the two equations. technique’s ease of interpretation. In general log-linear regression coefficients can be inter- Results preted as the percentage change in the The results of Table 1 show that the mean rate dependent variable associated with a one-unit for a one-night stay in a single room is 1,001 increase in the independent variable. Unfortu- Norwegian crowns (NOK). The similar nately, dummy coefficients and large coeffi- rate for a double room is 1,247 NOK (as of cients (ie, coefficients >0.20) do not permit this 11 May 2005, 1,000 NOK was the equivalent straightforward interpretation. In this case, the of 80.09 h). The respective median values are percentage difference between the character- 1,000 and 1,250 (results not shown). The istic of interest and the reference category is bivariate correlation between the two price obtained by taking the antilog of the coefficient variables is 0.89 (Po0.01). In other words, the minus 1 (Halvorsen and Palmquist, 1980). rate differences between single and double A priori it was expected that the two rooms are fairly constant for the hotels in dependent price variables in this study were question. Table 2 shows how the two room likely to be correlated. In such instances, a joint rates are affected by the various attributes listed estimation of the two regression equations is in Table 1. 318 Journal of Revenue and Pricing Management Vol. 5, 4 315–323 & 2007 Palgrave Macmillan Ltd, 1476-6930
Thrane Starting with the single room rates (ie, As to the coefficients for double room prices LogRSING), we see that the room rates in (ie, LogRDOUB), we see that, contrary to the hotels associated with chains are about 15% results for single rooms, a chain association does more expensive than the room rates in non- not have an effect on the rates for double chained hotels, ceteris paribus. Similarly, the rooms. By contrast, the effects of MINIBAR, rates for rooms that include mini-bars are about PARKFREE, HAIRDRY and ROOMSERV 39% higher than the rooms not offering are very similar for single and double room this facility (exp. (0.329)1 ¼ 0.389). Also, rates. In passing, it is also interesting to note room rates in hotels where free parking is that the effect of MINIBAR on LogRSING is available are about 19% higher than they are in significantly larger (F ¼ 3.04, Po0.10) than the hotels that do not provide such a service, analogues effect on LogRDOUB. The greater whereas hairdryers in hotel rooms drive up the the distance of hotels from Oslo Central Station room rates by 44% compared with rooms (ie, downtown Oslo), the lower the prices for where hairdryers are not available (exp. double rooms. For example, the room rate in a (0.364)1 ¼ 0.439). Finally, room rates are, hotel located 30 km from Oslo Central Station perhaps somewhat surprisingly, about is approximately 11% lower than the rate in a 12% lower in hotels offering room service hotel located only 10 km from this location, than those that do not provide this feature. ceteris paribus (0.056 2 ¼ 0.112) (it is The following attributes have no effect on also worth mentioning that the combined the rates for single rooms: presence of swim- effect of DIST on LogRSING and Log- ming pool in hotel (POOL), presence of RDOUB is statistically significant (F ¼ 7.58, restaurant in hotel (RESTAU), distance to Po0.01)). All else being equal, the room rates Oslo Central Station (DIST) and number of in larger hotels (as indicated by a greater beds in hotel (BEDS). number of beds) are a bit more expensive than Table 2: Room rates for one-night stay in single room (logged) and double room (logged) by independent variables. Seemingly unrelated regression (N=74). Independent LogRSING LogRDOUB F-test of model differences variable Coefficient Standard error Coefficient Standard error CHAIN 0.147* 0.076 0.013 0.070 14.50*** MINIBAR 0.329*** 0.065 0.266*** 0.059 3.04* POOL 0.065 0.068 0.075 0.062 0.804 PARKFREE 0.193*** 0.047 0.186*** 0.043 0.794 RESTAU 0.005 0.059 0.021 0.054 0.613 HAIRDRY 0.364*** 0.076 0.387*** 0.070 0.584 ROOMSERV 0.123** 0.051 0.127*** 0.047 0.896 DIST ( 10) 0.026 0.016 0.056*** 0.015 10.90*** BEDS ( 100) 0.003 0.014 0.022* 0.013 5.58** Constant 6.13*** 0.068 6.50*** 0.063 91.40*** R2 0.703 0.705 Note: see Table 1 for variable definitions and descriptions. Breusch-Pagan test of independence between regression equations: w2 (1 df)=51.75; Po0001, *Po10, **Po05, ***Po01 (two-tailed tests). & 2007 Palgrave Macmillan Ltd, 1476-6930 $30.00 Vol. 5, 4 315–323 Journal of Revenue and Pricing Management 319
Examining the determinants of room rates for hotels those in smaller hotels. For example, the room variables are set to their modal values, whereas rate in a hotel with 500 beds is about 9% higher BEDS is set to its mean value (see Wooldridge than the rate in a similar hotel with 100 beds (2000, p. 203–204) on how to transform a (0.022 4 ¼ 0.088). It is also worth pointing logged dependent variable into a level-form out that the effects of DIST and BEDS are dependent variable). significantly different for the two regression The explained variances in both of the equations (F ¼ 10.90, Po0.01 for DIST; regression equations reported in Table 2 F ¼ 5.58, Po0.05 for BEDS). Just as for the (R2 ¼ 0.70) show acceptable model fit. Yet, statistical explanation of LogRSING, the since the SUR regression was based on some- independent variables POOL and RESTAU what few observations, a number of diagnostic have no effects on LogRDOUB. Finally, the tests were carried out in order to examine the Breusch-Pagan statistic is clearly significant. robustness of the findings. First, insignificant This confirms that LogRSING and Log- RESET tests for misspecification (P>0.05 for RDOUB are indeed correlated, and that the both equations) indicated that the models two regression equations are not independent lacked no important explanatory variables. of each other. In this way, the use of the SUR Second, insignificant Breusch-Pagan tests for approach is clearly justified. heteroskedasticity (P>0.05 for both equations) Especially two attributes have pronounced suggested that the variance in the data did not effects on the rate for a stay in a double room: increase (or decrease) with price. Third, MINIBAR and HAIRDRY. Figure 1, which both plots (available on request) and formal also includes the variable DIST, displays this tests (Shapiro-Wilk W; P>0.05 for both more clearly: The room rate in a hotel located equations) suggested that a normal distribution 50 km from Oslo Central Station in which for the residuals could not be rejected for any neither a mini-bar nor a hairdryer are present is of the models. Fourth, multicollinearity is about 650 NOK. By contrast, the analogues often a concern in regression models where room rate in a hotel situated 10 km from Oslo many binary independent variables are present. Central Station in which both of these As no Variance Inflations Scores (VIF) ex- accessories are present is about 1,600 NOK. ceeded 2.06 for any of the independent In Figure 1, the remaining binary independent variables and the mean VIF-score was 1.63, multicollinearity was not a problem in the present analysis. Finally, the estimations 1800 were also carried out using the so-called ‘robust 1600 regression’ procedure, and the results were 1400 similar to the reported ones. Thus, the results Room rates for double rooms 1200 of these tests were all satisfactory. 1000 800 DISCUSSION AND CONCLUSION 600 During the last 15 years a number of hospitality 400 and tourism studies have been concerned about 200 how the utility-bearing attributes embedded in 0 lodging facilities are associated with the overall 0 10 20 30 40 50 price for such products. The theoretical frame- Distance to Oslo Central Station in kilometers work for these studies has generally been the No mini-bar, no hairdryer Mini-bar, no hairdryer seminal work by Rosen (1974) on price No mini-bar, hairdryer Mini-bar and hairdryer hedonic theory. In general the results of this Figure 1: Room rates for double rooms by three type of research tell consumers which attributes hotel attributes they have to pay extra for, which attributes 320 Journal of Revenue and Pricing Management Vol. 5, 4 315–323 & 2007 Palgrave Macmillan Ltd, 1476-6930
Thrane result in discount, and which attributes located farther away from Oslo downtown (ie, do not affect the price. At the same time, Oslo Central Station). Espinet et al. (2003) also however, the providers of hospitality and found that parking was associated with hotel tourism products may use such results room prices. as a means for strategic pricing. The ambition Variables measuring aspects of location have of the present research has been to examine in most of the prior research shown predictive how certain hotel attributes are related ability with respect to lodging prices (Bull, to the room rates for single and double rooms 1994; White and Mulligan, 2002; Monty and in the region surrounding Norway’s capital, Skidmore, 2003; Thrane, 2005). In this study, a Oslo. Most of the findings of this study locational effect was supported in the sense that have been in general accordance with previous longer distances to Oslo Central Station (ie, studies. However, some results were also downtown Oslo) were associated with less unexpected. Below follows a recapitulation of expensive room rates (cf. Figure 1). Prior the main findings and some comments in this studies have also found that hotels associated regard. with a hotel chain (ie, corporate affiliation) The two most important attributes in terms charge higher room rates than hotels with no of explaining variance in room rates were the such associations (Wu, 1998; Israeli, 2002). The presence or absence of the amenities mini-bars results of this study partially supported this and hairdryers, as Figure 1 displays. It is also notion, as a chain-association only affected the interesting to note that these two attributes are room rates in the expected way for single the only ones that refer to the actual hotel rooms. rooms, and not to the hotels themselves. Another novel, and to some extent un- While the effect of hairdryers must be expected, result of this study was the effect of considered as a novel feature of this room service on room rates. In this respect, the study, Aguilo et al. (2001, 2003) also observed room rates in hotels offering room service were the effect with respect to mini-bars. The about 12% less expensive than similar rates in next interesting question to ask is whether or hotels not providing this service. At this point, not consumers have the opportunity to the question of why this pattern was observed is choose between hotels that do, or do not, only a matter of speculation. A preliminary offer these amenities in their rooms. In conjecture would be that hotels offering room other words, do consumers have real choices service anticipate that the income generated by as regards saving money by choosing a this service will likely compensate for the lower hotel that does not offer these facilities? A room rates. In addition, lower room rates will quick glance at the data suggests that consumers by themselves likely attract more customers. for most practical purposes face such a choice Obviously, however, more research is needed to a somewhat limited degree, since as many 76 before any certain conclusions can be drawn in and 84% of the hotels in the data offer, this regard. respectively, mini-bars and hairdryers in their The following attributes did not appear to rooms. have a bearing on room rates in any of the Free parking was another attribute with a models: presence of swimming pool, presence noticeable effect on room rates in this study. In of restaurant and, for all practical purposes, all, 50% of the hotels in the data offered this hotel size (ie, number of hotel beds). As these facility, so in this regard consumers have a real characteristics have all been identified as salient choice as to saving money if they want to. in many prior studies, the lack of these However, further analysis (results available on associations also beg an explanation. Based on request) showed that presence of free parking the reasonable assumption that the hotels was more frequently associated with hotels analysed in this study are characterised by a & 2007 Palgrave Macmillan Ltd, 1476-6930 $30.00 Vol. 5, 4 315–323 Journal of Revenue and Pricing Management 321
Examining the determinants of room rates for hotels stronger element of business stays than the ones REFERENCES analysed in previous research (which has had an Aguilo, E., Alegre, J. and Sard, M. (2003) orientation towards vacation settings), a quali- ‘Examining the market structure of the German fied guess could be that these attributes are and UK tour operating industries through an simply more relevant to the vacation travel analysis of package holiday prices’, Tourism market. Economics, 9, 255–278. Some differences between the regression Aguilo, P. M., Alegre, J. and Riera, A. (2001) models for single and double room rates were ‘Determinants of the price of German tourist found. One possible explanation for the packages on the Island of Mallorca’, Tourism divergences between the two regression mod- Economics, 7, 59–74. els’ results could be that the ratio between Andersson, D. E. (2000) ‘Hypothesis testing in hedonic single and double room offerings varies among price estimation – on the selection of independent variables’, Annals of Regional Science, 34, 293–304. the hotels in the data. Another possibility is that Bull, A. O. (1994) ‘Pricing a motel’s location’, these differences could reflect that hotel or International Journal of Contemporary Hospitality hotel chain directors have different anticipa- Management, 6, 10–15. tions as regards the demand for single or double Carvell, S. A. and Herrin, W. E. (1990) ‘Pricing in rooms. Future research should address this the hospitality industry: an implict market more carefully. approach’, Florida International University Hospi- This study suffered from two main limita- tality Review, 8, 27–37. tions. First, the analysis was based on a limited Coenders, G., Espinet, J. M. and Saez, M. (2003) number of observations. Nevertheless, the ‘Predicting random level and seasonality of hotel explanatory power of this study’s multivariate prices: a latent growth curve approach’, Tourism models was in line with most previous research. Analysis, 8, 15–31. Furthermore, a number of diagnostic tests Combris, P., Lecocq, S. and Visser, M. (1997) ‘Estimation of a hedonic price equation for revealed no problems with the model specifica- bordeaux wine: does quality matter?’ The tions. The second limitation, which is more Economic Journal, 107, 390–402. severe, was related to the possibility for general- Espinet, J. M., Saez, M., Coenders, G. and Fluiva, isations to hotel areas around other capitals. A M. (2003) ‘Effect on prices of the attributes of priori it is difficult to have an informed opinion holiday hotels: a hedonic price approach’, in this regard; this is something future research Tourism Economics, 9, 165–177. should shed more light on. Fujita, M. (1989) Urban Economic Theory, Land Use and In conclusion, this study has within a price City Size, Cambridge University Press, Cambridge. hedonic framework examined how certain Halvorsen, R. and Palmquist, R. (1980) ‘The hotel attributes are related to the room rates interpretation of dummy variables in semiloga- in and around Norway’s capital, Oslo. Using a rithmic equations’, American Economic Review, seemingly unrelated regression (SUR) techni- 70, 474–475. que, a number of associations between hotel Henkel, R. E. (1976) Test of Significance, Sage attributes and room rates were observed. In Publications, Newbury Park, CA. particular, room amenities such as mini-bars Israeli, A. A. (2002) ‘Star rating and corporate affiliation: their influence on room price and and hairdryers drove up the room rates in the performance of hotels in Israel’, International same way as free parking. Furthermore, a Journal of Hospitality Management, 21, 405–424. location close to downtown Oslo meant that Jeffrey, M. Wooldridge (2000) Introductory Econo- consumers must pay more for hotel accom- metrics. A Modern Approach, South-Western modation. By contrast and somewhat surpris- College Publishing, USA. ingly, the room rates in hotels offering room Lancaster, K. J. (1966) ‘A new approach to service were lower than the rates in hotels not consumer theory’, Journal of Political Economy, providing this service. 74, 132–157. 322 Journal of Revenue and Pricing Management Vol. 5, 4 315–323 & 2007 Palgrave Macmillan Ltd, 1476-6930
Thrane Monty, B. and Skidmore, M. (2003) ‘Hedonic Marketing Tourism Places, Routledge, London, pricing and willingness to pay for bed and 85–103. breakfast amenities in Southeast Wisconsin’, Thrane, C. (2005) ‘Hedonic price models and sun- Journal of Travel Research, 42, 195–199. and-beach package tours: the Norwegian case’, Papatheodorou, A. (2002) ‘Exploring competitive- Journal of Travel Research, 43, 302–308. ness in mediterranean resorts’, Tourism Economics, White, P. J. and Mulligan, G. F. (2002) 8, 133–150. ‘Hedonic estimates of lodging rates in the four Rosen, S. (1974) ‘Hedonic prices and implicit corners region’, The Professional Geographer, 54, markets: product differentiation in pure compe- 533–543. tition’, Journal of Political Economy, 82, 34–55. Wooldridge, J. M. (2000) Introductory Econo- Roubi, S. and Litteljohn, D. (2004) ‘What metrics. A Modern Approach, South-Western makes hotel values in the UK? A hedonic College Publishing, USA. valuation approach’, International Journal of Wu, L. (1998) ‘The pricing of a brand-name Contemporary Hospitality Management, 16, product: franchising in the motel service in- 175–181. dustry’, Journal of Business Venturing, 14, 87–102. Sinclair, M. T., Clewer, A. and Pack, A. (1990) Zellner, A. (1962) ‘An efficient method of estimat- ‘Hedonic prices and the marketing of ing seemingly unrelated regressions and tests for package holidays: the case of tourism resorts in aggregation bias’, Journal of the American Statistical Malaga’, in Ashworth, G. and Goodall, B. (eds) Association, 99, 348–366. & 2007 Palgrave Macmillan Ltd, 1476-6930 $30.00 Vol. 5, 4 315–323 Journal of Revenue and Pricing Management 323
You can also read