Asymmetric Tourist Response to Price

Abstract

The objective of this article is to look into heterogeneity in loss aversion in order to detect how dispersed loss aversion is in tourism and to observe whether different degrees of loss aversion can lead to the identification of loss-aversion–based segments. Loss aversion is a prominent psychological human trait that causes asymmetric price reactions. Tourism literature has shown it is a critical characteristic with a significant, but heterogeneous, effect on tourist destination choice. However, so far no attempt has been made to look into loss aversion heterogeneity, and this article contributes to the literature by exploring, for the first time, the potential existence of groups of tourists that show differentiated asymmetric responses to price. The empirical application estimates the individual degree of loss aversion for each tourist, and detects five segments with different sensitivities. Relevant managerial implications are drawn in terms of implementing pricing strategies.

Keywords

segmentation loss aversion heterogeneity destination choice price

Introduction

There is an increasing body of evidence supporting the idea that the carriers of utility are generally not states but rather changes relative to a reference point (Köbberling, Schwieren, and Wakker 2007; Masiero and Hensher 2010, 2011). This is a basic pillar of Prospect theory, which posits that people evaluate outcomes not on their absolute level but rather on their deviation from some reference level (Kahneman and Tversky 1979). Derived from this theory, loss aversion has been one of the most relevant psychological effects in marketing: evidence of loss aversion implies that changes from reference points may be valued differently depending on whether they are gains or losses; in particular, people tend to be more sensitive to losses relative to their reference point than to gains. In practical terms, this property has important implications in markets in which individuals show themselves as loss averse: given that their final choice is greatly influenced by it, organizations can develop actions based on this phenomenon (e.g., by implementing activities to modify their reference points). Among the different approaches to loss aversion, pricing is the most notable research field in marketing (Klapper, Ebling, and Temme 2005).

In tourism, this interest is not an exception. On one hand, implications of prices are multiple because of their influence on current issues such as competitiveness (Hiemstra and Wong 2002; Lee 1996; Tan, McCahon, and Miller 2002; Stevens 1992), business-to-business relationships (Law and Lau 2004; Pearce 2007), future intentions such as repurchase (Feng, Cai, and Zhu 2006; Kao et al. 2008; Miller and Grazer 2003; Petrick, Morais, and Norman 2001; Petrick and Backman 2002; Petrick 2004; He and Song 2009), the design of bundling strategies (Hooper 1995), and their dual function in terms of sacrifice incurred and quality expected (Murphy and Pritchard 1997; Park et al. 2007); and on the other hand, today’s research trends tend to introduce psychological elements in the traditional, expected rational consumption of tourism, such as risk and uncertainty avoidance (Quintal, Lee, and Soutar 2010), or more controversially, the introduction of emotional aspects like “affect” that lead traditional cognitive-dominated theories to be revised (Walls, Okumus, and Wang 2011).

Therefore, there is a body of tourism research that has focused on the psychological impact of prices (Al-Sabbahy, Ekinci, and Riley 2004; Crompton and Love 1994; Lawson, Gnoth, and Paulin 1995). In this context, the idea of potentially asymmetric responsiveness to prices derived from the existence of reference points (i.e., price loss aversion) is especially relevant as the increased promotional activity by competing destinations is likely to raise price elasticities (Crouch 1994) and, as Kim and Crompton (2002) indicate, the challenge is to find ways to reduce individuals’ reluctance to accept increases in price. In this line, several authors have examined loss aversion in tourism: Oh (2003) analyzes room prices of a specific upscale hotel operating in a U.S. city. This author estimates reference prices as the average value of (1) the fair rate suggested by the sample individuals for the room in which they were staying and (2) the mean market room rate for hotels similar to the sample hotel (also estimated by the sample individuals). This study does not find that asymmetric effects of price deviations exist in individual judgments of price perceptions. Although the study of Kim and Crompton (2002) in the context of admission fees to a Texas state park is based on reference prices, they do not estimate them. They operationalize the perceptions of the admission price to a Texas state park by asking sample visitors whether a specific admission fee is “much too low, too low, about right, too high, much too high” and recode them on a 5-point Likert-type scale. They use this measure as the dependent variable in a regression model so as to find which independent variables have an influence on it. Their main result is that economic factors are better explanatory variables for perceptions of admission price than behavioral factors. Nicolau (2008) tests the existence of reference dependence and loss aversion in Spanish tourism, showing that tourists use reference prices to make their decisions; that is, they take into account the magnitude of the difference between reference price and actual price, rather than absolute prices. This study also finds that tourists react more strongly to price increases than to price decreases relative to the reference price, which represents evidence in favor of loss aversion. It is important to note, however, that this author finds significant heterogeneity parameters, which reveals the existence of different degrees of tourist loss aversion.

Precisely, heterogeneity has been claimed to be a critical aspect in the analysis (and detection) of loss aversion, not only in tourism but in general consumption too. Klapper, Ebling, and Temme (2005) empirically demonstrate that the inconsistencies found in empirical research on the relative size of loss aversion (e.g., Hardie, Johson, and Fader [1993] find evidence of loss aversion, while Kalyanaram and Little [1994] do not) are a consequence of not adequately accounting for consumers’ heterogeneity in their response.

Stemming from previous empirical evidence of tourist loss aversion found by Nicolau (2008), the aim of this article is to look into heterogeneity in loss aversion in order to detect how dispersed loss aversion is in tourism and to observe whether different degrees of loss aversion can lead to the identification of loss-aversion-based segments. To the author’s knowledge, this is the first attempt to segment the tourism market through loss aversion intensity.

This study argues that the existence of the aforementioned heterogeneity in tourism demand, through which individuals display different behaviors (Dolnicar and Grün 2008), has led segmentation to become a strategic approach and a critical long-term tool for destination decision makers (Dolnicar 2004; Dolnicar and Leisch 2003; Jeng and Fesenmaier 2002; Laesser and Crouch 2006; Sarigöllü and Huang 2005). Paralleling this practical interest, academics have looked into heterogeneous preferences from different perspectives to segment the tourism market, such as demographics (Mudambi and Baum 1997), behavioral dimensions (Pritchard and Howard 1997; Laesser and Crouch 2006), psychographics (Sirakaya, Uysal, and Yoshioka 2003), and benefits (Kastenholz, David, and Paul 1999). Note, however, that the formation of segments through price sensitivities is scarcer, and no attempt has been made to segment the market through loss aversion. The importance of segmenting the tourism market through loss aversion revolves around the idea that promotions are mainly based on price comparisons like the “was” and “now” or a comparison to a competitor. Knowing which segment can be more influenced by these comparisons would allow managers to clearly determine their target group.

Research Design

Method

The method that allows us to estimate individual loss aversions based on real decisions consists of two stages: (1) estimation of individual loss aversion through a logit model with random coefficients and (2) application of a cluster analysis.

Estimation of Individual Loss Aversion

To analyze reference dependence and loss aversion in an empirical setting, it is necessary to specify the utility structure in such a way that it depends on gains and losses relative to a reference point. Following the standard procedures by Bell and Lattin (2000) and Klapper, Ebling, and Temme (2005), the utility function U_int for alternative i and individual n on occasion t is expressed as

U_{int} = α_{i} + β_{n} G A I N_{int} + γ_{n} L O S S_{int} + ε_{int},

where, RP_nt is the reference price for individual n on occasion t and PRICE_it is the actual price of alternative i on occasion t, GAIN_int and LOSS_int are defined as follows: GAIN_int = (RP_nt- PRICE_it)D₁, where D₁ = 1 if RP_nt- PRICE_it > 0 and D₁ = 0 otherwise; and LOSS_int = (RP_nt – PRICE_it)D₂, where D₂ = 1 if RP_nt – PRICE_it < 0 and D₂ = 0 otherwise. Note that the prices of all alternatives are compared to a common reference price RP_nt for each individual, as each person has one reference point for all the alternatives (Tversky and Kahneman 1991). Finally, α_i, γ _n , and β _n are coefficients¹ to be estimated and ε _int is a random term. Loss aversion is detected if γ _n /β _n >1; that is, if the parameter associated with losses is greater than the parameter related to gains.

We assume that ε _int is a random term that is IId extreme value, so that logit-type models are used (Costa and Manente 1996; Tsaur and Wu 2005; Winzar, Pidcock, and Johnson 1993). In particular, we use the random parameter logit model (RPL) because of the following reasons: (1) as we are interested in modeling heterogeneity, the RPL explicitly models price response heterogeneity and, in line with Klapper, Ebling, and Temme (2005), allows us to account for heterogeneity to the fullest possible extent; (2) RPL models permit the estimation of the proportion of sample individuals who show positive or negative preferences toward an attribute—parameters greater or lesser than zero, respectively—through the normalization b/√W ~ N(0, 1) pointed out by Train (1998), where b and W are the mean and variance of the normal distribution φ(β _nh | b, W), with β _nh being the parameter for individual n that measures the effect of attribute h; (3) this model allows us to estimate individual parameters for each subject in the sample; and (4) RPL models do not have the restrictive substitution patterns of the traditional logit model, thus avoiding the assumption of Independence from Irrelevant Alternatives (IIA). Note in the formula below that the ratio of probabilities P(i)/P(j) depends on all the data and, contrary to the traditional logit, this ratio includes the attributes of alternatives other than i and j.

The introduction of heterogeneity in the choice model leads coefficients θ to vary over decision makers with density f(θ) and θ is not observable, the probability P_nt(i) of an individual n choosing alternative i on occasion t is the integral of P_nt(i/θ) over all the possible values of θ:

P_{n t} (i) = \int_{θ} \frac{exp {U_{int}}}{\sum_{j = 1}^{J} \exp {U_{j n t}}} φ (θ | b, W) d θ,

where J is the number of alternatives and φ is the density function of θ, assuming that θ is distributed normal with average b and variance W. To estimate the individual loss aversion parameters, we apply Bayesian estimation methods.

Cluster Analysis

We use the individual loss aversion parameters as inputs for a subsequent cluster analysis. To this end, we apply the Ward hierarchical algorithm to the matrix of these individual parameters. The final number of segments is reached when the segments observed explain at least 65% of the global variance and, when another segment is added, the increase in the total variance is less than 5% (Lewis and Thomas 1990). According to Grande and Abascal (2009), Gené (2002) and Sorensen (2003), this method is efficient and appropriate when using variables derived from previous statistical procedures. Also, an ANOVA is performed to confirm the segment differences.

Sample, Data, and Variables

To reach the objective, we use information on tourist choice behavior obtained from the national survey “Spanish Holidaying Behavior (III),” which was carried out by the Spanish Centre for Sociological Research. This is due to the following reasons: (1) The availability of information on individual tourist destination choice behavior in terms of types of destinations, in particular, the types “coastal” and “inland.” The examination of destination choices of a “coastal-inland” type is relevant because of the tendency of people to look for alternatives to the sun, sea, and sand–type holiday that predominates in countries like Spain. Moreover, the development of these alternatives is largely found in inland areas, as it allows a destination typically known for its coast to diversify its “product portfolio” as well as an inland economy to be revitalized. In this context, the study of prices is crucial for the development of tourism policies by public bodies and for the implementation of strategies in the tourism industry. And (2) the survey is directed at a sample (older than 18 years) obtained at each individual’s home, which avoids the characteristic selection bias of destination- collected samples, leading to a more precise analysis of tourist demand. The sample is taken by using multistage sampling, stratified by conglomerations, with proportional selection of primary units—cities—and of secondary units—censorial sections. The strata are formed by crossing the 17 Spanish autonomous regions and the city size, resulting in seven categories: less than 2,000 inhabitants; between 2,001 and 10,000; between 10,001 and 50,000; between 50,001 and 100,000; between 100,001 and 400,000; between 400,001 and 1,000,000; and more than 1,000,000. The individuals were selected through random routes and through quotas defined by gender and age. The information was collected through personal, at home, interviews with a structured questionnaire. By considering individuals who provide information on at least two consecutive holiday periods (regardless of whether they went, after the first time, on holiday or not), the sample size is of 410 individuals. There were no list of destinations constrained to certain locations; rather the respondents were free to list any destination. We base the analysis on domestic destinations because (1) considering international destinations would notably increase the number of locations to be taken into account and (2) the distinction between “coastal” and “inland” for Spanish destinations is deeply rooted in people’s minds. In fact, even though there might be differences between destinations that can be attributed to more than just being “coastal” or “inland” locations, the impact on the Spanish case is minimized on account of this traditional distinction.

Variables: To make the choice model operative, we will define the variables used and identify the dependent and independent variables.

Dependent variable

To represent the set of alternatives (destination types) available to the individual, we use the following three dummy variables: (1) coastal, which takes a value of 1 when this type of destination is chosen and 0 if not; (2) inland, where a value of 1 shows that this kind of destination has been selected and 0 if not; (3) not going on holiday (at the last vacation occasion), which takes a value of 1 when chosen and 0 if not.

Independent variables for the choice model

(a) Prices. Since the alternatives are “types of destinations” (coastal and inland) we have to build up a price index for each type. We measure prices of destination types using the specific cost index for each type of destination and each individual proposed by Eymann and Ronning (1997). The procedure used to form this index has sometimes been called “quasi-hedonic” regression technique because of its resemblance to the hedonic regression introduced by Rosen (1974). In fact, the index proposed by Eymann and Ronning (1997) is an application to tourism destinations of the well-known hedonic price index widely used in the literature in different fields (Izquierdo and Matea 2004). It implies following a two-stage procedure (Eymann and Ronning 1997): (1) A regression model is estimated $E_{INT} = δ_{i 1} + δ_{i 2} X_{INT}^{(1)} + δ_{i 3} X_{n t}^{(2)} + ε_{INT}$ where E_int are the tourism costs (expenditures) of each individual n in each destination type i on occasion t, $X_{n t}^{(1)}$ is the consumption intensity in the corresponding destination type i based on the number of days the individual n spent there on occasion t, and $X_{n t}^{(2)}$ are the sociodemographic characteristics of individual n on occasion t (household size, marriage status, education, and income). Note that although the utility function does not directly use the variable “income,” we are indirectly dealing with it by inserting income in this price index, which in turn will be ultimately introduced in the utility function. (2) The estimated parameters δ_i1, δ_i2, and δ_i3 are used to construct the specific cost indices—or quasi-hedonic prices QHP_int—for each type of destination and each individual at a specific occasion using the expression $Q H P_{INT} = {\hat{δ}}_{i 1} + {\hat{δ}}_{i 2} {\bar{X}}_{i t}^{(1)} + {\hat{δ}}_{i 3} X_{n t}^{(2)}$ where ${\bar{X}}_{i t}^{(1)}$ represents the average consumption of variable $X_{i t}^{(1)}$ in destination i in period t².

(b) Reference prices. It is important to stress that reference prices are not only quantities generally unavailable from conventional data sources, but they are difficult to measure (Winer 1986). We propose and try one internal reference price and two external reference prices, to empirically determine which one is best. We define the one internal memory-based reference price as the price a consumer paid at the last purchase incidence (Klapper, Ebling, and Temme 2005; Mazumdar, Raj, and Sinha 2005). On the other hand, we determine the two stimulus-based reference prices as (1) the current price of the last product purchased (Hardie, Johson, and Fader 1993; Bell and Lattin 2000), as it is easier for the consumer to remember the product bought at the last purchase occasion than to remember the last price paid, and (2) the average of the current prices of the available alternatives (Moon, Russell, and Duvvuri 2006), as individuals may observe to what extent a price stands out in comparison with other product prices.

Regarding their measurement, the reference prices for the destination types coastal and inland are, as in the case of prices, measured using the quasi-hedonic prices of Eymann and Ronning (1997) obtained from the two-stage procedure laid out before. Note that by using this technique, we are able to estimate the price QHP_int for each destination type i, each individual n and every purchase occasion t. Therefore, the internal reference price, defined as the price a consumer paid at the last purchase incidence is expressed as RP_nt= QHP_jnt-₁, where j is the alternative bought at the last occasion; the external reference price defined as the current price of the last alternative purchased as RP_nt= QHP_jnt; and the external reference price defined as the average of the current prices of the available alternatives as $R P_{n t} = {\bar{Q H P}}_{n t} .$

(c) Descriptive variables for the segments: (1) Income—monthly income levels are placed into the following categories: Income 1, up to €600 per month; Income 2, between €600 and €1200; Income 3, between €1200 and €2400; Income 4, between €2400 and €4500; and Income 5, more than €4500. (2) Age—apart from its quantitative measurement, four categorical variables are used: Age 1, less than 25 years old; Age 2, between 26 and 45; Age 3, between 46 and 65; and Age 4, more than 65 years old. (3) Household size—this is measured by the number of people living in the house. (4) Single—whether the interviewee is single or not. (5) Secondary home—whether the interviewee owns a secondary home or not. (6) Organization—whether the tourist uses a travel agent and 0 if he or she organizes his or her own vacation. (7) Length of stay—a quantitative variable of the number of days that a tourist spends outside the usual place of residence. (8) Accommodation type—the type of accommodation selected by the tourist is classified as “hotel,” “camping site,” “own apartment or villa,” “rented apartment or villa,” and “family or friends’ house.” (9) The variable relative to tourism expenditures is found by a quantitative variable that represents costs incurred during the holiday. (10) The motivations “resting,” “climate,” “cultural interest,” “interest in new places,” “doing sports,” and “visiting friend and relatives” are measured through dummy variables.

Results

The first action is to empirically test the best reference price alternative. We estimate—with each of them—the equation (2), which incorporates the effects of gain and loss. Regarding the internal reference price, Price paid at the last purchase incidence, the likelihood function is −408.54. As for external reference prices, Current price of the last product purchased and Average of current prices of the available alternatives, the likelihood functions arrived at are −413.13 and −418.52, respectively. These results show that the internal reference price measured by the price paid at the last purchase incidence presents the best fit (this superiority of the internal reference price is also found when quadratic terms are introduced (see Nicolau [2008]). This result is in accordance with the widely found evidence that the last price paid takes part in the formation of the reference price (Briesch et al. 1997; Kalyanaram and Winer 1995; Mazumdar, Raj, and Sinha 2005). Having empirically determined that in this application, the internal memory-based reference price is best, Table 1 presents the global results for an average tourist. The parameter estimates refer to the utility function shown in the method section, where the gain and loss parameters are assumed to follow a normal distribution; accordingly, the second row in Table 1 presents their mean parameters and the third row their standard deviations (the parameters of the constants are assumed to be fixed).

Table 1.

Loss Aversion Estimates (STANDARD ERROR IN PARENTHESES)

	Gain	Loss	Coastal Constant	Inland Constant	ML
Parameter	0.002 (0.004)	0.021^* (0.006)	1.205^* (0.263)	0.955^* (0.265)	−408.54
Standard Deviation of β	0.027^** (0.011)	0.028^* (0.006)	−	−

p < .001, **p < .05.

The global result for an average tourist shows that the mean loss aversion is 0.021 and a standard deviation standing at 0.028, both being significant at 1%. Concerning the parameter associated with gains, it is not significantly different from zero. The fact that the loss parameter is greater than the gain parameter supports the idea that tourists are loss averse. In real terms, this means that when individuals encounter actual prices above their reference prices they opt for a cheaper alternative (for a broader discussion of this result, see Nicolau [2008]).

At this point, however, the important result that we are focusing on and further exploring is the significance of the standard deviations SD(β) of loss aversion. Its significance shows that the loss aversion parameter follows a distribution, wherein the theoretical proportions of people that have a higher-than-zero parameter stands at 77.3% [φ(0.021/0.028) = 0.773], and the effect of the loss aversion parameter is different for each individual, which reflects the existence of heterogeneity in price responsiveness to the negative difference between reference and actual prices. By applying Bayesian estimation methods we estimate the individual parameters of loss aversion³, which are used as inputs for the cluster analysis.

To this end, we apply the Ward hierarchical algorithm to the matrix of the positive loss aversion parameters of each individual (as the negative parameters represent a loss aversion reversal, we treat this segment separately). By applying the two criteria described in the method (explanation of at least 65% of the global variance and increase in less than 5%), we select five segments. Table 2 summarizes the results of the application; the shaded area represents the number of segments selected. Note, however, that one segment contained only one individual (clearly an outlier), so we have added this subject to the nearest segment. Therefore, we are left with four segments with a positive loss aversion parameter, plus the segment with a negative loss aversion parameter.

Table 2.

Segments Based on Loss Aversion

No. of Segments	σ ² ^a	σ ² (%)^a	Explained Variance	Δσ ² ^a
10	0.005	0.718	0.226	99.28
9	0.007	0.945	0.404	99.05
8	0.010	1.349	0.573	98.65
7	0.014	1.922	1.018	98.07
6	0.021	2.941	2.201	97.05
5	0.037	5.142	3.718	94.85
4	0.064	8.860	15.889	91.14
3	0.180	24.749	29.041	75.25
2	0.391	53.790	46.210	46.21
1	0.727	100	0	0

Intragroup variance.

These five segments are described in Table 3 and are significantly distinct at a level of 1% (F = 74.27). This confirms the existence of differences in the price responsiveness of the sample individuals. In particular, and focusing on the input used to obtain the segments (i.e., the average loss aversion identified for each cluster [second column in Table 3]), there are 28% of the sample with reverse loss aversion and 72% with positive loss aversion (quite in line with the theoretical proportion of 77.3%). The latter can be divided into four groups depending on the degree of loss aversion intensity: very little loss aversion (13.4%), little loss aversion (9.7%), normal loss aversion (45.2%), and high loss aversion (3.7%).

Table 3.

Segment Characterization through Loss Aversion Degree

Segment	Average Loss Aversion	Number of Tourists	Percentage
Reverse loss aversion	−0.114	61	28.0
Very little loss aversion	0.006	29	13.4
Little loss aversion	0.054	21	9.7
Loss aversion	0.106	98	45.2
High loss aversion	0.290	8	3.7

Table 4 describes the descriptive statistics for each segment, in terms of demographics, tourist behavior, and motivations. Among the four demographics, income is not significantly different among the segments but age, household size, and being single show differences in loss aversion sensitivity: the youngest group is the “normal loss aversion” segment, who live in the largest household and has a greater proportion of singles; and the oldest group is the “high loss aversion” segment, with the lowest number of people at home and the lowest proportion of singles.

Table 4.

Segment Characterization by Demographics, Tourist Behavior, and Motivations

	Reverse LA	Very Little LA	Little LA	Normal LA	High LA	Global	χ^2a
Segment characterization by tourist demographics
Income 1	13%	23%	6%	12%	0%	13%	8.91
Income 2	43%	27%	47%	47%	60%	44%
Income 3	35%	50%	35%	35%	40%	37%
Income 4	9%	0%	12%	6%	0%	7%
Income 5	0%	0%	0%	0%	0%	0%
Age	41.56	41.55	44.52	34.21	49.63	38.82	4.72^b
Age 1: less than 25	21%	21%	19%	29%	13%	24%	30.45^b
Age 2: between 26 and 45	34%	45%	38%	52%	13%	43%
Age 3: between 46 and 65	33%	24%	19%	16%	75%	24%
Age 4: Over 65	11%	10%	24%	3%	0%	8%
Household size	3.51	3.28	3.05	3.87	3	3.58	2.72^c
Single	33%	31%	29%	45%	25%	37%	19.07^d
Segment characterization by tourist behavior							χ²
Own organization	83%	83%	94%	83%	80%	85%	1.40
Length of stay (days)	12.95	10	10.83	11.18	11.63	11.62	0.123
Accommodation type 1: hotel	28%	41%	29%	40%	40%	34%	19.50
Accommodation type 2: camping site	7%	14%	6%	11%	-	9%
Accommodation type 3: own apartment	33%	5%	29%	11%	20%	22%
Accommodation type 4: rented apartment	2%	14%	-	6%	-	4%
Accommodation type 5: friends and relatives	29%	27%	35%	29%	40%	30%
Tourist expenditures, €	2365.40	506.59	325.75	675.24	168.28	1114.28	5.63^b
Segment characterization by tourist motivations
Resting	52%	55%	62%	56%	25%	54%	3.481
Climate	43%	45%	52%	41%	63%	44%	2.166
Cultural interest	20%	17%	0%	10%	13%	13%	6.720
Interest in new places	33%	41%	24%	43%	25%	37%	4.182
Doing sports	3%	17%	10%	3%	13%	6%	9.875^c
VFR	34%	21%	14%	22%	50%	26%	7.242

Test statistics for continuous variables rely on F-test (ANOVA) whereas for all the other categorical variables χ²-test applies. LA = loss aversion; VFR = visiting friends and relatives.

prob < 1%

prob < 5%

prob < 10%

As for tourist behavior variables, “organization,” “length of stay,” and “accommodation type” do not show differences. However, expenditures are clearly contingent on loss aversion sensitivity: the “reverse loss aversion” segment presents the highest level of expenses at the destination; as expected, the “high loss aversion” segment has the lowest level of expenditures. Regarding motivations, only “doing sports” shows differences, with the “very little loss aversion” segment having the highest proportion of people taking part in sports activities.

Conclusions

The prevalent use of psychological characteristics in tourism behavior and the growing relevance of segmentation on account of the extant heterogeneity have led this article to look for segments built on loss aversion sensitivities. The empirical application has been carried out by using random parameter logit models that allow the analyst to capture the individual heterogeneity in people’s choice behavior.

The results show that three quarters of the sample behave in line with the loss aversion principle when making decisions on destination types. Most important, however, is the fact that these loss-averse people display different degrees of sensitivities: very little loss aversion (13.4%), little loss aversion (9.7%), normal loss aversion (45.2%), and high loss aversion (3.7%).

In an attempt to describe these segments in terms of demographics, tourist behavior, and motivations, we find that income is not relevant to elicit differences in loss aversion, while age, household size, and being single are. Also, we observe that different sensitivities to loss aversion produce distinct levels of expenditures, and regarding motivations, only “doing sports” seems to be relevant.

These findings have relevant practical implications: (1) loss aversion is expected to appear in tourism consumption; however, consideration of its different degrees of intensity is important as it implies differentiation in certain variables. For example, we have found that expenditures at the destination vary according to loss aversion sensitivity. Consequently, detecting people with different behavioral patterns is per se fundamental for destinations, but when it comes to price-related aspects, this relevance becomes higher as it implies dealing with one of the most determining factors when choosing a destination and an important driver of income generation for the destinations. (2) For promotional price strategies, the result that tourists looked at reference points indicates that lowering price can generate new reference prices in people’s price schemes; thus, when putting the prices back up to normal levels, it could produce negative reactions in demand. Nevertheless, these negative reactions are not of the same intensity among the tourist segments. Managers are sometimes afraid to raise prices, but the potential negative impact they could expect might be overestimated as they should first look at the type of segment that goes to their destinations to gage the impact an increase in price can have.

Among the limitations of the study, three stand out: First, the use of secondary information sources, as it does not allow us to work with dimensions specific to our investigation (especially regarding alternative reference price proposals). Second, despite the deeply rooted traditional distinction between “coastal” and “inland” for Spanish destinations, some differences between both destination types might have been overlooked. Third, using price indexes for coastal and inland destinations can be problematic as prices can vary within these categories. Even though this is the only alternative we have to generate proxies of prices for “destination types” and its use is standard in the literature, it is ultimately a proxy, and as such, it is bound to have measurement errors.

For further research would remain testing the range-frequency theory by which people make judgments on a value in comparison to the most extreme values and considering its rank in a series of prices (Moon and Voss 2009; Niedrich et al. 2009). Also, recent research (Saini et al. 2010) shows that the dominance of referent thinking over relative thinking and vice versa depends on the existence and degree of the deviation between actual and expected prices; thus, it would be relevant to analyze whether the referent thinking tested in this article holds when relative thinking is included. It would provide a comprehensive perspective of tourists’ reactions, especially in the context of promotional activities. Finally, Abe (1998) suggests that the price factor in the realm of packaged goods and services is critical and the findings of people’s reactions provide marketers with distinct advantages. Therefore, identifying nonlinear responses to the price of a product included in a package can help managers make efficient decisions. Consequently, it would be relevant to explore people’s subjective price perceptions in the context of tourism packages in order to design efficient promotional programs for bundling strategies. Obviously, all of these future lines of research could be complemented by testing different “reactions to prices” in international destinations (Eugenio-Martin and Campos-Soria 2011).

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Bio

Juan L. Nicolau, PhD, is a Professor of Marketing at the University of Alicante. His research covers the analysis of tourist choice behavior and the assessment of firm value. His work has appeared in Strategic Management Journal, Omega, Journal of Travel Research, Annals of Tourism Research, Tourism Management, Tourism Economics, Tourism Geographies, European Journal of Operational Research, Journal of Business Research, European Journal of Marketing, among others.

References

Abe

(1998). “Measuring Consumer, Nonlinear Brand Choice Response to Price.” Journal of Retailing, 74 (4): 541-68.

Al-Sabbahy

H. Z.

Ekinci

Riley

(2004). “An Investigation of Perceived Value Dimensions: Implications for Hospitality Research.” Journal of Travel Research, 42 (3): 226-34.

Bell

D. R.

Lattin

J. M.

(2000). “Looking for Loss Aversion in Scanner Panel Data: The Confounding Effect of Price Response Heterogeneity.” Marketing Science, 19 (2): 185-200.

Briesch

Krishnamurthi

Mazumdar

Raj

S. P.

(1997). “A Comparative Analysis of Reference Price Models.” Journal of Consumer Research, 24 (September): 202-14.

Costa

Manente

(1996). “Venice and Its Visitors: A Survey and a Model of Qualitative Choice.” Journal of Travel & Tourism Marketing, 4 (3): 45-69.

Crompton

J. L.

Love

L. L.

(1994). “Using Inferential Evidence to Determine Likely Reaction to a Price Increase at a Festival.” Journal of Travel Research, 32(4): 32-36.

Crouch

G. I.

(1994). The Study of International Tourism Demand: A Review of Findings. Journal of Travel Research, 33 (Spring): 12-23.

Dolnicar

(2004). “Beyond ‘Commonsense Segmentation’: A Systematics of Segmentation Approaches in Tourism.” Journal of Travel Research, 42 (3): 244-50.

Dolnicar

Grün

(2008). “Challenging ‘Factor-Cluster Segmentation.’” Journal of Travel Research, 47 (1): 63-71.

10.

Dolnicar

Leisch

(2003). “Winter Tourist Segments in Austria: Identifying Stable Vacation Styles Using Bagged Clustering Techniques.” Journal of Travel Research, 41 (3): 281-92.

11.

Eugenio-Martin

J. L.

Campos-Soria

J. A.

(2011). “Income and the Substitution Pattern between Domestic and International Tourism Demand.” Applied Economics, 43 (20): 2519-31.

12.

Eymann

Ronning

(1997). “Microeconometric Models of Tourists’ Destination Choice.” Regional Science and Urban Economics, 27:735-61.

13.

Feng

Cai

L. A.

Zhu

(2006). “Long-Haul Travel Attitude Construct and Relationship to Behavior—The Case of French Travelers.” Journal of Travel & Tourism Marketing, 20 (1): 1-11.

14.

Gené

(2002). “Construcción de una Tipología y Caracterización. Aplicación a los Turistas de la Costa Dorada.” Investigación y Marketing, 74: 50-55.

15.

Grande

Abascal

(2009). Fundamentos y Técnicas de Investigación Comercial. Madrid: Esic.

16.

Hardie

B. G. S.

Johson

E. J.

Fader

P. S.

(1993). “Modeling Loss Aversion and Reference Dependence Effects on Brand Choice.” Marketing Science, 12 (4): 378-94.

17.

Song

(2009). “A Mediation Model of Tourists’ Repurchase Intentions for Packaged Tour Services.” Journal of Travel Research 47(3): 317-31.

18.

Hiemstra

Wong

K. K. F.

(2002). “Factors Affecting Demand for Tourism in Hong Kong.” Journal of Travel & Tourism Marketing, 13 (1-2): 41-60.

19.

Hooper

(1995). “Evaluation Strategies for Packaging Travel.” Journal of Travel & Tourism Marketing, 4 (2): 65-82.

20.

Izquierdo

Matea

M. L. L.

(2004). “Índices de precios hedónicos para ordenadores personales en España.” Investigaciones Económicas, XXVIII (2): 377-96.

21.

Jeng

Fesenmaier

D. R.

(2002). “Conceptualizing the Travel Decision-Making Hierarchy: A Review of Recent Developments.” Tourism Analysis, 7 (1): 15-32.

22.

Kahneman

Tversky

(1979). “Prospect Theory: An Analysis of Decision under Risk.” Econometrica, 47 (2): 263-91.

23.

Kalyanaram

Little

J. D. C.

(1994). “An Empirical Analysis of Latitude of Price Acceptance in Consumer Package Goods.” Journal of Consumer Research, 21: 408-18.

24.

Kalyanaram

Winer

(1995). “Empirical Generalizations from Reference Price Research.” Marketing Science, 14 (3): 161-69.

25.

Kao

M. C.

Patterson

Scott

C. K.

(2008). “Motivations and Satisfactions of Taiwanese Tourists Who Visit Australia: An Exploratory Study.” Journal of Travel & Tourism Marketing, 24 (1): 17-33.

26.

Kastenholz

David

Paul

(1999). “Segmenting Tourism in Rural Areas: The Case of North and Central Portugal.” Journal of Travel Research, 37 (4): 353-63.

27.

Kim

Crompton

J. L.

(2002). “The Influence of Selected Behavioral and Economic Variables on Perceptions of Admission Price Levels.” Journal of Travel Research 41(2): 144-52.

28.

Klapper

Ebling

Temme

(2005). “Another Look at Loss Aversion in Brand Choice Data: Can We Characterize the Loss Averse Consumer?” International Journal of Research in Marketing, 22: 239-54.

29.

Köbberling

Schwieren

Wakker

(2007). “Prospect-Theory’s Diminishing Sensitivity versus Economics’ Intrinsic Utility of Money: How the Introduction of the Euro Can Be Used to Disentangle the Two Empirically.” Theory and Decision, 63 (3): 205-31.

30.

Laesser

Crouch

G. I.

(2006). “Segmenting Markets by Travel Expenditure Patterns: The Case of International Visitors to Australia.” Journal of Travel Research, 44 (4): 397-406.

31.

Law

Lau

(2004). “A Study of the Perceptions of Hong Kong Hotel Managers on the Potential Disintermediation of Travel Agencies.” Journal of Travel & Tourism Marketing, 17 (2-3): 117-31.

32.

Lawson

Gnoth

Paulin

(1995). “Tourists’ Awareness of Prices for Attractions and Activities.” Journal of Travel Research, 34 (1): 3-10.

33.

Lee

C. K.

(1996). “Major Determinants of International Tourism Demand for South Korea: Inclusion of Marketing Variables.” Journal of Travel & Tourism Marketing, 5 (1-2): 101-18.

34.

Lewis

Thomas

(1990). “The Linkage between Strategy, Strategic Groups, and Performance in the UK Retail Grocery Industry.” Strategic Management Journal, 11: 385-97.

35.

Masiero

Hensher

D. A.

(2010). “Analyzing Loss Aversion and Diminishing Sensitivity in a Freight Transport Stated Choice Experiment.” Transportation Research Part A: Policy and Practice, 44 (5): 349-58.

36.

Masiero

Hensher

D. A.

(2011). “Shift of Reference Point and Implications on Behavioral Reaction to Gains and Losses.” Transportation, 38 (2): 249-71.

37.

Mazumdar

Raj

S. P.

Sinha

(2005). “Reference Price Research: Review and Propositions.” Journal of Marketing, 69 (October): 84-102.

38.

Miller

A. R.

Grazer

W. F.

(2003). “Complaint Behavior as a Factor in Cruise Line Losses: An Analysis of Brand Loyalty.” Journal of Travel & Tourism Marketing, 15 (1): 77-91.

39.

Moon

Voss

(2009). “How Do Price Range Shoppers Differ from Reference Price Point Shoppers?” Journal of Business Research, 62 (1): 31-38.

40.

Moon

Russell

G. J.

Duvvuri

S. D.

(2006). “Profiling the Reference Price Consumer.” Journal of Retailing, 82: 1-11.

41.

Mudambi

Baum

(1997). “Strategic Segmentation: An Empirical Analysis of Tourist Expenditure in Turkey.” Journal of Travel Research, 36 (1): 29-34.

42.

Murphy

P. E.

Pritchard

(1997). “Destination Price/Value Perceptions: An Examination of Origin and Seasonal Influences.” Journal of Travel Research, 35 (3): 16-22.

43.

Nicolau

J. L.

(2008). “Testing Reference Dependence, Loss Aversion and Diminishing Sensitivity in Spanish Tourism.” Investigaciones Económicas, 32 (2): 231-55.

44.

Niedrich

R. W.

Weathers

Hill

R. C.

Bell

D. R.

(2009). “Specifying Price Judgments with Range-Frequency Theory in Models of Brand Choice.” Journal of Marketing Research, 46: 693-702.

45.

(2003). “Price Fairness and Its Asymmetric Effects on Overall Price, Quality, and Value Judgment: The Case of an Upscale Hotel.” Tourism Management, 24: 387-99.

46.

Park

Ellis

G. D.

Kim

S. S.

Ruddel

E. J.

Agrusa

(2007). “Predictors of Social Equity and Price Acceptability Judgments of User Fees.” Journal of Travel & Tourism Marketing, 21 (2-3): 89-103.

47.

Pearce

D. G.

(2007). “Supplier Selection in the New Zealand Inbound Tourism Industry.” Journal of Travel & Tourism Marketing, 23 (1): 57-69.

48.

Petrick

J. F.

(2004). “First Timers’ and Repeaters’ Perceived Value.” Journal of Travel Research 43(1): 29-38.

49.

Petrick

J. F.

Backman

S. J.

(2002). “An Examination of the Construct of Perceived Value for the Prediction of Golf Travelers’ Intentions to Revisit.” Journal of Travel Research, 41 (1): 38-45.

50.

Petrick

J. F.

Morais

D. D.

Norman

W. C.

(2001). “An Examination of the Determinants of Entertainment Vacationers’ Intentions to Revisit.” Journal of Travel Research, 40 (1): 41-48.

51.

Pritchard

M. P.

Howard

D. R.

(1997). “The Loyal Traveller: Examining a Typology of Service Patronage.” Journal of Travel Research, 35 (4): 2-10.

52.

Quintal

V. A.

Lee

J. A.

Soutar

G. N.

(2010). “Tourists’ Information Search: The Differential Impact of Risk and Uncertainty Avoidance.” International Journal of Tourism Research, 12 (4): 321-33.

53.

Rosen

(1974). “Hedonic Prices and Implicit Markets: Product Differentiation in Pure Competition.” The Journal of Political Economy, 82 (1): 34-55.

54.

Saini

Rao

R. S.

Monga

(2010). “Is that deal worth my time? The interactive effect of relative and referent thinking on willingness to seek a bargain.” Journal of Marketing, 74: 34-48.

55.

Sarigöllü

Huang

(2005). “Benefits Segmentation of Visitors to Latin America.” Journal of Travel Research, 43 (3): 277-93.

56.

Sirakaya

Uysal

Yoshioka

C. F.

(2003). “Segmenting the Japanese Tour Market to Turkey.” Journal of Travel Research, 41 (3): 293-304.

57.

Stevens

B. F.

(1992). “Price Value Perceptions of Travelers.” Journal of Travel Research, 31 (2): 44-48.

58.

Sorensen

(2003). “The Science of Shopping.” Marketing Research, 15 (3): 30-35.

59.

Tan

A. Y. F.

McCahon

Miller

(2002). “Modeling Tourist Flows to Indonesia and Malaysia.” Journal of Travel & Tourism Marketing, 13 (1-2): 61-82.

60.

Train

K. E.

(1998). “Recreation Demand Models with Taste Differences over People.” Land Economics, 74 (2): 230-239.

61.

Train

(2009). Discrete Choice Methods with Simulation, 2nd ed., Cambridge: Cambridge University Press.

62.

Tsaur

S. H.

D. H.

(2005). “The Use of Stated Preference Model in Travel Itinerary Choice Behavior.” Journal of Travel & Tourism Marketing, 18 (4): 37-48.

63.

Tversky

Kahneman

(1991). “Loss Aversion in Riskless Choice: A Reference-Dependent Model.” The Quarterly Journal of Economics, 106 (4): 1039-61.

64.

Walls

A. R.

Okumus

Wang

(2011). “Cognition and Affect Interplay: A Framework for the Tourist Vacation Decision-Making Process.” Journal of Travel & Tourism Marketing, 28 (5): 565-80.

65.

Winer

R. S.

(1986). “A Reference Price Model of Brand Choice for Frequently Purchased Products.” Journal of Consumer Research, 13 (September): 250-56.

66.

Winzar

Pidcock

Johnson

(1993). “Modelling Long Distance Pleasure Travel Mode Using Perceived Modal Attributes.” Journal of Travel & Tourism Marketing, 2 (1): 53-68.